Dwarf Fortress Wiki - User contributions [en]

Minecart

2015-03-23T17:26:55Z

Pirate Bob: added reference for high speed minecarts from collisions

{{quality|Fine}}{{av}}
A '''minecart''' is a [[tool]] intended for [[hauling]], introduced in version 0.34.08. It can be made of [[wood]] at a [[carpenter's workshop]] or [[metal]] at a [[metalsmith's forge]] (using the [[Metal crafter|metalcrafting]] labor.) Minecarts store up to five times as many items as [[wheelbarrow]]s and are quite a bit faster than dwarves hauling objects by hand, but have the disadvantages of requiring a dedicated track network, a complex route planning phase, and the possibility of dwarves [[Fun|blundering into the path of carts filled with lead ore]]. Tracks may be carved into stone, or [[Construction|constructed]]; the latter allows above-ground routes, but these are more difficult to set up due to their additional [[building material|material requirements]].

Just like wheelbarrows, minecarts are considered [[item]]s and are stored in a [[furniture]] [[stockpile]]. Despite their five-times-greater capacity, they are only 33% larger than wheelbarrows and are identical in base [[item value|value]] when made from the same [[material]] (the value may differ due to the [[item quality]]). [[thief|Thieves]] or even mischievous animals can steal minecarts, even when they are moving on a track{{cite forum|109460/3289070}}. However, minecarts moving fast enough or being ridden cannot be stolen.

Although most of the utility of minecarts is in [[fortress mode]], an [[adventure mode|adventurer]] can also ride in a minecart. Adventurers can also pick up and relocate minecarts.

The invention of minecarts revolutionized the [[minecart logic|Science of Dwarfputing]] by enabling smaller, faster logic systems to be built.

== Basic Minecart Usage ==
Minecarts can be used to swiftly transport dwarves, [[flow|fluids]], and/or large amounts of items, but before you have a functional minecart there are several preconditions that need to be met. First of all you need an actual minecart, constructed either in a [[carpenter's workshop]] or [[metalsmith's forge]]. For the minecart to be able to move you also need to carve (with {{k|d}} {{k|T}}) or construct (with {{k|b}} {{k|C}} {{k|T}}) a track, which could be as simple as a straight line. Finally you need to construct stops on your track (with {{k|b}} {{k|C}} {{k|S}}) where the minecart will start and stop.

After you have created the stops and assigned a cart to the track, you must create logic routes connecting several stops and designate starting conditions for each stop. This is done with the {{k|h}}auling key. The most basic conditions are how the cart's movement is initiated and in which direction the cart should start moving. Carts can be either be Pushed (a dwarf stands at a stop and gives the cart a single push) or Guided (a dwarf continually pushes the cart forward, guiding it along the track). The [[hauling]] [[labor]] required for pushing and guiding carts is called "Push/Haul Vehicles" and is turned on by default.

To control which items to transport you can add conditions specifying: (1) which kind of items to be loaded, and unloaded, (2) stockpile links to define which stockpile(s) the items should be un/loaded to and from.

===Capacity and weights ===
Minecarts have five times the [[Weight|capacity]] of [[wheelbarrow]]s.

'''Examples of the capacity of one cart'''
{| class="wikitable"
|-
! Item
! Amount
|-
| [[stone]]
| 5
|-
| [[wood|log]]
| 10
|-
| [[block]]/[[bar]]
| 83
|-
| [[Kitchen|prepared meals]]
| 500
|-
| [[Trap_component#Spiked_ball|spiked balls]]
| 500
|-
| [[Weapon#Native_weapons|mace]]
| 625
|-
| [[Weapon#Native_weapons|spears]]
| 1250
|-
| [[cloth]]
| 2500
|}

The weight of the loaded minecart does not affect the initial velocity received from pushing or launching from a roller. However, the load of a minecart ''does'' affect whether a [[pressure plate]] triggers or not, based on the pressure plate's setting.

'''Weights of different carts'''

{| class="wikitable"
|-
! Type of cart
! Empty cart
! Fully loaded (items)
|-
| oaken minecart
| 28Γ
| 378Γ (10 oak logs)
|-
| platinum minecart
| 856Γ
| 10482Γ (83 gold bars)
|}

The weight of a minecart is one twenty-fifth (1/25) the [[density]] of its material in Urists. Because pressure plates can be set to trigger at intervals of 50 Urists, minecarts with weights just under a multiple of 50 are ideal for switching based on whether they're full or empty. The best minecart materials for full/empty switching are as follows:

{| class="wikitable sortable"
|-
! Material !! Minecart weight !! Content weight required to trigger !! Banana roasts required to trigger (for scale)
|-
| [[Bismuth]] (moods only) || 391 || 9 || 15
|-
| [[Brass]] || 342 || 8 || 14
|-
| [[Electrum]] || 596 || 4 || 7
|-
| [[Fine pewter]] || 291 || 9 || 15
|-
| [[Glumprong]] || 48 || 2 || 4
|-
| [[Lay pewter]] || 291 || 9 || 15
|-
| [[Nickel silver]] || 346 || 4 || 7
|-
| [[Tin]] || 291 || 9 || 15
|-
| [[Trifle pewter]] || 291 || 9 || 15
|}

=== Creating tracks ===
Minecart tracks are made up of contiguous track, tracked ramp, or bridge tiles. Track tiles and tracked ramp tiles have a direction or series of directions associated with them. These directions dictate which directions a minecart on a given tile may move from that tile. For example, a Track NE (northeast) tile allows a minecart on it to move either north or east from its present position. Therefore, if you want your minecart to move east along a straight piece of track, then return west using that same track, you would need to use EW tracks so that the cart could travel east initially, then return west over the same track. Excluding designs in which the cart will "jump" tracks via a drop or other ramp, tracks must be valid end to end to work for most looped or straight-track applications. A single east only track tile in your line of east-west tracks will cause any route using the track to fail the moment it tries to go the wrong way over that tile. Minecart tracks can be built in two ways: Engraved/carved or constructed. A given minecart track need not use engraved or constructed elements exclusively, as the two methods can be used interchangeably depending on the needs of a given section of track. The way the tracks are built is slightly different between the two, as explained below.

====Simple tracks====

'''Carved'''

A single-tile wide strip of natural stone can be designated to be [[Engraver|carved]] (with {{K|d}} {{k|T}}), which will create a straight two-way track. The creation of corners, crossings, and T-junctions is as simple as designating another strip of track that overlaps an existent or newly designated track. Engraved tracks are removed by [[smoothing]] the rock they're on, which results in a smooth floor (that can be re-engraved if necessary), or by building a [[floor]] on top and subsequently removing it. Dwarves can carve corner tracks in one pass by designating the track carving twice and canceling unwanted carvings (with {{K|d}} {{K|x}}). Tracks can be engraved in any natural floor tile, rough, smooth and even over engravings, providing an easy method to remove low-quality or undesired floor engravings. Once a track has been engraved, it's important to check the track directions for each tile in the route carefully to make sure no mistakes were made by yourself or the game's track engraving logic.

'''Constructed'''

Tracks can also be built as regular [[construction]]s (through {{K|b}} {{K|C}} {{K|T}}). This method is resource-expensive, since each track tile requires one stone, [[bar]], or [[block]] for construction, and time-consuming, since you can't designate strips longer than 10 tiles at a time. Corners, crossings, T-junctions, and ramps also have to be designated individually. However, it is usually the only way to build tracks above ground or on soil (barring the [[Obsidian farming|creation of obsidian]]). Constructed tracks are designated for removal like any regular construction; be aware that removing track ramps built on top of natural ones will also remove the original ramp, leaving a flat floor.

====Ramps====

'''Carved'''

The carving of natural ramps is a little more confusing: to carve a two-way track on a ramp (natural only, does not work on constructed ramps), you must designate the track '''starting on the ramp and one square beyond''' in the direction you want the track to go. For the side of the ramp square you want to head upward, there '''must''' be either a natural or constructed wall in the square next to it, otherwise the game assumes you are trying to carve it on the same level -- this can result in the track being carved underneath a door or other object. If you have accidentally done this, you can correct it by smoothing the ramp and constructing a single square of wall next to it, then re-carving the ramp correctly. (However, the wall must stay there permanently; removing it will disconnect the track.)

'''Constructed'''

When constructing track ramps, the stated direction should be the same as the connected tracks. For example, a track going up from West to East would require, starting from the West, a Track (EW), a Track/Ramp (EW) and a Wall behind the ramp, underneath the section of track above it. Incorrectly placed ramps result in minecarts ignoring the ramp and crashing into the supporting wall. They will not, however, display as unusable as when the supporting wall is missing.

'''Examples of ramps'''

A simple ramp would look like this:
{{diagram|spaces=yes|\
z +0 z +1
░░░░ ░░░░
═▲o ░▼═
░░░░ ░░░░
o : wall
}}

Carving track corners into ramps is rather unintuitive and complicated. Since engraving tracks always requires two tiles to connect in a straight line as input, you have to give two separate designations for a single job: a track bit from the ramp tile to the "below" direction and another one to the wall of the "upward" direction. If you wanted to change direction on a ramp from east to north:
{{diagram|spaces=yes|\
z +0 z +1
░░░░░ ░░░░░
░░░░░ ══╗░░
=▲░░ ░░▼░░
░░░░░ ░░░░░
}}
you would need to connect the ramp on z +0 both to the west and to the north by issuing two "carve track" commands, one selecting the ramp and the track tile to the west, and another connecting the ramp tile with the wall to the north. An engraver would then carve a NW track corner into the ramp, allowing carts to pass the corner correctly both going up and down. Such track corners are perfectly serviceable for guided carts, but moving down a route of several of them by pushed or ridden cart is problematic - ramps on corners behave very counter-intuitively, resulting in loss of speed when going down and diagonal movement when going up.

Moving to and from ramps (or between ramps "pointing" in different directions0 causes some non-trivial adjustments to speed and even moving along the tiles at a fixed speed ''unrelated to the entry/exit velocity values'', because transitions to/from ramps are processed differently and are not to be "skipped". This affects compact track/ramp combinations (such as e.g. a simple 2x2 ramp spiral) most, and combined with bouncing often makes them work not in the way one could expect. {{cite forum|144328/5705102}}

{{anchor|Tracks}}

=== Hauling route ===
A hauling route is a list of directions describing how and under what conditions a minecart will move. The proper setting up of routes is essential for a working rail system. Routes, stops, departure conditions and stockpile links are managed from the {{k|h}}auling menu.

==== Route ====
A route defines the path a minecart will take along a track, as well as under what conditions it will move or stop moving. A route is made up of stops. Stops are precisely what they sound like, a position on the track at which you want a minecart to stop. A minecart track might use as little as a single stop for a looped track, which will serve as both a starting and stopping point for the cart, or it could contain many stops, perhaps to load supplies or wait for a bridge to be manually lowered, before reaching its destination or returning to its starting point. It is important to note that you only need to place stops on a route where you actually want the cart to stop and wait for some action to occur. They are not needed to help navigate the cart along the track beyond telling it where on the track to stop.

New routes are created with the {{k|h}}auling key. Existing ones can be removed (without confirmation) with the {{k|x}} key, and also {{k|n}}icknamed. Before operating, the route must have a {{k|v}}ehicle assigned to it (this can be done with either the route or a stop selected). Assigning a full minecart to a route may result in a slow hauling job if the contents are heavy.

==== Stops ====
Stops are the individual waypoints that make up a hauling route. A given stop consists of the location of a tile, as well as conditions describing when, where, and how a cart should be moved after being stopped at that tile. Stops can be created from within the {{k|h}}auling menu, by placing the cursor over a tile and hitting {{k|s}} while highlighting the route (or a stop within) you've already designated. A minecart will begin its route at the first stop created, and continue through each subsequent stop, being guided, pushed, or ridden from each stop to the next depending on the conditions specified. In many basic minecart applications, the cart will end up at the same stop it began at, though this is not always the case. It is important to note that hauling stop order is enforced, even if there is no track. A dwarf will drag the cart overland back to a skipped stop in the route's list if your tracks bypass it somehow.

Once a stop has been placed, it is given a default set of conditions under which to move the minecart if it is stopped there. Each new stop gets the same default conditions regardless of the track it is placed upon (e.g. guide the cart to the north). For this reason new stops might get marked by yellow exclamation marks ({{DFtext|!|#ff0}}) due to invalid directions. One important thing to note is that as you place additional stops, the display will show paths between the stops you have defined. However, this is '''not''' necessarily the actual route the minecart will take once the route is in operation. For example, if a route were defined with two stops at opposite ends of a track with many twists and turns, a line will be drawn directly between those stops to show the order in which they will be visited. These route lines may crisscross all over the tracks, but so long as the track is valid end to end, the cart will follow the track from one stop to the next, even across twists, turns, and z-level changes. Stops, which are the steps that make up a route, should not be confused with Track Stops, described below.

===== Stockpile links =====
By placing the cursor on top of a stockpile and using {{k|s}}, you can create stockpile links while defining a hauling stop. Links can also be redefined by selecting them, placing the cursor over a different stockpile, and pressing {{k|p}}. The cart will then be filled by items present in its various linked stockpiles in preference to other items. Note that bins should be used with caution in stockpiles that are linked to minecarts. Bins cause problems when used with the "Desired Items" list in a stop's conditions. For example, if a minecart is set to accept only granite blocks, and to depart north when it is 100% full of granite blocks, it will not depart if any of those granite blocks are in bins, even if bins are also included in the desired items list. Two solutions to this problem exist as of v0.40.24. First, bins can be disallowed in stockpiles that are linked to stops. Alternatively, bins '''can''' be used in conjunction with minecarts provided that the minecart's departure conditions use only "any items" instead of "desired items." This option can be toggled in the advanced conditions menu for a stop, accessible via the {{key|C|}} key. The cart's contents can still be controlled by specifying what items are allowed in the linked stockpile.

===== Departure condition =====
Departure conditions involve setting conditions in which the minecart will leave on the route. Each condition includes:
# A departure mode (Guide, Ride or Push).
# An initial departure direction (NSEW). Note that this defines the initial direction of movement only. Even if a track includes many turns, as long as the initial movement direction is valid the cart will follow the minecart track thereafter.
# A timer, before which the departure condition cannot be met.
# Conditions on the amount of items in the cart.
Departure conditions are created with the {{k|n}} key. A new departure condition will read: "guide north immediately when empty of desired items". This condition can be changed between basic presets with {{k|c}}. "Advanced" mode ({{k|C}}) allows for more precise control over departure conditions: fine tuning the percentage from 0 to 100 in 25% steps ({{k|f}} and {{k|F}}), switching it being either the maximum or the minimum amount of items for the condition to be met ({{k|m}}), and whether the cart accepts all or only a specific set of items ({{k|l}}). Common to both screens are the departure mode ({{k|p}}, Push, Ride or Guide), {{k|d}}irection, and timer ({{k|t}} and {{k|T}}) options.

To have a cart only carry a specific set of items, the stop can be set to only carry "desired" items, opening the selection screen with the {{k|Enter}} key while having said stop condition selected, and toggling as desired, or it can simply be linked to a stockpile and set to depart once it is full of items from its linked stockpiles, regardless of type.

=== Track Stops ===
A Track Stop, not to be confused with a route stop, is an optional, single-tile construction which serves two purposes. First, it can be used to cancel a cart's momentum in order to slow or stop it as it passes over the Track Stop. This might be necessary if a cart were pushed down a series of ramps to its destination. Second, a Track Stop can cause a cart to automatically dump its contents as it passes over the Track Stop. Track Stops are constructed via {{k|b}} {{k|C}} {{k|S}}, and must be constructed atop an existing piece of track. If a Track Stop has been set to automatically dump a cart's contents, the cart will dump its contents in the direction indicated when it passes over the Track Stop. Depending on the friction settings chosen for the Track Stop, the cart might then stop after dumping, or it might continue on its route to another destination.

Track Stops are not mandatory; in fact, their main use is in automated rail systems. However, even in basic rail systems it can be useful to set a Track Stop to dump items: this saves time that dwarves would otherwise spend in removing items from the cart, time that is better spent driving the cart back to where it's needed. Dumping will occur even with a guided cart. '''Take care not to set Track Stops at a loading site to dump their contents''', or dwarves will never be able to fill the cart. It will dump any contents the moment they are loaded.

Counter-intuitive to their construction method, Track Stops are considered [[building]]s and must be removed by {{k|q}} {{k|x}}.
* See [[#More_on_Track_stop |More on Track Stops]]

=== Step-by-step tutorial ===

Let's construct a simple minecart route. This route will move stone blocks from an input stockpile to an output stockpile. We'll begin by creating the stockpiles:

[[File:minecart-example-1.png|Stockpiles designated.]]

The input stockpile is on the left; the output stockpile is on the right. We'll be moving blocks from left to right. Disable bins in both stockpiles, and set the input stockpile to accept only from links. Then make the stockpile take from the mason's workshop where the blocks are being produced.

Next, carve the track:

[[File:minecart-example-2.png|Track carving designation.]]

Note that the ends of the designation are uniquely shaped; this is automatic, and not anything you need to control. Now, wait for your engravers to come along and carve the track into the stone. (Your haulers will probably also fill up the input stockpile while you wait.)

In addition, while we're waiting for that to happen, we'll build an iron minecart in the forge.

[[File:minecart-example-3.png|Track carved.]]

When the track has been carved, it will look like the above (the track will be solid instead of flashing). Now, order a track stop to be constructed next to the output stockpile:

{|
| [[File:minecart-example-4.png|Track stop designation.]]
| [[File:minecart-example-5.png|Select dumping direction.]]
|}

You must press {{k|d}} three times to select the dumping direction ''before'' placing the track stop. We want our blocks to be dumped into the output stockpile east of the track stop. Then wait for a mechanic to come along and build the track stop.

[[File:minecart-example-6.png|Track stop constructed.]]

Now we'll define the actual ''route''. This is done in the {{k|h}}auling menu. Press {{k|r}} to begin defining a route. Next, move the cursor to the input end of the track, and then press {{k|s}} to define the first stop:

{|
| [[File:minecart-example-7.png|Stop 1 designation.]]
| [[File:minecart-example-8.png|Route definition, in progress.]]
|}

Move the cursor again, to the output end of the track, and press {{k|s}} again to define the second stop:

{|
| [[File:minecart-example-9.png|Stop 2 designation.]]
| [[File:minecart-example-10.png|Route definition, two stops.]]
|-
|
| [[File:minecart-example-11.png|Stops are not defined yet.]]
|}

There are several user interface features to note at this point. The stops have been positioned, but they haven't been ''defined'' yet, so there is a warning {{DFtext|!|#ff0}} symbol by each of them. In the lower right corner, we see what the {{DFtext|!|#ff0}} means. Also, note that the second stop is labeled in white, while the other two lines are grey. The white text is a selection indicator, and can be moved up and down by pressing {{k|+}}/{{k|-}}.

Next we need to define what our stops do. We want the minecart to be filled with blocks at the first stop, then travel to the second stop where it will dump its cargo, and then return. Press {{k|-}} to move the selection up to stop 1, and {{k|Enter}} to open it up. By default, the stop has three conditions:

[[File:minecart-example-12.png|Default stop definition.]]

We don't want any of these, so press {{k|x}} three times to delete them. This leaves us with a blank stop. Now we can add the conditions we actually want. Press {{k|n}} to begin adding the first condition, then {{k|d}} twice to change the direction from north to east. Then press {{k|c}} to change the condition from empty to full. This will instruct the minecart to be guided east when full of desired items.

To set the desired items, we create a stockpile link. Press {{k|s}}, then move the cursor to the input stockpile, then press {{k|p}} to select that stockpile. Now press {{k|Enter}}; this opens up a selection screen that resembles the stockpile customization screen. Move down to Blocks, {{k|e}}nable them, then (if you wish) restrict it to stone blocks.

When you've done all that, stop 1 should look like this:

[[File:minecart-example-13.png|Stop 1, defined.]]

Stop 2 is much simpler. All we need to do is have the minecart return to the input stop. So, make a condition and change the direction:

[[File:minecart-example-14.png|Stop 2, defined.]]

Finally, we just have to assign our minecart. Go back to the route definition screen, and press {{k|v}}. Select the minecart, and press {{k|Enter}}.

Now we've got everything set up:

[[File:minecart-example-15.png|Route, fully defined.]]

The V is red because the minecart hasn't been moved onto the track yet. Some dwarf will have to haul it from the forge to the first stop, by hand; this will take a while, especially if the forge is far away.

Once the minecart is in place, dwarves should fill it with blocks from the input stockpile, which will in turn be filled with blocks from the workshop where your mason has been toiling dutifully. When the minecart is full, the blocks will be dumped into the 1x1 stockpile on the right. Automatic quantum dumping!

=== Troubleshooting ===

Because of the complexity of the system, all but the most careful and experienced minecart users will encounter issues. Most route issues can be diagnosed and fixed from the {{k|h}}auling menu.

'''Symptom:''' {{DFtext|! Set dir/connect track|6:1}} message appears to the right of one or more stops

'''Possible Causes:'''
* The departure direction of the stop might be invalid. Edit the stop using {{k|Enter}} and press{{k|d}} until it is pointing in a valid direction.
* The track stop might not be built on top of a track. The track stop must be deconstructed to remedy this issue.
* Your track might not be built correctly. Make sure all connected tracks between destinations are not one-way tracks.
** This can be especially confusing with ramps. To carve a two-way track on a (natural) ramp, you must designate the ramp and one square beyond in the direction you want the track to go.
** Ramps '''must''' have a solid block on the side opposite to the track, or they will neither work nor be marked as "unusable". The solid block can be natural or constructed.
* The desired/kept items might not be configured correctly.

'''Symptom:''' The status '''0% V''' always appears to the right of one stop.

'''Possible Causes:'''
* The stop may not be set to take from a stockpile. Edit the Stop using {{k|Enter}} and make sure you see a message like "Take from Stockpile #1".
* The take conditions must correspond with the contents of the stockpile.
* The track stop may be set to dump. A track stop set to dump cannot be filled. You must either set the stop to a time-based departure or deconstruct the track stop and rebuild it without dumping. (Alternatively, with [[DFHack]] you can modify "Dump on arrival" to "No" using the {{key|q}} menu without rebuilding the stop.)
* Make sure the minecart itself has not been designated to be dumped (such as when using mass-dump).

'''Symptom:''' Dwarves fill the minecart properly, but will not move it thereafter.

'''Possible Causes:'''
* The minecart may contain items which are not included in its current stop's desired items. Check inside the minecart using the {{key|k}} and {{key|z}} keys and ensure that all items in the cart are desired items.
* The minecart may contain desired items in bins. Minecarts seem to have problems realizing that they are in fact full of desired items if some of those items are in bins, even if bins are also among the desired items for that stop. '''This cannot be solved by adding the appropriate bins to the stop's desired items.''' Either disallow bins in stockpiles you intend to load minecarts from, or set the departure conditions to rely only on percentage of total load rather than percentage of desired items using the advanced conditions menu ({{key|C}} key).

'''Symptom:''' Dwarves repeatedly attempt to load the minecart, but no items are ever loaded into it.

'''Possible Causes:'''
* This can be caused by using a Track Stop with autodumping enabled at a loading site. Every time a dwarf places an item into a cart resting on such a track stop, the item will be immediately dumped, causing unlimited, useless cart loading jobs. Autodumping Track Stops should never be used at a loading site.

'''Symptom:''' A dwarf picks up the minecart and carries it to its destination.
* See [[#Quirks|Quirks]]

=== Danger ===
Minecarts are not without <strike>danger</strike> [[fun]]. Although designating a track automatically sets the [[traffic]] designation to low, dwarves ''may'' still walk on them, and [[creature]]s ignore traffic designations altogether. If an unlucky dwarf or creature fails to [[dodger|dodge]] a minecart, they can be injured. Most of this danger can be avoided by setting the minecart {{k|h}}auling commands to guide instead of push or ride (dwarves guiding minecarts will ignore traffic restrictions), as well as by [[pasture|pasturing]] domestic animals and preventing the access of other creatures to the tracks. Note that removing the track doesn't reset that tile back to normal traffic priority, so you may wish to manually clean up traffic designation afterward. Also note that bridges that are used as tracks don't have their traffic priority changed automatically (since they're just normal bridges), which could cause dwarves to pathfind normally through dangerous minecart entrances in your fort's walls if you're not careful.

The only <s>fool</s>''dwarf''-proof method is to make the tracks inaccessible. There are several ways to create a track which works for minecarts but doesn't allow creature-traversal; the simplest is perhaps building a [[statue]] on the tracks. Other options include adding single-tile holes (minecarts moving at reasonable speed will jump the gap), vertical drops, minecart-triggered doors, small pools of liquid (4/7 water or 2/7 magma), and hostile creatures overlooking the tracks. For safety, both ends of the track should be isolated, making the dangerous center sections completely inaccessible (though maintenance access can be provided by a locked door).

Danger does not always involve living victims: careless route designation can also result in minecarts careening off tracks or colliding with each other. If this occurs, the [[item]]s may be scattered; this can cause even more hauling jobs than the minecart aimed to eliminate. Even <s>better</s> worse, scattered items, especially [[weapon]]s, can injure passing [[dwarf|dwarves]] or other [[creature]]s; in the words of Toady One the Great, "Accidental grapeshotting of the dining room should be possible now."

Of course, the danger of using minecarts means they can also be [[Trap_design#Minecarts|used as weapons]] by imaginative players.

== Advanced usage and automation ==
Minecart-specific effects are implemented via track stops, rollers and [[pressure plate]]s with "track" condition set. Since all three are considered [[building]]s, they can't be built on the same square (however convenient track stop + pressure plate would be) nor a simple ramp, and are removed by {{k|q}} {{k|x}}.

=== More on Track stop ===
Track stops are constructions that allow further automation of minecart systems via adjustable features such as braking by friction and automatic dumping of contents. They can be built from logs, bars and blocks through {{K|b}} {{K|C}} {{K|S}}; friction amount, dumping toggle and dumping direction must be set '''before''' construction, and these settings can be neither changed nor seen thereafter; however, track stops can be linked to [[pressure plate]]s or [[lever]]s to toggle friction and dumping On or Off (trigger state is inverted: switch On = track stop Off).

If a [[stockpile]] is placed on the tile that a track stop is set to dump to, it can act as a [[Exploit#Quantum_stockpiles|quantum stockpile]] and any items dumped from a minecart that match the storage settings of the stockpile will remain there and accumulate. Normally trackstops are built on top of existing track to operate on moving minecarts, but they can also be used without tracks to create [[Exploit#The_Minecart_Stop|automatic quantum stockpiles]] (see also [[#Step-by-step_tutorial|step-by-step tutorial]]). It is not always desirable to collect ALL of certain items into one quantum stockpile, such as when distributing a material to multiple separate industries. You can link your quantum stockpile to various other stockpiles, ensuring that your dwarves will keep them supplied as necessary. Because quantum stockpiles never fill up like regular stockpiles, it may be a good idea to add a switch to turn them off.

Items dumped from a minecart at a track stop (or dumped by any other means) into open space fall through z-levels until they land on a solid surface. Items falling onto a designated [[stockpile]] will automatically be considered part of that stockpile, even if the stockpile is set to disallow those items (they will, however, be automatically moved to a more appropriate stockpile, if available). Items falling on top of a minecart will '''not''' fall "inside" the minecart. Use with caution; dwarves have fragile skulls.{{bug|5945}}

=== Automated propulsion ===
==== Roller ====
{{Main|Roller}}

A '''roller''' is a [[power]]ed [[machine component]] for the automated propulsion of minecarts. They are built over the top of existing tracks with {{K|b|M|r}}, requiring a [[mechanic]], ''(length/4)+1'' [[mechanism]]s and a [[rope]]. Rollers may also be placed directly on ramps to help pull carts up Z levels. Rollers are very useful to maintain a cart's momentum along long routes, to get them to climb Z-levels without dwarfpower involved, and to get them to reach speeds unattainable by guiding dwarves. These devices are variable-length (1-10), variable-direction and variable-speed ([[Minecart#Numbers_behind_the_scene|see below]]), all traits that can be set at construction time; a roller uses two units of power per tile it is long.

Single-tile rollers transfer power in all four cardinal directions, while other rollers generally only transfer power perpendicular to their activity direction. Longer rollers can also transfer power along their activity direction if built in the correct order, although this can be hard to accomplish and is easily broken. Rollers cannot be powered from above.

Rollers have great acceleration and capped speed. Carts going faster than the roller are unaffected. If a cart moves across an active roller in the direction the roller works and moves slower than the roller's specified speed, the cart will be set to the roller's speed. A cart going against a roller's movement direction will be sent back the way it came (once again at the roller's speed), unless it was moving extremely fast: speed increment of 100000 allows to reverse carts from the full "highest" (50000) speed roller to full "highest" speed back, but ramps can accelerate a cart beyond this. {{cite forum|144328/5702453}}
A cart crossing over a roller perpendicular to its current movement direction will gain the roller's amount of speed in the perpendicular direction without directly changing its forward motion. Without an adjacent wall to constrict its movement, this will typically send a cart off the rails on a diagonal path, completely unable to follow any tracks until it collides with a wall or is otherwise brought to rest. However, if the roller is placed over a track turn and pushes ''from'' the direction of that turn's track, the turn affects carts ''after'' the roller, so they will be forced into the turn rather than derailed in a diagonal direction. {{cite forum|144328/5702453}}

{{diagram|spaces=yes|\
tracks: full:
║ ║
═╗═ ═╢═
║ ║

╢ : roller pushing from W to E
}}
If the roller is powered, carts from ''all'' directions (unless too fast) exit S, because speed imparted by the roller forces carts toward E and ''then'' into the turn.
If not powered, carts from W and N exit S, carts from E and S exit W. Carts above derail speed will ignore the turn, of course.

{{diagram|spaces=yes|\
║ ║
═╗═ ═╟═
║ ║

╟ : Roller pushing from E to W
}}
Carts from the E or W: exit W.
Carts from N: derailed diagonally, exit SW.
Carts from S: derailed diagonally, exit NW.

Rollers affects carts on a track - if placed on a floor or ramp without any tracks, they are ignored. Depowered rollers are also ignored, friction is determined by the tiles underneath.

Because of their one-way nature, rollers are unsuitable for most two-way minecart tracks (unless you set gears toggling roller A->B off while toggling A<-B rollers on). However, a minecart set to be ''guided'' is not affected by rollers at all{{cite forum|109460/3286235}} — this allows a one-way track to be used in both directions. In addition, unpowered rollers do not affect minecarts.

Care must be taken in [[glacier]]s and other extremely cold [[biome]]s, since rollers (and the machinery used to power them) will not operate when constructed on natural [[ice]] floors.

==== Impulse ramps ====
Carts can be given momentum without rollers or changing z-level through a phenomenon called "impulse ramps". A track ramp which is connected both to a wall and to a floor will ''always'' accelerate a cart towards the connected floor tile, no matter where the cart enters the tile from. This means carts can be accelerated as though dropping z-levels, even if the cart doesn't actually change z-level at all. If a track ramp faces three directions such as ╩, then two of those directions need to be facing walls for the cart to be accelerated towards the remaining direction.

Example of straight impulse acceleration:

{{diagram|spaces=yes|\
▒▒▒▒▒▒▒▒▒▒ ▒▒▒▒▒▒▒▒▒▒
═▲▲▲▲▲▲▲▲▲▲═ ═╚╚╚╚╚╚╚╚╚╚═
▒ : Wall
═ : Normal track
▲/╚ : N/E Track/Ramp
}}

If a cart enters from the left, it will speed up on every track/ramp and exit to the right going very very fast - more than one tile every step. If it enters from the right then it will bounce back impulsed by the ramp if it's going slow enough.

As another oddity, carts coming from ramps will in some cases "teleport" through most of the next tile. This can be used to send carts up levels with minimal loss of speed. You can just make an upward spiral alternating impulse ramps and regular upward ramps. It takes no power, is quick and cheap to build, requiring only channeling and track carving, and the cart goes up fast, but not so fast that it launches its contents.

Example of an impulse elevator:

{{diagram|spaces=yes|\
z +0 z +1 z +2 z +3
░░░░░ ░░░░░ ░░░░░ ░░░░░
░╔░░░ ░▼╚╗░ ░░▼▼░ ░░░░░
░╝░░░ ░▼░░░ ░░░╔░ ░░░▼░
░▼▼░░ ░░░░░ ░░░╝░ ░╚╗▼░
░░░░░ ░░░░░ ░░░░░ ░░░░░

░ : Wall
╔,╚,╗,╝ : Track/Ramp
▼ : Down Ramp (empty space)
}}

A sort of opposite effect to impulse ramps also exists: ramps lacking the proper "up" and "down" connections are treated as flat track, even if they actually go up or down z-levels. This allows building "anti-impulse" slopes consisting entirely of ramps only connected up, which a minecart can travel up forty levels and more, needing no more than a single push.

=== Controlling traffic ===

==== Switching ====

As constructions or tile features, [[door]]s and other furniture can be built on tracks. A [[door]] or [[floodgate]] can be turned on or off by a [[lever]], effectively controlling the flow of automated minecarts. This may be <s>dangerous</s> [[fun]], however.
{{diagram|spaces=yes|\
->
A ════┤≡════ B
┤ : roller pushing to East
≡ : door
}}
The roller pushes the cart east, but until the "departure condition" is fulfilled, the door remains closed and blocks the path.

[[Bridge]]s can also act as tracks, but only if they're lowered or not retracted. This property can enable levers to turn tracks on and off. However, care should be taken to ensure that such bridges are never operated while a cart is on top of them, as the cart will be flung off the track. It's worth noting that it's often faster, and cheaper, to construct large bridges than long sections of constructed track.

A powered track switch can be constructed by building an "inverted" corner as illustrated below.
{{diagram|spaces=yes|\
B B
║ -> ║
║ ║
════╚═══ ════├════
A C A C
├ : roller pushing to West.
}}

If the cart is pushed East from the stop at 'A' while the roller is activated, it will arrive at 'B'. If the roller is not running, it will arrive at 'C'. The switch works by the roller first reversing the incoming cart's movement and the cart ''then'' following the track corner.

This switch is very reliable, reacts instantly to on/off signals, and carts of any speed can be switched by this design, although very fast carts will require rollers that are several tiles long, up to three. The requirement for power can be inconvenient or impractical. Non-powered solutions may use controlled derailment, or a connecting bridge.
{{diagram|spaces=yes|\
B ╥
║
║
╞════╝ ════╡
A D C
}}
Here the track between A and C is not continuous. The only continuous track is A->B, with a corner (not a T section). Fast moving carts will tend to derail at D and rejoin the track to C. Placing a door at D will prevent the derailment, so the cart continues to B. The door is operated by mechanisms elsewhere (typically, a lever, but some fun can be had with pressure plates).

Since it depends on derailing, this switch requires a very fast cart, faster than what can be achieved with rollers alone. To gain sufficient speed, a cart must be accelerated further, usually by descending several levels or through impulse ramps. The high speed makes the cart much more dangerous and harder to control.

If carts are moving too slowly to derail at the corner, a retractable bridge may be used as a connector between A and C.
{{diagram|spaces=yes|\
B╥
║
║
A╞════bbb════╡C
}}
The bridge must overlap the corner. Bridges behave like a track crossing, allowing carts to pass in a straight line. When retracted, the corner reappears, so the carts will continue to B. Bridges take 100 steps to react to a signal, necessitating rather long "lead times" when switching tracks via bridge.

As mentioned above, special care must be taken to make sure the bridge doesn't change state while the cart is passing over it. Retracting bridges will throw the cart, causing it to stop dead. Raising bridges can even crush the cart.

==== Controlling Speed ====


Minecarts can reach extremely high speeds, especially when descending multiple Z-levels. A minecart will derail at a track corner if its speed exceeds 0.5 t/st (tiles per step), '''unless''' the route in the direction of travel is blocked:

Will derail at > 0.5 t/st:

{{diagram|spaces=yes|\
in ══╗ -> derailing
║
out
}}

Will not derail at > 0.5 t/st:
{{diagram|spaces=yes|\
in ══╗O
║
out

O : wall/column.
}}

This behavior can be used to build a "speed limiter", that will ensure that when a minecart exits it is traveling below derail speed:
{{diagram|spaces=yes|\
░░░░ ░░░░░ ░░░░░
in ═╔═╗░ ░╔S╗░ ░╔S╗░
out ═╬═╝░ out ═╗═╝░ out ═╗═╝░
░╚S╝░ ░╚═╝═ in ░╚S╝░
░░░░░ ░░░░ ║░░░
in
░ : wall
S : Track Stop (High Friction or lower)
}}
If the minecart is traveling below derailment speed, it will not be affected; if above, will be slowed down and checked again. Granted, you could do the same just with track turns, but it may take a lot of turns and time.

Since all the derailings, bounces and ramps can impart a sideway component of speed small enough to start visible drift many tiles away (say, [[Fun|in the middle of a bridge]]), track turns have one more use: forcing the carts to move strictly along the grid directions. Carts passing a turn below derailing speed convert one component of velocity into another, thus eliminating the drift.

=== Loading liquids ===
[[Water]] and [[magma]] can also be loaded into minecarts by submerging them to a depth of at least 6/7 while standing still or moving at speeds of at most 10000. Loading fluids onto minecarts can be difficult because the added friction provided by fluids can stop a cart in a submerged tile. Curiously, filling a minecart with magma does not injure a dwarf ''riding'' it. A minecart will hold enough fluid to increase the depth of a single tile by 2. This amount is listed as 833 units, which weigh 459Γ (water) or 999Γ (magma). An iron or steel cart filled with magma weighs 1313Γ, while an adamantine cart filled with magma weighs 1007Γ. Since you need a minecart above the liquid's level, possible arrangements may include pressure-activated sluices, rollers (with magma-safe chains for magma), pouring from above to "submerge" it briefly on the same level and drain excess away (dig deeper and leave a vaporizer, though if you could have power for rollers, may as well use a pump) and exploits with ramps (not necessarily impulse ramps, "same height" passing dip does it).
The liquids can be dumped by a constructed track stop.

== Quirks ==
This little quirk concerns dwarf-managed minecarts. If a track which was previously open becomes blocked (ex. flipping a switch connected to a floodgate you've built on the track to raise it) and the conditions for departure are met, instead of refusing to ride/guide the minecart or ride/guide it until it reaches the obstacle, the dwarf will pick up the minecart off the tracks and haul it to its scheduled destination on foot. If the distance is long enough and the weight of the cart heavy enough (due to being filled with heavy items such as stones), the dwarf may drop the cart because of fatigue/hunger/thirst before reaching the destination. This will cancel that vehicle setting job and make another dwarf come by and attempt to haul the cart to the nearest appropriate stockpile where another dwarf will pick up the cart and attempt to haul it to its initial stop. If the stockpile is far enough from initial stop, this second dwarf who is attempting to place the minecart on its tracks may also drop the minecart out of fatigue/hunger/thirst creating a loop that will go on until a dwarf with enough endurance manages to place the minecart where it belongs.

In fact, it seems dwarves are more than happy to attempt to carry a minecart from one stop to another even if just waiting until the track is open again would be the more sane option.

Dwarves will also carry a minecart to its next stop if the direction specified is incorrect (or invalid). This can often occur when using the default departure settings and forgetting to set the direction of each condition.

Dwarves can admire buildings while riding mine carts. Dwarves will not fall asleep during a ride (at least not from being drowsy). If riding on a continuous powered track loop, the dwarf will die of dehydration/starvation as they can not jump off to get sustenance{{cite forum|109460/3377228}}. Dwarves riding in submerged minecarts will gain experience in [[swimming]].{{cite forum|129889}}

Tracks block wagon access to trade depots, unless they're on a ramp. [[Bridge]]s can also be used, as they function as tracks but do not block wagons.

== Physics ==


Minecart physics depend greatly on the departure mode set in the route stop conditions.

When set to "Push" or "Ride", minecarts will move according to the regular laws of momentum, gaining speed when going downhill, losing it slowly due to friction when on a flat plane, and more quickly when going uphill. In these modes, minecarts will move in a straight line until they either are brought to a stop by friction or an obstacle, or until they encounter a turn. A minecart will roll straight past "blocked" ends of T-junctions or track ends, they have no power to restrict a cart's movement. The cart's behavior is largely independent of the weight of its contents (including fluids and dwarves): heavily loaded carts gain more momentum when accelerating, but this only plays a role in collisions: a heavy cart gains just as much speed and is as easy to stop as a light one. In either case, dwarves can not push nor ride an unpowered cart up a ramp, bouncing back the direction it came. At best, this is a waste of time; at worst, it will give your cart-pushing dwarf a [[fun|fun surprise]]. To solve this, the player can either use Rollers (see below) or set the cart to be Guided.

The difference between "Push" and "Ride" is whether the dwarf will go along with the cart or not.
{{DFtext|Push}}: the dwarf will give the cart an initial push, not enough to go up a ramp, but enough to go some way along flat track, and the dwarf will remain at the first stop, ready for a new job.
{{DFtext|Ride}}: the dwarf will give the cart the same initial push and then hop aboard the cart riding with it to the next stop.
{{DFtext|Guide}}: minecarts seem to ignore all laws of physics. That is:
*Ignore the weight of any and all items inside. Therefore:
**Move at the speed of the dwarf that is guiding them. It is thus recommended to pick the most [[attribute#Agility|agile]] of your dwarves for cart-guiding tasks.
*Ignore working rollers.
*Will ''not'' collide with other guided carts even when a full frontal collision would be expected.
*Will go up ramps like nobody's business.
This is therefore the recommended method of transport for simple non-powered rail systems, despite it diverting a dwarf from other, potentially more important tasks.

Some samples with behavior:
{{diagram|spaces=yes|\
A <-> B A <-> C A <-> B
B B B
║ ║ ║
A══╝ A══╩══C A══╬╗
You can only go A->B ╚╝
Works when the cart Works
is in Guide mode.
}}
In the second example above, a cart "pushed" from B will go over the junction and roll off into the unknown south.

=== Skipping ===
If a minecart is moving fast enough, it can skip over [[water]] or [[magma]], making splashes of [[mist]] (or [[magma mist]]) as it attempts to move on them horizontally. This horizontal movement is independent of the minecart and its content's [[weight]].

=== Track Jumping ===
If a minecart encounters the end of the track or a T junction with no "exit" in its movement direction, it will simply leave the track and continue on its course in a straight line until it encounters an obstacle, slows to a stop, or encounters another (properly aligned) Track even if the tile at which it joins the new track instantly sends it around a corner.

=== Falling ===
When falling, a minecart appears to cause no damage upon collision, possibly to allow cart "stacking" across Z-levels.{{cite devlog|2012|04|06}} A dwarf riding in a minecart that is dropped multiple z-levels suffers normal fall damage. Minecarts can fall through up/down stairs.
The cart falling (from a hatch, thus with no horizontal speed) onto a track ramp is accelerated as if starting from the middle of the ramp - i.e. to the same speed, no matter how many Z-levels it was dropped, vertical velocity is negated. {{cite forum|144328/5701211}}

=== Stacking ===
If a minecart lands on top of another minecart, they may form a stack, with the upper cart on the z-level above the lower. Subsequent carts do not form a stack, but rather quantum stockpile in the same space. This behaviour is useful for [[megaprojects]] and [[trap design]] with minecarts as the weaponry. Moderation should still be exercised: carts take longer to fall into a "stacking" tile already occupied by other carts and will spend that time "hanging" in the air above the stack. This can lead to following carts striking them, which can cause all kinds of malfunctions. The extra time is two game steps for every cart already in the stack, which doesn't hurt stacks of ten carts very much but makes stacks of 100+ rather impractical.

These minecarts on the upper level generally need to be struck with another minecart to move out, or have their support removed. The latter option is safest done by shooting it away with another minecart, manual removal of a stack-supporting cart typically causes the next cart from the stack to [[fun|fall on top]] of the hauler.

=== Numbers behind the scene ===

According to early research by '''expwnent'''{{cite forum|112831/3536975}}:

The minecart has a variable for speed. Speed is measured in tiles/100000 ticks, so a speed of one hundred thousand means one tile per tick. By going down a large number of ramps, a maximum speed of 270,000 can be reached, which presents the limit for most practical applications. Short bursts of (much) higher speeds are possible through carefully planned collisions of high-speed carts {{cite forum|137557/5145499}}.

Every tick the cart accumulates distance units, as well as slows down depending on current tile (speed is reduced by "friction" of the tile). The cart will move to the next tile the tick before accumulating 100000 distance units, (or several tiles in case of great speed), then the leftover distance units are added to the default 100000 distance u. of the next tile. Since most deceleration and acceleration is applied per step, with the notable exception of corners, a cart going at twice the speed of another one can cover about four times the distance in a straight line, but only twice the distance along a winding track with very many corners.

A push will teleport a cart to the beginning of the next tile (NOT the middle!) in one tick with 19990 speed (10 speed is lost due to track friction), while a roller will give a cart the roller's set speed, and it will start to accumulate regular track friction past the middle of the roller tile. Some track features will affect a minecart when it is past the middle of the previous tile: entering a ramp or a hole/drop will happen when the cart has left the middle of the previous tile, and the ramp will gain additional distance unit depending on the leftover units from the previous tile. When a cart leaves a ramp it will emerge after one tick in the middle of the next regular tile, so its entry coordinate is "50000-speed+friction". Rollers also affect the speed of minecart from the middle of the previous tile.

Friction of tiles:
{| class="wikitable"
|-
! Tile
! Friction
! Comment
|-
| Tracks
| 10
|-
| Ground/Floor
| 200
|-
| Unusable ramp
| 10
|-
| Upwards ramp
| 4910 (10+4900)
|-
| Downwards ramp
| -4890 (10-4900)
|-
| Roller
| ±100000 (but capped by the set speed)
|-
| Corner track
| 10
| Speed reduced by 1000 upon leaving the corner tile
|-
| Track stop (highest)
| 50000
|-
| Track stop (high)
| 10000
|-
| Track stop (medium)
| 500
|-
| Track stop (low)
| 50
|-
| Track stop (lowest)
| 10
|-
| Water 1-6
| Additional (WaterLevel - 1) * 100
| rowspan="2" | [[#Skipping|See Skipping]]
|-
| Magma 1-6
| Additional (WaterLevel - 1) * 500
|-
| Empty space
| 0
|}

Water of depth 7/7 provides a friction of about 10000 per step, as does maximum-depth magma. This higher friction may not apply to very slow-moving carts.

Impulse sources:
{| class="wikitable"
|-
! Feature
! Speed
|-
| Push
| 20000
|-
| Roller lowest
| 10000
|-
| Roller low
| 20000
|-
| Roller medium
| 30000
|-
| Roller high
| 40000
|-
| Roller Highest
| 50000
|-
|}

== Non-standard uses ==
Minecarts include some interesting characteristics that have motivated uses beyond hauling. They can be useful for creating fully-automated [[exploit|quantum stockpiles]] and [[garbage disposal]]s. Storing perishable goods (meat, meals, etc.) inside a minecart appears to guard against rot and vermin.
Minecarts can be [[Trap_design#Minecarts|used as weapons]], or as (hopefully non-fatal) triggers to restart stalled [[healthcare]]. They can also be used to time/control game events, either using a basic [[repeater]] or much more advanced [[minecart logic]].
Minecarts trigger [[pressure plate]]s, which means a trap can be designed to trigger when a thief attempts to steal a minecart.
A pressure plate can be used as automatic and more precise custom "launch when full enough" system - as long as weight of your minecarts stays the same. You cannot build a hatch or roller on the same tile, so launch by bumping with another cart. {{cite forum|15096/4580050}}
Dwarves riding minecarts can attack enemies within reach (which goes back to dev log). This applies to shooting, and they actually can hit targets while riding by.{{cite forum|109460/5266119}} Whether a minecart protects the rider and how it interacts with dodging is not known yet. Minecart riders can also [[Swimming#Minecart_training|train swimming]] and [[Megaprojects#Surveillance_Track|detect ambushers]].

== Adventure mode ==
In addition to being used for hauling, minecarts can also be ridden in [[adventure mode]]. (Adapted from forum thread {{cite forum|122903/4258212}})

# If the minecart is in your inventory, drop it. If it is already on the ground, proceed to step 2.
# Press {{k|u}} when you are 1 tile away from the minecart (or standing on the same tile as the minecart).
# You will be presented with the following options:
[[File:minecart adventure mode menu.png|left]]
{{clear}}
* If you {{DFtext|Push}} the minecart, it will move a few tiles in the direction you chose. Physics comes into play here, so it will gain/lose speed depending on the usual factors.
* If you {{DFtext|Ride}} the minecart, you will hop into the minecart, even if you were a tile away, and it will move in the chosen direction with you in it. It will gain/lose speed depending on the usual factors. Whilst the minecart is in motion, you should press {{k|.}} to skip your turn; if you attempt to move whilst the minecart is still in motion, the laws of physics come into play, and you will take [[wound|damage]]. Alternatively, you can push the minecart whilst it's still in motion (although it's unclear how one can bend [[physics]] so as to push a moving minecart whilst inside the minecart). If you push it in the same direction you are already travelling in, you will greatly increase the minecart's velocity. You can also push it in different directions, and this will cause it to gradually change direction-the amount of pushes this requires depends on the minecart's velocity. Once the minecart has stopped moving, you may move out of it safely, or you may want to give it another push. Note that if you push a minecart right after having ridden it (still on the same tile as the minecart), it will act as though you chose to ''ride'' it.

If you want to test this out without creating an adventurer, the [[object testing arena]] allows you to spawn minecarts ({{k|k}}-{{k|c}}-{{k|n}})

== Forging and Melting ==
* Metal minecarts cost '''two''' [[metal]] bars to forge, or '''six''' [[adamantine]] wafers.
* When a non-adamantine metal minecart is melted down, it will return '''1.8''' metal bars, for an '''efficiency of 90%'''.
* When an adamantine minecart is melted down, it will produce '''1.8''' wafers, for an '''efficiency of 30%'''.

== See also ==
* [http://www.bay12forums.com/smf/index.php?topic=109460.0 The "How Does Minecart" Thread] by '''Girlinhat''' et al.
* [http://www.bay12forums.com/smf/index.php?topic=112831.0 SCIENCE: Quantifying minecart physics] by '''Snaake''' et al.
* [http://www.bay12forums.com/smf/index.php?topic=129676.0 How to build a Multi-cart Ore to Magma Minecart Project without needing power] by '''WanderingKid'''.
* [http://www.bay12forums.com/smf/index.php?topic=144328.0 My very own Minecart Education Thread. Ten Lessons, now complete.] by '''Larix'''.

== Bugs ==
*A dwarf will drop her [[child|baby]], if she has one, when boarding a minecart set to be ridden.
*Dwarves have no concept of traffic safety and will walk into busy minecart lines to retrieve objects, often with deadly consequences. This is especially problematic in [[Swimming#Minecart_training|clever applications]] depending on dwarves riding the carts very frequently, because they have a bad habit of dumping their worn clothes on the tracks after a minecart ride. Adding an automatically-operated [[hatch cover]] at the end of such a ride can help prevent [[unfortunate accident]]s.
*Dwarves cannot guide a minecart through an unlocked door unless another dwarf opens the door.{{bug|6056}}
*It is possible for a creature and minecart moving towards each other to pass without collision if they exchange tiles in the same tick.
*After a minecart ride, a dwarf will sometimes haul the minecart to a storage stockpile, leaving another dwarf to haul the vehicle back to the route.
*Minecarts falling onto a floor injure creatures in the tile below the floor.{{bug|6068}}
*A minecart's initial velocity is not affected by weight, when pushed or launched from rollers.{{bug|6296}}
*Removing a stop that has a vehicle waiting on it may cause the game to crash.{{bug|5980}}

{{Category|Fortress mode}}
{{Category|Interface}}

[[ru:Minecart]]

DF2014 Talk:Material science

2014-07-29T15:48:17Z

Pirate Bob: Created page with "The projectile physics changed completely in 40.05, so this page needs to be updated. I made a note of this on the page. I will try to do updates after I (or others) finish ..."

The projectile physics changed completely in 40.05, so this page needs to be updated. I made a note of this on the page. I will try to do updates after I (or others) finish testing the new projectile physics, unless someone more talented at writing would like to update the wiki. --[[User:Pirate Bob|Pirate Bob]] ([[User talk:Pirate Bob|talk]]) 15:48, 29 July 2014 (UTC)

v0.34:Material science

2013-06-14T00:37:57Z

Pirate Bob: adamantine bolts won't deflect off other armor, as bolts won't deflect if they are made out of a superior material

{{quality|Fine|02:27, 11 May 2012 (UTC)}}{{av}}
[[Material|Materials]] have a number of properties representing real world variables that describe how they respond to inputs. In particular, the game now has a number of variables that describe what happens to a material when it's put under stress.

==What is stress?==
In the real world, an object is stressed when a force is applied to the object. Depending on the nature of the force applied, this stress can take a number of forms, and the object can respond differently based on its material and how that material handles different stresses.

In the material raws, whenever you see 'yield', 'fracture', or '[[Strain at yield|strain at yield]]', that property is a stress-related quality.

==When does Dwarf Fortress make stress calculations?==
At present, DF seems to only apply forces during combat, and thus only stresses objects (generally armor and various body layers) at that time.

==There's a lot of stress-related properties, what do they mean?==
The first thing you'll notice is that the second word in each stress variable is one of Yield, Fracture, or strain at yield. These are mechanical performance terms.

The first set of words are things like Impact, Bending, and so forth. These describe modes of applying force.

The following explanations assumes real world physics sort of apply (since [[Main:Toady One|Toady One]] chose real world properties). The game doesn't use all of these properties yet, and may not be applying them according to real world physics.

===Mechanical Performance Properties===
Yield: This is almost certainly 'Yield Strength', which is the amount of stress needed to cause a material to go from elastic deformation to plastic deformation. (That is, if you cease stressing the object, does it revert to its original shape or not). Since most objects only elastically deform over small distances of deformation, high Yield values generally means it takes a lot of force to noticeably 'stretch' them (but see strain at yield).

Fracture: The fracture point is the amount of stress or force necessarily to cause the material to fail, or in other words, to break.

Strain at yield (sometimes incorrectly referred to as 'elasticity'): This variable tells you how much deformation occurs to the material while it is deforming elastically. That is, as long as the force is less than the yield strength, stress * strain at yield = deformation distance. The smaller the strain at yield, the less deformation occurs under stress.

Note: Strain at yield is the inverse of the Elastic Modulus. Thus a highly elastic material has low elastic modulus, and engages in less elastic collisions.

===Modes of Applying Force===
Impact: Force applied by a sudden strike, like a hammer.

Compressive: Force applied by exerting pressure on an object, like trying to squish something between your hands.

Tensile: Force applied by pulling on something, like suspending one object via another. (e.g., if you suspend an [[elf]] from a metal pole, you are applying a tensile force to the pole).

Torsion: Force applied by twisting something. Note that you're twisting some portion of the object relative to itself to cause a torsion stress to be applied to it. (Consider trying to twist a metal rod by grasping at either end and attempting to wring it - yes, you'd have to apply a lot of force to succeed).

Shear: Force applied by pushing part of the material so it tries to slide relative to another part of it. Ie, pushing at the top of an object when the bottom part is fixed to the ground is going to primarily apply a shear stress to it (the top part will try to move in the direction you push, and the lower part will resist this shear stress).

Bending: Force applied by bending a material.

==Effects on Combat==
The Dwarf Fortress combat system does not use all material properties at present (0.34.11). Weapon and armor damage/wear/decay is not tracked.

The mechanics governing how material properties impact the protection armor provides against projectiles have been [http://www.bay12forums.com/smf/index.php?topic=116151.0 studied reasonably carefully], and are described below. Armor protection from melee weapons has not been studied in detail, although it likely shares many similarities with ranged weapons. '''''Note: all of the below is based on recent (0.34.11) experiments, and many results may be approximate or incomplete.'''''

===Projectile Properties===
The main factor determining if a projectile penetrates armor is its momentum. Material properties of the bolt/arrow (except for IMPACT_YIELD, which will be described below), as well as bolt quality and marksdwarf/archer skills do not appear to matter (the skill of the marksdwarf determines if the bolt misses, hits, is dodged, or is blocked with a shield). Dwarven weapons launch projectiles with roughly constant momentum. However rounding of projectile masses can result in some very strange behavior. The complete calculation used to determine projectile momentum is as follows:
*Find the mass of your projectile. Mass=SOLID_DENSITY*SIZE/1,000,000.
*Floor that number to an integer.
*Divide the SHOOT_FORCE of the weapon by the integer, and round the result to the nearest integer.
*Cap the results by the SHOOT_MAXVEL for the weapon.
*Multiply the velocity by the ''unrounded'' mass of the projectile to get the momentum.
For DF34.11 values, iron, bronze, copper, silver, and steel bolts/arrows all have mass between 1 and 2, which will be floored to 1, meaning that they will have velocity equal to SHOOT_FORCE (so long as it is less than or equal to SHOOT_MAXVEL). Wood, adamantine, and bone bolts/arrows, as well as all blowdarts, have mass less than 1, and will have velocity which is always equal to SHOOT_MAXVEL and does not depend on SHOOT_FORCE.

===Interactions Between Projectiles and Armor===
[[File:projectile.png|200px|thumb|right|A simplified flowchart showing how material properties are used as a projectile contacts armor.]]
When a projectile strikes armor, there are several possible outcomes:
*Conversion of edged damage to blunt damage (chain mail does this).
*The projectile absorbs the force of the collision, and is deflected.
*The armor absorbs all or part of the force of the blow.
*The armor is fractured and does nothing to stop the projectile.
For armor to be at all effective at stopping projectiles, the armor material must have SHEAR_YIELD and/or SHEAR_FRACTURE greater than or equal to the projectile material. Otherwise, the projectile just cuts through the armor like it isn’t there. Additionally, the IMPACT_FRACTURE of the armor must be large relative to the projectile momentum, or the armor fractures and the projectile passes through without slowing down significantly (this happens for metal bolts against any plate armor in 0.34.11).

====Conversion of Edged Damage to Blunt Damage====
If the momentum of the projectile is not too high, then chain armor can convert the edged damage normally caused by projectiles to blunt damage, resulting in chips, fractures, jams and bruises, but no tears or cuts. Plate armor does not appear to provide this type of protection under any conditions tested so far. It is not yet known how the momentum needed for edged damage to penetrate chain armor is calculated, but for the cases examined the momentum needed is many times larger than those observed for in-game projectiles.

====The Projectile Absorbs the Force of the Collision====
If the projectile does not cut right through the armor (armor material must have SHEAR_YIELD and/or SHEAR_FRACTURE greater than or equal to the projectile material), then it must push in a chunk of the armor. If the armor is not strong enough to resist being pushed in (armor resisting the blow is covered in the next section), then the force to push in the armor is proportional to the mass of the chunk of armor. If the projectile material is not strong enough to exert this force, then it is deformed and deflects. The following algorithm determines if this type of deflection occurs:
*Calculate the volume of armor moved by the projectile. Multiply armor LAYER_SIZE by projectile CONTACT_AREA, and then round down to the nearest 100. If the result is less than 100, then instead round up to 100.
*Mass=armor SOLID_DENSITY times the volume from the last step
*If the IMPACT_YIELD of the projectile is less than mass*(800/157)/PROJECTILE_SIZE, then the projectile deforms and is reported as “deflected”.
In-game, this type of deflection is observed for wood bolts impacting metal armor, and explains why adamantine is observed to be the worst armor for deflecting wood bolts, as it has the lowest density. As explained below, adamantine can barely stop wood bolts using the "Armor Absorbs the Force of the Collision" mechanism, and wood bolts would easily fracture copper or steel armor were it not for the higher densities of those metals causing the bolt to absorb the force of the collision instead.

====The Armor Absorbs the Force of the Collision====

This form of deflection depends only on the momentum, but not the material, of the projectile. The armor will provide 50% protection from projectiles if
:momentum = (IF-IY/2)*round100(C*S)/(2400000-AU*10000-Q*30000)
where IF=IMPACT_FRACTURE and IY=IMPACT_YIELD of the armor material, C is CONTACT_AREA of the projectile (although it is the smaller of the projectile's contact area and the body part struck's contact area), and S=LAYER_SIZE of the armor item(s) covering the body part struck, AU is the target's armor user skill (0=none, 16=legendary) and Q is the armor quality (0=none, 5=masterwork). The function round100(x) is equal to 100 if x is less than 100, and 100*floor(x/100) otherwise (rounded down to the nearest 100). It unclear if or how the thicknesses of multiple armor layers covering one body part stack. This formula is approximate, and the exact values of the coefficients in the denominator may not be accurate. The amount of protection provided drops sigmoidally from 100% to zero over a fairly narrow range. Furthermore, negative status of the target creature (fallen over, unconscious, etc) are known to significantly reduce armor protection provided. These numbers will roughly predict the results for an ideal, uninjured dwarf, and worse protection should be expected in other cases. The exact manner in which status affects impact armor protection has not yet been explored.

The manner in which armor offers protection is different depending on whether the bolt momentum is greater than 50000/IMPACT_STRAIN_AT_YIELD. If momentum is less than this threshold, then armor can completely deflect projectiles, while if it is larger, armor can significantly slow projectiles so that they cause only bruises (not tears, chips, fractures, or jams).

This type of collision is only observed in-game for bolts with densities less than about 750 (most wood, adamantine) deflecting off adamantine armor. In all other cases the bolt momentum is far to high to be stopped by any armor using this mechanism.

{{Category|Materials|*}}

v0.34 Talk:Pet

2013-05-20T01:28:20Z

Pirate Bob:

The first part of the article says: "...being fed by their owners if they are grazers...", but there is also the statement: "...and the fact that owners will not specifically feed their pets...". Does the owner of a grazing pet feed it, or does it have to be pastured?
: If a pet becomes hungry then a job is definitely generated for their owner to feed them. This job is only assigned to the pet's owner, not to any idle dwarf. I assume that the job is only generated once the pet is hungry (similar to "give water" jobs being generated when dwarves are thirsty) but am not confident enough on the details to edit the article. [[User:Pushy|Pushy]] 12:20, 8 November 2012 (UTC)
:: I have a fortress that has had no pasture for going on a year now. My dwarves have several grazer pets (a couple rabbits and a billy goat) who all seem perfectly fine without being pastured. In the same timeframe, my other grazers have all starved. Clearly, the owners take care of their pets. --[[User:LackofCertainty|LackofCertainty]] 18:39, 22 January 2013 (UTC)
:::My current best guess is that pets follow owners outside where grass is, non-assigned stay in the meeting areas where no gras grows and die. --[[User:Old Ancient|Old Ancient]] 18:28, 2 April 2013 (UTC)

==Fights==
I found in reports that various animals attacked my dwarves (a broken ear, really?) or other animals. It appears that these animals perceive(d) the dining room as crowded, even though no pasture was placed there. Might also be the mix of animals (ducks, rooster, dogs...) Is this known/covered in a different article? It appears no serious injuries happened so far but the thought of my dwarves being attacked by war dogs while dining - not funny. --[[User:Old Ancient|Old Ancient]] 04:12, 28 March 2013 (UTC)
:According to the [[DF2012:Quickstart guide#Temporary Meeting Area|Quickstart guide]], animals will fight if they are in meeting areas as well, which is probably what happened in your case. If they aren't forced to stay in a specific area, they would probably just drift away from each other. --{{User:Lethosor/sig}} 20:25, 28 March 2013 (UTC)
::Thanks. Now, why did I post this to "pets"?? --[[User:Old Ancient|Old Ancient]] 19:40, 30 March 2013 (UTC)

==Feeding==
"Pasture pets" ''do'' get hungry. Contact with the owner does not per se feed them. They don't seek out a pasture on their own. They might be feeding on grass whenever they follow their owner outside. No "feed" job observed so far. Cats appear to cause 'give water' jobs occasionally that are executed by a random dwarf, not the owner. Research ongoing. --[[User:Old Ancient|Old Ancient]] 19:40, 30 March 2013 (UTC)

==Chains/Cages==
I removed the mention of assigning migrant pets to chains, as I believe it is not possible to assign pets to chains in DF2012. If this is incorrect, this should of course be reverted, but it also might be good to add a discussion of how to assign pets to chains, as this is not obvious. Only stray animals show up when I try to assign an animal to a chain. Also, it might not hurt to mention somewhere that pets cannot be caged.

User talk:Pirate Bob

2013-05-20T01:25:04Z

Pirate Bob: Blanked the page

User talk:Pirate Bob

2013-05-20T01:24:15Z

Pirate Bob: Created page with "I may be missing something, but I am pretty sure in DF2012 you cannot assign pets to chains (nothing but stray animals is showing up for me when I try to assign an animal to a..."

I may be missing something, but I am pretty sure in DF2012 you cannot assign pets to chains (nothing but stray animals is showing up for me when I try to assign an animal to a chain). If this is incorrect, maybe there should be some discussion of how to assign pets to chains, as it is not obvious? It also might not hurt to mention that pets cannot be assigned to cages (assuming this is true...)

v0.34:Pet

2013-05-20T01:22:01Z

Pirate Bob: /* Migrant Pets */

{{quality|Exceptional|03:25, 16 June 2012 (UTC)}}{{av}}

'''Pets''' are [[creatures]] that have been adopted by your [[dwarves]]. All creatures that can be [[animal training|tamed]] can be made into pets; that is, most creatures that are both [[:Category:Humanoids|non-sapient]] and [[Evil|non-evil]]. Pets will follow their owners around, providing a companionship [[happiness]] bonus (based on the creature's pet value), being fed by their owners if they are [[grazer]]s, and, in many cases, protecting them from harm, either by actively attacking enemies or by acting as meat shields that allow their charges to scramble away to safety.

== Adopting ==
Pets cannot be assigned; rather, dwarves will adopt them on an ''ad hoc'' basis. Dwarves that are idling near a wandering [[stray]] animal may chose to make them their pet, giving them a name, causing the animal to follow them around, and giving the owner a happy "got a pet" [[thought]]. The chances of this event happening are strongly influenced by your dwarves' personal preferences; while they will rarely go out of their way to adopt animals, every dwarf has a preference for a certain creature, and will select them as pets when available, often resulting in the fortress engraver being followed around by a family of [[cavy|cavies]].

You can control what animals are up for adoption through the animal [[status]] screen, the first tab on the {{k|z}} menu. There you can toggle available/unavailable on each tame individual, freeing them up for adoption. Animals will never be adopted out of a [[cage]], and must be let out to roam for adoption to take place. [[Stray]] animals tend to accumulate at [[meeting area]]s, which facilitates adoption. Be aware that while this is unproblematic for animals that don't need to be fed (e.g. carnivores), [[Pasture#Grazing animals and pasture size|grazers]] however tend to starve to death in your [[dining room]] if you do not assign them to a [[pasture]] instead.

In the case of grazers, owners take care of feeding their pets.{{verify}} details are fuzzy?

[[Cat]]s are notable for being an exception to the rule; they will assign ''themselves'' to dwarves, and cannot be made available or unavailable. Vermin pets will perch on their masters' shoulders, protecting them from those darned cats.

== Utility ==
Pets will provide a happiness bonus to their owner for the duration of their or their owner's lives. If the dwarf dies, the animal becomes a stray, and is available for adoption again. If the pet dies from anything besides old age, the owner will get a negative thought, made worse if the pet isn't given a proper [[burial]]. As long as a creature is somebody's pet, they cannot be butchered; this can lead to a [[catsplosion]] when there are several crazy old cat ladies in your fortress and you don't immediately butcher all the kittens.

Another possible use is for large grazers, which are unable to keep themselves fed via grazing. By turning them into a pet, you can essentially have a dwarf dedicated to keeping them fed. This could open up some interesting options. (such as utilizing tame [[elephant]]s without worrying about them starving themselves)

== Good pets ==
The value of a pet is based upon the species' internal pet value, a number in turn based on a number of factors: body size, exotic-ness, and intelligence being among them. It also helps if the animals in question aren't useful for anything else, although assigning pets to pastures to keep them in place works fine, allowing such thing as pet [[alligator]] [[egg production|egg farms]]. Milkable and shearable pets are excellent, since they can still be taken to a [[farmer's workshop]] to get their goods. The best pets of all are [[cave dragon]]s and [[dragon]]s, if the world is old enough and/or you are lucky enough to cage some.

== Migrant Pets ==
[[Migrant]] waves will bring stray animals, but may also bring pets along. While this can be a boon, some players may find it a hindrance; animals with no value beyond [[butchery]], such as pigs or cavies, may be viewed as clutter. If so, you can assign a pasture to put them somewhere conveniently dangerous, such as under a [[bridge|atom smasher]] or in a field to trip up [[ambush|ambushers]]. Be warned that dead pets generate negative [[thought]]s.

v0.34 Talk:Clothing

2013-01-10T16:05:13Z

Pirate Bob: /* Bad thoughts from lack of headwear */

==Armor and bad thoughts==
I'm still using 34.07 so I can't confirm, but didn't Toady state for 34.11 that he "made armor stop lack of clothing thoughts" at http://www.bay12games.com/dwarves/index.html?
The top of this page still says only "Clothing' stops nudity thoughts. [[User:Xen0n|Xen0n]] 17:27, 23 June 2012 (UTC)
:It says that clothing prevents bad thoughts from nudity, it doesn't say that '''only''' clothing does that nor does it say that armor does '''not''' do that. I don't see anything that needs correcting there. -- [[User:HiEv|Hi]][[User talk:HiEv|Ev]] 00:09, 25 June 2012 (UTC)
:I can confirm that armor does prevent bad thoughts from nudity. Or, at least wearing armor plus a robe does. I will have to try getting rid of the robes and seeing what happens. I agree that if this is the case the page should be updated, as I also believed from reading the current page that at least some regular clothing was needed.--[[User:Pirate Bob|Pirate Bob]] 15:35, 8 January 2013 (UTC)

==Bad thoughts from lack of headwear==
I believe that if dwarves are not wearing a cap or helm they get the "was worried not to have adequate protection lately" thought. If this is true, it should definitely be listed here, and also probably added to the main [[Thought]] page. This thought is listed on [[DF2012:List_of_Dwarven_Thoughts]], but I can't find anywhere that explains how it works. I will try to confirm the cause of this thought, and then add it (unless someone else can confirm it first). --[[User:Pirate Bob|Pirate Bob]] 15:35, 8 January 2013 (UTC)
:It seems that headwear is not the cause of this thought. From [http://www.bay12forums.com/smf/index.php?topic=118497.msg3751594#msg3751594 Quietust] "The "lack of protection" thought is normally only received by the Tax Collector when you don't have enough royal guards, but the DFHack 'siren' script uses it as a punishment for forcefully waking up your units...". I believe I used the "siren" when I got annoyed that my stupid broker was refusing to trade again, so that explains that. I guess headwear is not required at all?--[[User:Pirate Bob|Pirate Bob]] 16:05, 10 January 2013 (UTC)

v0.34 Talk:Clothing

2013-01-08T15:35:30Z

Pirate Bob: /* Armor and bad thoughts */

v0.34:Material science

2012-12-02T22:16:00Z

Pirate Bob: /* The Armor Absorbs the Force of the Collision */

{{quality|Fine|02:27, 11 May 2012 (UTC)}}{{av}}
[[Material|Materials]] have a number of properties representing real world variables that describe how they respond to inputs. In particular, the game now has a number of variables that describe what happens to a material when it's put under stress.

==What is stress?==
In the real world, an object is stressed when a force is applied to the object. Depending on the nature of the force applied, this stress can take a number of forms, and the object can respond differently based on its material and how that material handles different stresses.

In the material raws, whenever you see 'yield', 'fracture', or '[[Strain at yield|strain at yield]]', that property is a stress-related quality.

==When does Dwarf Fortress make stress calculations?==
At present, DF seems to only apply forces during combat, and thus only stresses objects (generally armor and various body layers) at that time.

==There's a lot of stress-related properties, what do they mean?==
The first thing you'll notice is that the second word in each stress variable is one of Yield, Fracture, or strain at yield. These are mechanical performance terms.

The first set of words are things like Impact, Bending, and so forth. These describe modes of applying force.

The following explanations assumes real world physics sort of apply (since [[Main:Toady One|Toady One]] chose real world properties). The game doesn't use all of these properties yet, and may not be applying them according to real world physics.

===Mechanical Performance Properties===
Yield: This is almost certainly 'Yield Strength', which is the amount of stress needed to cause a material to go from elastic deformation to plastic deformation. (That is, if you cease stressing the object, does it revert to its original shape or not). Since most objects only elastically deform over small distances of deformation, high Yield values generally means it takes a lot of force to noticeably 'stretch' them (but see strain at yield).

Fracture: The fracture point is the amount of stress or force necessarily to cause the material to fail, or in other words, to break.

Strain at yield (sometimes incorrectly referred to as 'elasticity'): This variable tells you how much deformation occurs to the material while it is deforming elastically. That is, as long as the force is less than the yield strength, stress * strain at yield = deformation distance. The smaller the strain at yield, the less deformation occurs under stress.

Note: Strain at yield is the inverse of the Elastic Modulus. Thus a highly elastic material has low elastic modulus, and engages in less elastic collisions.

===Modes of Applying Force===
Impact: Force applied by a sudden strike, like a hammer.

Compressive: Force applied by exerting pressure on an object, like trying to squish something between your hands.

Tensile: Force applied by pulling on something, like suspending one object via another. (e.g., if you suspend an [[elf]] from a metal pole, you are applying a tensile force to the pole).

Torsion: Force applied by twisting something. Note that you're twisting some portion of the object relative to itself to cause a torsion stress to be applied to it. (Consider trying to twist a metal rod by grasping at either end and attempting to wring it - yes, you'd have to apply a lot of force to succeed).

Shear: Force applied by pushing part of the material so it tries to slide relative to another part of it. Ie, pushing at the top of an object when the bottom part is fixed to the ground is going to primarily apply a shear stress to it (the top part will try to move in the direction you push, and the lower part will resist this shear stress).

Bending: Force applied by bending a material.

==Effects on Combat==
The Dwarf Fortress combat system does not use all material properties at present (0.34.11). Weapon and armor damage/wear/decay is not tracked.

The mechanics governing how material properties impact the protection armor provides against projectiles have been [http://www.bay12forums.com/smf/index.php?topic=116151.0 studied reasonably carefully], and are described below. Armor protection from melee weapons has not been studied in detail, although it likely shares many similarities with ranged weapons. '''''Note: all of the below is based on recent (0.34.11) experiments, and many results may be approximate or incomplete.'''''

===Projectile Properties===
The main factor determining if a projectile penetrates armor is its momentum. Material properties of the bolt/arrow (except for IMPACT_YIELD, which will be described below), as well as bolt quality and marksdwarf/archer skills do not appear to matter (the skill of the marksdwarf determines if the bolt misses, hits, is dodged, or is blocked with a shield). Dwarven weapons launch projectiles with roughly constant momentum. However rounding of projectile masses can result in some very strange behavior. The complete calculation used to determine projectile momentum is as follows:
*Find the mass of your projectile. Mass=SOLID_DENSITY*SIZE/1,000,000.
*Floor that number to an integer.
*Divide the SHOOT_FORCE of the weapon by the integer, and round the result to the nearest integer.
*Cap the results by the SHOOT_MAXVEL for the weapon.
*Multiply the velocity by the ''unrounded'' mass of the projectile to get the momentum.
For DF34.11 values, iron, bronze, copper, silver, and steel bolts/arrows all have mass between 1 and 2, which will be floored to 1, meaning that they will have velocity equal to SHOOT_FORCE (so long as it is less than or equal to SHOOT_MAXVEL). Wood, adamantine, and bone bolts/arrows, as well as all blowdarts, have mass less than 1, and will have velocity which is always equal to SHOOT_MAXVEL and does not depend on SHOOT_FORCE.

===Interactions Between Projectiles and Armor===
[[File:projectile.png|200px|thumb|right|A simplified flowchart showing how material properties are used as a projectile contacts armor.]]
When a projectile strikes armor, there are several possible outcomes:
*Conversion of edged damage to blunt damage (chain mail does this).
*The projectile absorbs the force of the collision, and is deflected.
*The armor absorbs all or part of the force of the blow.
*The armor is fractured and does nothing to stop the projectile.
For armor to be at all effective at stopping projectiles, the armor material must have SHEAR_YIELD and/or SHEAR_FRACTURE greater than or equal to the projectile material. Otherwise, the projectile just cuts through the armor like it isn’t there. Additionally, the IMPACT_FRACTURE of the armor must be large relative to the projectile momentum, or the armor fractures and the projectile passes through without slowing down significantly (this happens for metal bolts against any plate armor in 0.34.11).

====Conversion of Edged Damage to Blunt Damage====
If the momentum of the projectile is not too high, then chain armor can convert the edged damage normally caused by projectiles to blunt damage, resulting in chips, fractures, jams and bruises, but no tears or cuts. Plate armor does not appear to provide this type of protection under any conditions tested so far. It is not yet known how the momentum needed for edged damage to penetrate chain armor is calculated, but for the cases examined the momentum needed is many times larger than those observed for in-game projectiles.

====The Projectile Absorbs the Force of the Collision====
If the projectile does not cut right through the armor (armor material must have SHEAR_YIELD and/or SHEAR_FRACTURE greater than or equal to the projectile material), then it must push in a chunk of the armor. If the armor is not strong enough to resist being pushed in (armor resisting the blow is covered in the next section), then the force to push in the armor is proportional to the mass of the chunk of armor. If the projectile material is not strong enough to exert this force, then it is deformed and deflects. The following algorithm determines if this type of deflection occurs:
*Calculate the volume of armor moved by the projectile. Multiply armor LAYER_SIZE by projectile CONTACT_AREA, and then round down to the nearest 100. If the result is less than 100, then instead round up to 100.
*Mass=armor SOLID_DENSITY times the volume from the last step
*If the IMPACT_YIELD of the projectile is less than mass*(800/157)/PROJECTILE_SIZE, then the projectile deforms and is reported as “deflected”.
In-game, this type of deflection is observed for wood bolts impacting metal armor, and explains why adamantine is observed to be the worst armor for deflecting wood bolts, as it has the lowest density. As explained below, adamantine can barely stop wood bolts using the "Armor Absorbs the Force of the Collision" mechanism, and wood bolts would easily fracture copper or steel armor were it not for the higher densities of those metals causing the bolt to absorb the force of the collision instead.

====The Armor Absorbs the Force of the Collision====

This form of deflection depends only on the momentum, but not the material, of the projectile. The armor will provide 50% protection from projectiles if
:momentum = (IF-IY/2)*round100(C*S)/(2400000-AU*10000-Q*30000)
where IF=IMPACT_FRACTURE and IY=IMPACT_YIELD of the armor material, C is CONTACT_AREA of the projectile (although it is the smaller of the projectile's contact area and the body part struck's contact area), and S=LAYER_SIZE of the armor item(s) covering the body part struck, AU is the target's armor user skill (0=none, 16=legendary) and Q is the armor quality (0=none, 5=masterwork). The function round100(x) is equal to 100 if x is less than 100, and 100*floor(x/100) otherwise (rounded down to the nearest 100). It unclear if or how the thicknesses of multiple armor layers covering one body part stack. This formula is approximate, and the exact values of the coefficients in the denominator may not be accurate. The amount of protection provided drops sigmoidally from 100% to zero over a fairly narrow range. Furthermore, negative status of the target creature (fallen over, unconscious, etc) are known to significantly reduce armor protection provided. These numbers will roughly predict the results for an ideal, uninjured dwarf, and worse protection should be expected in other cases. The exact manner in which status affects impact armor protection has not yet been explored.

The manner in which armor offers protection is different depending on whether the bolt momentum is greater than 50000/IMPACT_STRAIN_AT_YIELD. If momentum is less than this threshold, then armor can completely deflect projectiles, while if it is larger, armor can significantly slow projectiles so that they cause only bruises (not tears, chips, fractures, or jams).

This type of collision is only observed in-game for bolts with densities less than about 750 (most wood, adamantine) deflecting off adamantine armor, or adamantine bolts off most metal armors. In all other cases the bolt momentum is far to high to be stopped by any armor using this mechanism.

{{Category|Materials|*}}

v0.34:Material science

2012-12-02T22:08:39Z

Pirate Bob: /* The Projectile Absorbs the Force of the Collision */

{{quality|Fine|02:27, 11 May 2012 (UTC)}}{{av}}
[[Material|Materials]] have a number of properties representing real world variables that describe how they respond to inputs. In particular, the game now has a number of variables that describe what happens to a material when it's put under stress.

==What is stress?==
In the real world, an object is stressed when a force is applied to the object. Depending on the nature of the force applied, this stress can take a number of forms, and the object can respond differently based on its material and how that material handles different stresses.

In the material raws, whenever you see 'yield', 'fracture', or '[[Strain at yield|strain at yield]]', that property is a stress-related quality.

==When does Dwarf Fortress make stress calculations?==
At present, DF seems to only apply forces during combat, and thus only stresses objects (generally armor and various body layers) at that time.

==There's a lot of stress-related properties, what do they mean?==
The first thing you'll notice is that the second word in each stress variable is one of Yield, Fracture, or strain at yield. These are mechanical performance terms.

The first set of words are things like Impact, Bending, and so forth. These describe modes of applying force.

The following explanations assumes real world physics sort of apply (since [[Main:Toady One|Toady One]] chose real world properties). The game doesn't use all of these properties yet, and may not be applying them according to real world physics.

===Mechanical Performance Properties===
Yield: This is almost certainly 'Yield Strength', which is the amount of stress needed to cause a material to go from elastic deformation to plastic deformation. (That is, if you cease stressing the object, does it revert to its original shape or not). Since most objects only elastically deform over small distances of deformation, high Yield values generally means it takes a lot of force to noticeably 'stretch' them (but see strain at yield).

Fracture: The fracture point is the amount of stress or force necessarily to cause the material to fail, or in other words, to break.

Strain at yield (sometimes incorrectly referred to as 'elasticity'): This variable tells you how much deformation occurs to the material while it is deforming elastically. That is, as long as the force is less than the yield strength, stress * strain at yield = deformation distance. The smaller the strain at yield, the less deformation occurs under stress.

Note: Strain at yield is the inverse of the Elastic Modulus. Thus a highly elastic material has low elastic modulus, and engages in less elastic collisions.

===Modes of Applying Force===
Impact: Force applied by a sudden strike, like a hammer.

Compressive: Force applied by exerting pressure on an object, like trying to squish something between your hands.

Tensile: Force applied by pulling on something, like suspending one object via another. (e.g., if you suspend an [[elf]] from a metal pole, you are applying a tensile force to the pole).

Torsion: Force applied by twisting something. Note that you're twisting some portion of the object relative to itself to cause a torsion stress to be applied to it. (Consider trying to twist a metal rod by grasping at either end and attempting to wring it - yes, you'd have to apply a lot of force to succeed).

Shear: Force applied by pushing part of the material so it tries to slide relative to another part of it. Ie, pushing at the top of an object when the bottom part is fixed to the ground is going to primarily apply a shear stress to it (the top part will try to move in the direction you push, and the lower part will resist this shear stress).

Bending: Force applied by bending a material.

==Effects on Combat==
The Dwarf Fortress combat system does not use all material properties at present (0.34.11). Weapon and armor damage/wear/decay is not tracked.

The mechanics governing how material properties impact the protection armor provides against projectiles have been [http://www.bay12forums.com/smf/index.php?topic=116151.0 studied reasonably carefully], and are described below. Armor protection from melee weapons has not been studied in detail, although it likely shares many similarities with ranged weapons. '''''Note: all of the below is based on recent (0.34.11) experiments, and many results may be approximate or incomplete.'''''

===Projectile Properties===
The main factor determining if a projectile penetrates armor is its momentum. Material properties of the bolt/arrow (except for IMPACT_YIELD, which will be described below), as well as bolt quality and marksdwarf/archer skills do not appear to matter (the skill of the marksdwarf determines if the bolt misses, hits, is dodged, or is blocked with a shield). Dwarven weapons launch projectiles with roughly constant momentum. However rounding of projectile masses can result in some very strange behavior. The complete calculation used to determine projectile momentum is as follows:
*Find the mass of your projectile. Mass=SOLID_DENSITY*SIZE/1,000,000.
*Floor that number to an integer.
*Divide the SHOOT_FORCE of the weapon by the integer, and round the result to the nearest integer.
*Cap the results by the SHOOT_MAXVEL for the weapon.
*Multiply the velocity by the ''unrounded'' mass of the projectile to get the momentum.
For DF34.11 values, iron, bronze, copper, silver, and steel bolts/arrows all have mass between 1 and 2, which will be floored to 1, meaning that they will have velocity equal to SHOOT_FORCE (so long as it is less than or equal to SHOOT_MAXVEL). Wood, adamantine, and bone bolts/arrows, as well as all blowdarts, have mass less than 1, and will have velocity which is always equal to SHOOT_MAXVEL and does not depend on SHOOT_FORCE.

===Interactions Between Projectiles and Armor===
[[File:projectile.png|200px|thumb|right|A simplified flowchart showing how material properties are used as a projectile contacts armor.]]
When a projectile strikes armor, there are several possible outcomes:
*Conversion of edged damage to blunt damage (chain mail does this).
*The projectile absorbs the force of the collision, and is deflected.
*The armor absorbs all or part of the force of the blow.
*The armor is fractured and does nothing to stop the projectile.
For armor to be at all effective at stopping projectiles, the armor material must have SHEAR_YIELD and/or SHEAR_FRACTURE greater than or equal to the projectile material. Otherwise, the projectile just cuts through the armor like it isn’t there. Additionally, the IMPACT_FRACTURE of the armor must be large relative to the projectile momentum, or the armor fractures and the projectile passes through without slowing down significantly (this happens for metal bolts against any plate armor in 0.34.11).

====Conversion of Edged Damage to Blunt Damage====
If the momentum of the projectile is not too high, then chain armor can convert the edged damage normally caused by projectiles to blunt damage, resulting in chips, fractures, jams and bruises, but no tears or cuts. Plate armor does not appear to provide this type of protection under any conditions tested so far. It is not yet known how the momentum needed for edged damage to penetrate chain armor is calculated, but for the cases examined the momentum needed is many times larger than those observed for in-game projectiles.

====The Projectile Absorbs the Force of the Collision====
If the projectile does not cut right through the armor (armor material must have SHEAR_YIELD and/or SHEAR_FRACTURE greater than or equal to the projectile material), then it must push in a chunk of the armor. If the armor is not strong enough to resist being pushed in (armor resisting the blow is covered in the next section), then the force to push in the armor is proportional to the mass of the chunk of armor. If the projectile material is not strong enough to exert this force, then it is deformed and deflects. The following algorithm determines if this type of deflection occurs:
*Calculate the volume of armor moved by the projectile. Multiply armor LAYER_SIZE by projectile CONTACT_AREA, and then round down to the nearest 100. If the result is less than 100, then instead round up to 100.
*Mass=armor SOLID_DENSITY times the volume from the last step
*If the IMPACT_YIELD of the projectile is less than mass*(800/157)/PROJECTILE_SIZE, then the projectile deforms and is reported as “deflected”.
In-game, this type of deflection is observed for wood bolts impacting metal armor, and explains why adamantine is observed to be the worst armor for deflecting wood bolts, as it has the lowest density. As explained below, adamantine can barely stop wood bolts using the "Armor Absorbs the Force of the Collision" mechanism, and wood bolts would easily fracture copper or steel armor were it not for the higher densities of those metals causing the bolt to absorb the force of the collision instead.

====The Armor Absorbs the Force of the Collision====

This form of deflection depends only on the momentum, but not the material, of the projectile. The armor will provide 50% protection from projectiles if
:momentum = (IF-IY/2)*round100(C*S)/(2400000-AU*10000-Q*30000)
where IF=IMPACT_FRACTURE and IY=IMPACT_YIELD of the armor material, C is CONTACT_AREA of the projectile (although it is the smaller of the projectile's contact area and the body part struck's contact area), and S=LAYER_SIZE of the armor item(s) covering the body part struck, AU is the target's armor user skill (0=none, 16=legendary) and Q is the armor quality (0=none, 5=masterwork). The function round100(x) is equal to 100 if x is less than 100, and 100*floor(x/100) otherwise (rounded down to the nearest 100). It unclear if or how the thicknesses of multiple armor layers covering one body part stack. This formula is approximate, and the exact values of the coefficients in the denominator may not be accurate. The amount of protection provided drops sigmoidally from 100% to zero over a fairly narrow range.

The manner in which armor offers protection is different depending on whether the bolt momentum is greater than 50000/IMPACT_STRAIN_AT_YIELD. If momentum is less than this threshold, then armor can completely deflect projectiles, while if it is larger, armor can significantly slow projectiles so that they cause only bruises (not tears, chips, fractures, or jams).

{{Category|Materials|*}}

v0.34 Talk:Material science

2012-12-02T22:05:57Z

Pirate Bob: /* Updated page */

I don't understand what the first paragraph in the "Effects on Combat" section mean, so I am leaving them alone for now. If no one else can make sense of them they should probably be removed. I am starting to add the results of our ballistics testing to the combat section. --[[User:Pirate Bob|Pirate Bob]] 22:19, 29 November 2012 (UTC)

It would be very nice to add Urist's flowchart to the "Interactions between projectiles and armor" section. It might also be nice if the flowchart could somehow be highlighted to emphasize which sections of the text correspond to which parts of the flowchart, but I don't know if this is feasible. --[[User:Pirate Bob|Pirate Bob]] 22:49, 29 November 2012 (UTC)

The ratio involved in "The projectile absorbs the force of the collision" seems kind of odd. We should investigate if there are actually more parameters determining this at some point.--[[User:Pirate Bob|Pirate Bob]] 23:11, 29 November 2012 (UTC)
:Sorry about that, the fraction 800/157 is equal to 40000/7850. It is this way because a size 100 ammo will deflect off iron armor if the ammo has an IMPACT_YIELD less than 40000. We did investigate the other parameters and they don't apply to that equation. --[[User:UristDaVinci|UristDaVinci]] 06:11, 2 December 2012 (UTC)

Perhaps we should put links to Toady One posts on the forums in a section on this page, as "Word of Toady" on how the system works. --[[User:UristDaVinci|UristDaVinci]] 06:11, 2 December 2012 (UTC)

== Updated page ==

Good work updating the page, but I'm a little confused by "As explained below, wood bolts are capable of fracturing through the strongest metal armor, and would fracture copper or steel armor were it not for the higher densities of those metals." [[User:Emufarmers|Emufarmers]] 10:24, 2 December 2012 (UTC)
:Look at the flowchart. If copper, steel, bronze, or iron were not so dense (or wood so fragile), wood bolts would be able to fracture a hole in the armor and deal damage. Adamantine is stronger than steel, but provides less protection against wood bolts only because of the low density of adamantine. Can you think of a better way to explain that? --[[User:UristDaVinci|UristDaVinci]] 20:45, 2 December 2012 (UTC)
:I was just looking at the numbers, and it appears according to our formula that adamantine actually should be able to deflect most wood (density 600) with 34.11 parameter values. Adamantine should provide protection against bolts up to momentum ~110, and a density 600 bolt (mass 0.090 kg) will travel at SHOOT_MAXVEL=1000, and have momentum 90, and should be deflected. I used pine (density 510) in my original tests where I found only 58% deflection of wood bolts off adamantine armor, so I have no idea what's going on there. I guess I'll have to check deflection of wood off adamantine again...--[[User:Pirate Bob|Pirate Bob]] 22:05, 2 December 2012 (UTC)
:Update - I confirmed that addy can block tower cap bolts >90% of the time. I suspect what was going on before when I found 56% deflection is I was using normal dwarves, and the probability of armor penetration goes up dramatically if the dwarf is passed out (which happens rapidly due to hits to unarmored areas causing chips). Anyway, the correct thing to say is that adamantine barely stop wood bolts under ideal conditions, and no other armors would stop them using the "armor absorbs the collosion" mechanism.--[[User:Pirate Bob|Pirate Bob]] 22:05, 2 December 2012 (UTC)

v0.34 Talk:Material science

2012-12-02T22:05:43Z

Pirate Bob: /* Updated page */

v0.34 Talk:Material science

2012-12-02T21:36:34Z

Pirate Bob: /* Updated page */

v0.34:Material science

2012-12-02T21:26:25Z

Pirate Bob: Undo revision 179092 by Pirate Bob (talk)

{{quality|Fine|02:27, 11 May 2012 (UTC)}}{{av}}
[[Material|Materials]] have a number of properties representing real world variables that describe how they respond to inputs. In particular, the game now has a number of variables that describe what happens to a material when it's put under stress.

==What is stress?==
In the real world, an object is stressed when a force is applied to the object. Depending on the nature of the force applied, this stress can take a number of forms, and the object can respond differently based on its material and how that material handles different stresses.

In the material raws, whenever you see 'yield', 'fracture', or '[[Strain at yield|strain at yield]]', that property is a stress-related quality.

==When does Dwarf Fortress make stress calculations?==
At present, DF seems to only apply forces during combat, and thus only stresses objects (generally armor and various body layers) at that time.

==There's a lot of stress-related properties, what do they mean?==
The first thing you'll notice is that the second word in each stress variable is one of Yield, Fracture, or strain at yield. These are mechanical performance terms.

The first set of words are things like Impact, Bending, and so forth. These describe modes of applying force.

The following explanations assumes real world physics sort of apply (since [[Main:Toady One|Toady One]] chose real world properties). The game doesn't use all of these properties yet, and may not be applying them according to real world physics.

===Mechanical Performance Properties===
Yield: This is almost certainly 'Yield Strength', which is the amount of stress needed to cause a material to go from elastic deformation to plastic deformation. (That is, if you cease stressing the object, does it revert to its original shape or not). Since most objects only elastically deform over small distances of deformation, high Yield values generally means it takes a lot of force to noticeably 'stretch' them (but see strain at yield).

Fracture: The fracture point is the amount of stress or force necessarily to cause the material to fail, or in other words, to break.

Strain at yield (sometimes incorrectly referred to as 'elasticity'): This variable tells you how much deformation occurs to the material while it is deforming elastically. That is, as long as the force is less than the yield strength, stress * strain at yield = deformation distance. The smaller the strain at yield, the less deformation occurs under stress.

Note: Strain at yield is the inverse of the Elastic Modulus. Thus a highly elastic material has low elastic modulus, and engages in less elastic collisions.

===Modes of Applying Force===
Impact: Force applied by a sudden strike, like a hammer.

Compressive: Force applied by exerting pressure on an object, like trying to squish something between your hands.

Tensile: Force applied by pulling on something, like suspending one object via another. (e.g., if you suspend an [[elf]] from a metal pole, you are applying a tensile force to the pole).

Torsion: Force applied by twisting something. Note that you're twisting some portion of the object relative to itself to cause a torsion stress to be applied to it. (Consider trying to twist a metal rod by grasping at either end and attempting to wring it - yes, you'd have to apply a lot of force to succeed).

Shear: Force applied by pushing part of the material so it tries to slide relative to another part of it. Ie, pushing at the top of an object when the bottom part is fixed to the ground is going to primarily apply a shear stress to it (the top part will try to move in the direction you push, and the lower part will resist this shear stress).

Bending: Force applied by bending a material.

==Effects on Combat==
The Dwarf Fortress combat system does not use all material properties at present (0.34.11). Weapon and armor damage/wear/decay is not tracked.

The mechanics governing how material properties impact the protection armor provides against projectiles have been [http://www.bay12forums.com/smf/index.php?topic=116151.0 studied reasonably carefully], and are described below. Armor protection from melee weapons has not been studied in detail, although it likely shares many similarities with ranged weapons. '''''Note: all of the below is based on recent (0.34.11) experiments, and many results may be approximate or incomplete.'''''

===Projectile Properties===
The main factor determining if a projectile penetrates armor is its momentum. Material properties of the bolt/arrow (except for IMPACT_YIELD, which will be described below), as well as bolt quality and marksdwarf/archer skills do not appear to matter (the skill of the marksdwarf determines if the bolt misses, hits, is dodged, or is blocked with a shield). Dwarven weapons launch projectiles with roughly constant momentum. However rounding of projectile masses can result in some very strange behavior. The complete calculation used to determine projectile momentum is as follows:
*Find the mass of your projectile. Mass=SOLID_DENSITY*SIZE/1,000,000.
*Floor that number to an integer.
*Divide the SHOOT_FORCE of the weapon by the integer, and round the result to the nearest integer.
*Cap the results by the SHOOT_MAXVEL for the weapon.
*Multiply the velocity by the ''unrounded'' mass of the projectile to get the momentum.
For DF34.11 values, iron, bronze, copper, silver, and steel bolts/arrows all have mass between 1 and 2, which will be floored to 1, meaning that they will have velocity equal to SHOOT_FORCE (so long as it is less than or equal to SHOOT_MAXVEL). Wood, adamantine, and bone bolts/arrows, as well as all blowdarts, have mass less than 1, and will have velocity which is always equal to SHOOT_MAXVEL and does not depend on SHOOT_FORCE.

===Interactions Between Projectiles and Armor===
[[File:projectile.png|200px|thumb|right|A simplified flowchart showing how material properties are used as a projectile contacts armor.]]
When a projectile strikes armor, there are several possible outcomes:
*Conversion of edged damage to blunt damage (chain mail does this).
*The projectile absorbs the force of the collision, and is deflected.
*The armor absorbs all or part of the force of the blow.
*The armor is fractured and does nothing to stop the projectile.
For armor to be at all effective at stopping projectiles, the armor material must have SHEAR_YIELD and/or SHEAR_FRACTURE greater than or equal to the projectile material. Otherwise, the projectile just cuts through the armor like it isn’t there. Additionally, the IMPACT_FRACTURE of the armor must be large relative to the projectile momentum, or the armor fractures and the projectile passes through without slowing down significantly (this happens for metal bolts against any plate armor in 0.34.11).

====Conversion of Edged Damage to Blunt Damage====
If the momentum of the projectile is not too high, then chain armor can convert the edged damage normally caused by projectiles to blunt damage, resulting in chips, fractures, jams and bruises, but no tears or cuts. Plate armor does not appear to provide this type of protection under any conditions tested so far. It is not yet known how the momentum needed for edged damage to penetrate chain armor is calculated, but for the cases examined the momentum needed is many times larger than those observed for in-game projectiles.

====The Projectile Absorbs the Force of the Collision====
If the projectile does not cut right through the armor (armor material must have SHEAR_YIELD and/or SHEAR_FRACTURE greater than or equal to the projectile material), then it must push in a chunk of the armor. If the armor is not strong enough to resist being pushed in (armor resisting the blow is covered in the next section), then the force to push in the armor is proportional to the mass of the chunk of armor. If the projectile material is not strong enough to exert this force, then it is deformed and deflects. The following algorithm determines if this type of deflection occurs:
*Calculate the volume of armor moved by the projectile. Multiply armor LAYER_SIZE by projectile CONTACT_AREA, and then round down to the nearest 100. If the result is less than 100, then instead round up to 100.
*Mass=armor SOLID_DENSITY times the volume from the last step
*If the IMPACT_YIELD of the projectile is less than mass*(800/157)/PROJECTILE_SIZE, then the projectile deforms and is reported as “deflected”.
In-game, this type of deflection is observed for wood bolts impacting metal armor, and explains why adamantine is observed to be the worst armor for deflecting wood bolts, as it has the lowest density. As explained below, wood bolts are capable of fracturing through the strongest metal armor, and would fracture copper or steel armor were it not for the higher densities of those metals.

====The Armor Absorbs the Force of the Collision====

This form of deflection depends only on the momentum, but not the material, of the projectile. The armor will provide 50% protection from projectiles if
:momentum = (IF-IY/2)*round100(C*S)/(2400000-AU*10000-Q*30000)
where IF=IMPACT_FRACTURE and IY=IMPACT_YIELD of the armor material, C is CONTACT_AREA of the projectile (although it is the smaller of the projectile's contact area and the body part struck's contact area), and S=LAYER_SIZE of the armor item(s) covering the body part struck, AU is the target's armor user skill (0=none, 16=legendary) and Q is the armor quality (0=none, 5=masterwork). The function round100(x) is equal to 100 if x is less than 100, and 100*floor(x/100) otherwise (rounded down to the nearest 100). It unclear if or how the thicknesses of multiple armor layers covering one body part stack. This formula is approximate, and the exact values of the coefficients in the denominator may not be accurate. The amount of protection provided drops sigmoidally from 100% to zero over a fairly narrow range.

The manner in which armor offers protection is different depending on whether the bolt momentum is greater than 50000/IMPACT_STRAIN_AT_YIELD. If momentum is less than this threshold, then armor can completely deflect projectiles, while if it is larger, armor can significantly slow projectiles so that they cause only bruises (not tears, chips, fractures, or jams).

{{Category|Materials|*}}

v0.34:Material science

2012-12-02T21:16:41Z

Pirate Bob: /* The Armor Absorbs the Force of the Collision */

{{quality|Fine|02:27, 11 May 2012 (UTC)}}{{av}}
[[Material|Materials]] have a number of properties representing real world variables that describe how they respond to inputs. In particular, the game now has a number of variables that describe what happens to a material when it's put under stress.

==What is stress?==
In the real world, an object is stressed when a force is applied to the object. Depending on the nature of the force applied, this stress can take a number of forms, and the object can respond differently based on its material and how that material handles different stresses.

In the material raws, whenever you see 'yield', 'fracture', or '[[Strain at yield|strain at yield]]', that property is a stress-related quality.

==When does Dwarf Fortress make stress calculations?==
At present, DF seems to only apply forces during combat, and thus only stresses objects (generally armor and various body layers) at that time.

==There's a lot of stress-related properties, what do they mean?==
The first thing you'll notice is that the second word in each stress variable is one of Yield, Fracture, or strain at yield. These are mechanical performance terms.

The first set of words are things like Impact, Bending, and so forth. These describe modes of applying force.

The following explanations assumes real world physics sort of apply (since [[Main:Toady One|Toady One]] chose real world properties). The game doesn't use all of these properties yet, and may not be applying them according to real world physics.

===Mechanical Performance Properties===
Yield: This is almost certainly 'Yield Strength', which is the amount of stress needed to cause a material to go from elastic deformation to plastic deformation. (That is, if you cease stressing the object, does it revert to its original shape or not). Since most objects only elastically deform over small distances of deformation, high Yield values generally means it takes a lot of force to noticeably 'stretch' them (but see strain at yield).

Fracture: The fracture point is the amount of stress or force necessarily to cause the material to fail, or in other words, to break.

Strain at yield (sometimes incorrectly referred to as 'elasticity'): This variable tells you how much deformation occurs to the material while it is deforming elastically. That is, as long as the force is less than the yield strength, stress * strain at yield = deformation distance. The smaller the strain at yield, the less deformation occurs under stress.

Note: Strain at yield is the inverse of the Elastic Modulus. Thus a highly elastic material has low elastic modulus, and engages in less elastic collisions.

===Modes of Applying Force===
Impact: Force applied by a sudden strike, like a hammer.

Compressive: Force applied by exerting pressure on an object, like trying to squish something between your hands.

Tensile: Force applied by pulling on something, like suspending one object via another. (e.g., if you suspend an [[elf]] from a metal pole, you are applying a tensile force to the pole).

Torsion: Force applied by twisting something. Note that you're twisting some portion of the object relative to itself to cause a torsion stress to be applied to it. (Consider trying to twist a metal rod by grasping at either end and attempting to wring it - yes, you'd have to apply a lot of force to succeed).

Shear: Force applied by pushing part of the material so it tries to slide relative to another part of it. Ie, pushing at the top of an object when the bottom part is fixed to the ground is going to primarily apply a shear stress to it (the top part will try to move in the direction you push, and the lower part will resist this shear stress).

Bending: Force applied by bending a material.

==Effects on Combat==
The Dwarf Fortress combat system does not use all material properties at present (0.34.11). Weapon and armor damage/wear/decay is not tracked.

The mechanics governing how material properties impact the protection armor provides against projectiles have been [http://www.bay12forums.com/smf/index.php?topic=116151.0 studied reasonably carefully], and are described below. Armor protection from melee weapons has not been studied in detail, although it likely shares many similarities with ranged weapons. '''''Note: all of the below is based on recent (0.34.11) experiments, and many results may be approximate or incomplete.'''''

===Projectile Properties===
The main factor determining if a projectile penetrates armor is its momentum. Material properties of the bolt/arrow (except for IMPACT_YIELD, which will be described below), as well as bolt quality and marksdwarf/archer skills do not appear to matter (the skill of the marksdwarf determines if the bolt misses, hits, is dodged, or is blocked with a shield). Dwarven weapons launch projectiles with roughly constant momentum. However rounding of projectile masses can result in some very strange behavior. The complete calculation used to determine projectile momentum is as follows:
*Find the mass of your projectile. Mass=SOLID_DENSITY*SIZE/1,000,000.
*Floor that number to an integer.
*Divide the SHOOT_FORCE of the weapon by the integer, and round the result to the nearest integer.
*Cap the results by the SHOOT_MAXVEL for the weapon.
*Multiply the velocity by the ''unrounded'' mass of the projectile to get the momentum.
For DF34.11 values, iron, bronze, copper, silver, and steel bolts/arrows all have mass between 1 and 2, which will be floored to 1, meaning that they will have velocity equal to SHOOT_FORCE (so long as it is less than or equal to SHOOT_MAXVEL). Wood, adamantine, and bone bolts/arrows, as well as all blowdarts, have mass less than 1, and will have velocity which is always equal to SHOOT_MAXVEL and does not depend on SHOOT_FORCE.

===Interactions Between Projectiles and Armor===
[[File:projectile.png|200px|thumb|right|A simplified flowchart showing how material properties are used as a projectile contacts armor.]]
When a projectile strikes armor, there are several possible outcomes:
*Conversion of edged damage to blunt damage (chain mail does this).
*The projectile absorbs the force of the collision, and is deflected.
*The armor absorbs all or part of the force of the blow.
*The armor is fractured and does nothing to stop the projectile.
For armor to be at all effective at stopping projectiles, the armor material must have SHEAR_YIELD and/or SHEAR_FRACTURE greater than or equal to the projectile material. Otherwise, the projectile just cuts through the armor like it isn’t there. Additionally, the IMPACT_FRACTURE of the armor must be large relative to the projectile momentum, or the armor fractures and the projectile passes through without slowing down significantly (this happens for metal bolts against any plate armor in 0.34.11).

====Conversion of Edged Damage to Blunt Damage====
If the momentum of the projectile is not too high, then chain armor can convert the edged damage normally caused by projectiles to blunt damage, resulting in chips, fractures, jams and bruises, but no tears or cuts. Plate armor does not appear to provide this type of protection under any conditions tested so far. It is not yet known how the momentum needed for edged damage to penetrate chain armor is calculated, but for the cases examined the momentum needed is many times larger than those observed for in-game projectiles.

====The Projectile Absorbs the Force of the Collision====
If the projectile does not cut right through the armor (armor material must have SHEAR_YIELD and/or SHEAR_FRACTURE greater than or equal to the projectile material), then it must push in a chunk of the armor. If the armor is not strong enough to resist being pushed in (armor resisting the blow is covered in the next section), then the force to push in the armor is proportional to the mass of the chunk of armor. If the projectile material is not strong enough to exert this force, then it is deformed and deflects. The following algorithm determines if this type of deflection occurs:
*Calculate the volume of armor moved by the projectile. Multiply armor LAYER_SIZE by projectile CONTACT_AREA, and then round down to the nearest 100. If the result is less than 100, then instead round up to 100.
*Mass=armor SOLID_DENSITY times the volume from the last step
*If the IMPACT_YIELD of the projectile is less than mass*(800/157)/PROJECTILE_SIZE, then the projectile deforms and is reported as “deflected”.
In-game, this type of deflection is observed for wood bolts impacting metal armor, and explains why adamantine is observed to be the worst armor for deflecting wood bolts, as it has the lowest density. As explained below, wood bolts are capable of fracturing through the strongest metal armor, and would fracture copper or steel armor were it not for the higher densities of those metals.

====The Armor Absorbs the Force of the Collision====
This form of deflection depends only on the momentum, but not the material, of the projectile. With DF34.11 parameters, no armor actually uses this mechanism, as even wood bolts have too much momentum to be absorbed by any armor material.

The armor will provide 50% protection from projectiles if
:momentum = (IF-IY/2)*round100(C*S)/(2400000-AU*10000-Q*30000)
where IF=IMPACT_FRACTURE and IY=IMPACT_YIELD of the armor material, C is CONTACT_AREA of the projectile (although it is the smaller of the projectile's contact area and the body part struck's contact area), and S=LAYER_SIZE of the armor item(s) covering the body part struck, AU is the target's armor user skill (0=none, 16=legendary) and Q is the armor quality (0=none, 5=masterwork). The function round100(x) is equal to 100 if x is less than 100, and 100*floor(x/100) otherwise (rounded down to the nearest 100). It unclear if or how the thicknesses of multiple armor layers covering one body part stack. This formula is approximate, and the exact values of the coefficients in the denominator may not be accurate. The amount of protection provided drops sigmoidally from 100% to zero over a fairly narrow range.

The manner in which armor offers protection is different depending on whether the bolt momentum is greater than 50000/IMPACT_STRAIN_AT_YIELD. If momentum is less than this threshold, then armor can completely deflect projectiles, while if it is larger, armor can significantly slow projectiles so that they cause only bruises (not tears, chips, fractures, or jams).

{{Category|Materials|*}}

v0.34:Material science

2012-11-30T03:00:23Z

Pirate Bob: /* The Armor Absorbs the Force of the Collision */

{{quality|Fine|02:27, 11 May 2012 (UTC)}}{{av}}
[[Material|Materials]] have a number of properties representing real world variables that describe how they respond to inputs. In particular, the game now has a number of variables that describe what happens to a material when it's put under stress.

==What is stress?==
In the real world, an object is stressed when a force is applied to the object. Depending on the nature of the force applied, this stress can take a number of forms, and the object can respond differently based on its material and how that material handles different stresses.

In the material raws, whenever you see 'yield', 'fracture', or '[[Strain at yield|strain at yield]]', that property is a stress-related quality.

==When does Dwarf Fortress make stress calculations?==
At present, DF seems to only apply forces during combat, and thus only stresses objects (generally armor and various body layers) at that time.

==There's a lot of stress-related properties, what do they mean?==
The first thing you'll notice is that the second word in each stress variable is one of Yield, Fracture, or strain at yield. These are mechanical performance terms.

The first set of words are things like Impact, Bending, and so forth. These describe modes of applying force.

The following explanations assumes real world physics sort of apply (since [[Main:Toady One|Toady One]] chose real world properties).

===Mechanical Performance Properties===
Yield: This is almost certainly 'Yield Strength', which is the amount of stress needed to cause a material to go from elastic deformation to plastic deformation. (That is, if you cease stressing the object, does it revert to its original shape or not). Since most objects only elastically deform over small distances of deformation, high Yield values generally means it takes a lot of force to noticeably 'stretch' them (but see strain at yield).

Fracture: The fracture point is the amount of stress or force necessarily to cause the material to fail, or in other words, to break.

Strain at yield (sometimes incorrectly referred to as 'elasticity'): This variable tells you how much deformation occurs to the material while it is deforming elastically. That is, as long as the force is less than the yield strength, stress * strain at yield = deformation distance. The smaller the strain at yield, the less deformation occurs under stress.

Note: Strain at yield is the inverse of the Elastic Modulus. Thus a highly elastic material has low elastic modulus, and engages in less elastic collisions.

===Modes of Applying Force===
Impact: Force applied by a sudden strike, like a hammer.

Compressive: Force applied by exerting pressure on an object, like trying to squish something between your hands.

Tensile: Force applied by pulling on something, like suspending one object via another. (e.g., if you suspend an [[elf]] from a metal pole, you are applying a tensile force to the pole).

Torsion: Force applied by twisting something. Note that you're twisting some portion of the object relative to itself to cause a torsion stress to be applied to it. (Consider trying to twist a metal rod by grasping at either end and attempting to wring it - yes, you'd have to apply a lot of force to succeed).

Shear: Force applied by pushing part of the material so it tries to slide relative to another part of it. Ie, pushing at the top of an object when the bottom part is fixed to the ground is going to primarily apply a shear stress to it (the top part will try to move in the direction you push, and the lower part will resist this shear stress).

Bending: Force applied by bending a material.

===Hypotheses based on Real World===
*High strain at yield should reduce the effectiveness of a weapon. (the more it deforms, the longer it takes to break contact with the struck surface and thus the more inelastic the collision is).
*High strain at yield should be good for armor, because it decreases the force transferred by a weapon (to a point - it can't be so inelastic as to render the armor useless!)

==Effects on Combat==
Dwarf Fortress only features a limited combat system. Item decay does not seem to be simulated properly at all, so the hypotheses are largely incorrect.

The mechanics governing how material properties impact the protection armor provides against projectiles have been [http://www.bay12forums.com/smf/index.php?topic=116151.0 studied reasonably carefully], and are described below. Armor protection from melee weapons has not been studied in detail, although it likely shares many similarities with ranged weapons. '''''Note: all of the below is based on recent experiments, and many results may be approximate or incomplete.'''''

===Projectile Properties===
The main factor determining if a projectile penetrates armor is its momentum. Material properties of the bolt/arrow (except for IMPACT_YIELD, which will be described below), as well as bolt quality and marksdwarf/archer skills do not appear to matter. Dwarven weapons launch projectiles with roughly constant momentum. However rounding of projectile masses can result in some very strange behavior. The complete calculation used to determine projectile momentum is as follows:
*Find the mass of your projectile. Mass=SOLID_DENSITY*SIZE/1,000,000.
*Floor that number to an integer.
*Divide the SHOOT_FORCE of the weapon by the integer, and round the result to the nearest integer.
*Cap the results by the SHOOT_MAXVEL for the weapon.
*Multiply the velocity by the ''unrounded'' mass of the projectile to get the momentum.
For DF34.11 values, iron, bronze, copper, silver, and steel bolts/arrows all have mass between 1 and 2, which will be floored to 1, meaning that they will have velocity equal to SHOOT_FORCE (so long as it is less than or equal to SHOOT_MAXVEL). Wood, adamantine, and bone bolts/arrows, as well as all blowdarts, have mass less than 1, and will have velocity which is always equal to SHOOT_MAXVEL and does not depend on SHOOT_FORCE.

===Interactions Between Projectiles and Armor===
When a projectile strikes armor, there are several possible outcomes:
*Conversion of edged damage to blunt damage (chain mail does this).
*The projectile absorbs the force of the collision, and is deflected.
*The armor absorbs all or part of the force of the blow.
*The armor is fractured and does nothing to stop the projectile.
For armor to be at all effective at stopping projectiles, the armor material must have SHEAR_YIELD and/or SHEAR_FRACTURE greater than or equal to the projectile material. Otherwise, the projectile just cuts through the armor like it isn’t there. Additionally, the IMPACT_FRACTURE of the armor must be large relative to the projectile momentum, or the armor fractures and the projectile passes through without slowing down significantly (this happens for metal bolts against any plate armor in the current version). [Insert Urist's flowchart here]

====Conversion of Edged Damage to Blunt Damage====
If the momentum of the projectile is not too high, then chain armor can convert the edged damage normally caused by projectiles to blunt damage, resulting in chips, fractures, jams and bruises, but no tears or cuts. Plate armor does not appear to provide this type of protection under any conditions tested so far. It is not yet known how the momentum needed for edged damage to penetrate chain armor is calculated, but for the cases examined the momentum needed is many times larger than those observed for in-game projectiles. [Insert plot here]

====The Projectile Absorbs the Force of the Collision====
If the projectile does not cut right through the armor (armor material must have SHEAR_YIELD and/or SHEAR_FRACTURE greater than or equal to the projectile material), then it must push in a chunk of the armor. If the armor is not strong enough to resist being pushed in (armor resisting the blow is covered in the next section), then the force to push in the armor is proportional to the mass of the chunk of armor. If the projectile material is not strong enough to exert this force, then it is deformed and deflects. The following algorithm determines if this type of deflection occurs:
*Calculate the volume of armor moved by the projectile. Multiply armor LAYER_SIZE by projectile CONTACT_AREA, and then round down to the nearest 100. If the result is less than 100, then instead round up to 100.
*Mass=armor SOLID_DENSITY times the volume from the last step
*If the IMPACT_YIELD of the projectile is less than mass*(800/157)/PROJECTILE_SIZE, then the projectile deforms and is reported as “deflected”.
In-game, this type of deflection is observed for wood bolts impacting metal armor, and explains why adamantine is observed to be the worst armor for deflecting wood bolts, as it has the lowest density.

====The Armor Absorbs the Force of the Collision====

This form of deflection depends only on the momentum, but not the material, of the projectile. The armor will provide 50% protection from projectiles if
:momentum = (IF-IY/2)*round100(C*S)/(2400000-AU*10000-Q*30000)
where IF=IMPACT_FRACTURE and IY=IMPACT_YIELD of the armor material, C is CONTACT_AREA of the projectile (although it may be the smaller of the projectile and the body part struck), and S=LAYER_SIZE of the armor item(s) covering the body part struck, AU is the target's armor user skill (0=none, 16=legendary) and Q is the armor quality (0=none, 5=masterwork). The function round100(x) is equal to 100 if x is less than 100, and 100*floor(x/100) otherwise (rounded to the nearest 100). It unclear if or how the thicknesses of multiple armor layers covering one body part stack. This formula is approximate, and the exact values of the coefficients in the denominator may not be accurate. The amount of protection provided drops sigmoidally from 100% to zero over a fairly narrow range.

The manner in which armor offers protection is different depending on whether the bolt momentum is greater than 50000/IMPACT_STRAIN_AT_YIELD [I need to check if/how CONTACT_AREA influences this equation - this if for CONTACT_AREA=2]. If momentum is less than this threshold, then armor can completely deflect projectiles, while if it is larger, armor can significantly slow projectiles so that they cause only bruises (not tears, chips, fractures, or jams).

{{Category|Materials|*}}

v0.34:Material science

2012-11-30T02:52:37Z

Pirate Bob: /* The Armor Absorbs the Force of the Collision */

{{quality|Fine|02:27, 11 May 2012 (UTC)}}{{av}}
[[Material|Materials]] have a number of properties representing real world variables that describe how they respond to inputs. In particular, the game now has a number of variables that describe what happens to a material when it's put under stress.

==What is stress?==
In the real world, an object is stressed when a force is applied to the object. Depending on the nature of the force applied, this stress can take a number of forms, and the object can respond differently based on its material and how that material handles different stresses.

In the material raws, whenever you see 'yield', 'fracture', or '[[Strain at yield|strain at yield]]', that property is a stress-related quality.

==When does Dwarf Fortress make stress calculations?==
At present, DF seems to only apply forces during combat, and thus only stresses objects (generally armor and various body layers) at that time.

==There's a lot of stress-related properties, what do they mean?==
The first thing you'll notice is that the second word in each stress variable is one of Yield, Fracture, or strain at yield. These are mechanical performance terms.

The first set of words are things like Impact, Bending, and so forth. These describe modes of applying force.

The following explanations assumes real world physics sort of apply (since [[Main:Toady One|Toady One]] chose real world properties).

===Mechanical Performance Properties===
Yield: This is almost certainly 'Yield Strength', which is the amount of stress needed to cause a material to go from elastic deformation to plastic deformation. (That is, if you cease stressing the object, does it revert to its original shape or not). Since most objects only elastically deform over small distances of deformation, high Yield values generally means it takes a lot of force to noticeably 'stretch' them (but see strain at yield).

Fracture: The fracture point is the amount of stress or force necessarily to cause the material to fail, or in other words, to break.

Strain at yield (sometimes incorrectly referred to as 'elasticity'): This variable tells you how much deformation occurs to the material while it is deforming elastically. That is, as long as the force is less than the yield strength, stress * strain at yield = deformation distance. The smaller the strain at yield, the less deformation occurs under stress.

Note: Strain at yield is the inverse of the Elastic Modulus. Thus a highly elastic material has low elastic modulus, and engages in less elastic collisions.

===Modes of Applying Force===
Impact: Force applied by a sudden strike, like a hammer.

Compressive: Force applied by exerting pressure on an object, like trying to squish something between your hands.

Tensile: Force applied by pulling on something, like suspending one object via another. (e.g., if you suspend an [[elf]] from a metal pole, you are applying a tensile force to the pole).

Torsion: Force applied by twisting something. Note that you're twisting some portion of the object relative to itself to cause a torsion stress to be applied to it. (Consider trying to twist a metal rod by grasping at either end and attempting to wring it - yes, you'd have to apply a lot of force to succeed).

Shear: Force applied by pushing part of the material so it tries to slide relative to another part of it. Ie, pushing at the top of an object when the bottom part is fixed to the ground is going to primarily apply a shear stress to it (the top part will try to move in the direction you push, and the lower part will resist this shear stress).

Bending: Force applied by bending a material.

===Hypotheses based on Real World===
*High strain at yield should reduce the effectiveness of a weapon. (the more it deforms, the longer it takes to break contact with the struck surface and thus the more inelastic the collision is).
*High strain at yield should be good for armor, because it decreases the force transferred by a weapon (to a point - it can't be so inelastic as to render the armor useless!)

==Effects on Combat==
Dwarf Fortress only features a limited combat system. Item decay does not seem to be simulated properly at all, so the hypotheses are largely incorrect.

The mechanics governing how material properties impact the protection armor provides against projectiles have been [http://www.bay12forums.com/smf/index.php?topic=116151.0 studied reasonably carefully], and are described below. Armor protection from melee weapons has not been studied in detail, although it likely shares many similarities with ranged weapons. '''''Note: all of the below is based on recent experiments, and many results may be approximate or incomplete.'''''

===Projectile Properties===
The main factor determining if a projectile penetrates armor is its momentum. Material properties of the bolt/arrow (except for IMPACT_YIELD, which will be described below), as well as bolt quality and marksdwarf/archer skills do not appear to matter. Dwarven weapons launch projectiles with roughly constant momentum. However rounding of projectile masses can result in some very strange behavior. The complete calculation used to determine projectile momentum is as follows:
*Find the mass of your projectile. Mass=SOLID_DENSITY*SIZE/1,000,000.
*Floor that number to an integer.
*Divide the SHOOT_FORCE of the weapon by the integer, and round the result to the nearest integer.
*Cap the results by the SHOOT_MAXVEL for the weapon.
*Multiply the velocity by the ''unrounded'' mass of the projectile to get the momentum.
For DF34.11 values, iron, bronze, copper, silver, and steel bolts/arrows all have mass between 1 and 2, which will be floored to 1, meaning that they will have velocity equal to SHOOT_FORCE (so long as it is less than or equal to SHOOT_MAXVEL). Wood, adamantine, and bone bolts/arrows, as well as all blowdarts, have mass less than 1, and will have velocity which is always equal to SHOOT_MAXVEL and does not depend on SHOOT_FORCE.

===Interactions Between Projectiles and Armor===
When a projectile strikes armor, there are several possible outcomes:
*Conversion of edged damage to blunt damage (chain mail does this).
*The projectile absorbs the force of the collision, and is deflected.
*The armor absorbs all or part of the force of the blow.
*The armor is fractured and does nothing to stop the projectile.
For armor to be at all effective at stopping projectiles, the armor material must have SHEAR_YIELD and/or SHEAR_FRACTURE greater than or equal to the projectile material. Otherwise, the projectile just cuts through the armor like it isn’t there. Additionally, the IMPACT_FRACTURE of the armor must be large relative to the projectile momentum, or the armor fractures and the projectile passes through without slowing down significantly (this happens for metal bolts against any plate armor in the current version). [Insert Urist's flowchart here]

====Conversion of Edged Damage to Blunt Damage====
If the momentum of the projectile is not too high, then chain armor can convert the edged damage normally caused by projectiles to blunt damage, resulting in chips, fractures, jams and bruises, but no tears or cuts. Plate armor does not appear to provide this type of protection under any conditions tested so far. It is not yet known how the momentum needed for edged damage to penetrate chain armor is calculated, but for the cases examined the momentum needed is many times larger than those observed for in-game projectiles. [Insert plot here]

====The Projectile Absorbs the Force of the Collision====
If the projectile does not cut right through the armor (armor material must have SHEAR_YIELD and/or SHEAR_FRACTURE greater than or equal to the projectile material), then it must push in a chunk of the armor. If the armor is not strong enough to resist being pushed in (armor resisting the blow is covered in the next section), then the force to push in the armor is proportional to the mass of the chunk of armor. If the projectile material is not strong enough to exert this force, then it is deformed and deflects. The following algorithm determines if this type of deflection occurs:
*Calculate the volume of armor moved by the projectile. Multiply armor LAYER_SIZE by projectile CONTACT_AREA, and then round down to the nearest 100. If the result is less than 100, then instead round up to 100.
*Mass=armor SOLID_DENSITY times the volume from the last step
*If the IMPACT_YIELD of the projectile is less than mass*(800/157)/PROJECTILE_SIZE, then the projectile deforms and is reported as “deflected”.
In-game, this type of deflection is observed for wood bolts impacting metal armor, and explains why adamantine is observed to be the worst armor for deflecting wood bolts, as it has the lowest density.

====The Armor Absorbs the Force of the Collision====

This form of deflection depends only on the momentum, but not the material, of the projectile. The armor will provide 50% protection from projectiles if

momentum = (IF-IY/2)*round100(C*S)/(2400000-AU*10000-Q*30000)

where IF=IMPACT_FRACTURE and IY=IMPACT_YIELD of the armor material, C is CONTACT_AREA of the projectile (although it may be the smaller of the projectile and the body part struck), and S=LAYER_SIZE of the armor item(s) covering the body part struck, AU is the target's armor user skill (0=none, 16=legendary) and Q is the armor quality (0=none, 5=masterwork). The function round100(x) is equal to 100 if x is less than 100, and 100*floor(x/100) otherwise (rounded to the nearest 100). It unclear if or how the thicknesses of multiple armor layers covering one body part stack. This formula is approximate, and the exact values of the coefficients in the denominator may not be accurate. The amount of protection provided drops sigmoidally from 100% to zero over a fairly narrow range.

The manner in which armor offers protection is different depending on whether the bolt momentum is greater than 50000/IMPACT_STRAIN_AT_YIELD [I need to check if/how CONTACT_AREA influences this equation - this if for CONTACT_AREA=2]. If momentum is less than this threshold, then armor can completely deflect projectiles, while if it is larger, armor can significantly slow projectiles so that they cause only bruises (not tears, chips, fractures, or jams).

{{Category|Materials|*}}

v0.34:Material science

2012-11-29T23:29:10Z

Pirate Bob: /* Effects on Combat */

{{quality|Fine|02:27, 11 May 2012 (UTC)}}{{av}}
[[Material|Materials]] have a number of properties representing real world variables that describe how they respond to inputs. In particular, the game now has a number of variables that describe what happens to a material when it's put under stress.

==What is stress?==
In the real world, an object is stressed when a force is applied to the object. Depending on the nature of the force applied, this stress can take a number of forms, and the object can respond differently based on its material and how that material handles different stresses.

In the material raws, whenever you see 'yield', 'fracture', or '[[Strain at yield|strain at yield]]', that property is a stress-related quality.

==When does Dwarf Fortress make stress calculations?==
At present, DF seems to only apply forces during combat, and thus only stresses objects (generally armor and various body layers) at that time.

==There's a lot of stress-related properties, what do they mean?==
The first thing you'll notice is that the second word in each stress variable is one of Yield, Fracture, or strain at yield. These are mechanical performance terms.

The first set of words are things like Impact, Bending, and so forth. These describe modes of applying force.

The following explanations assumes real world physics sort of apply (since [[Main:Toady One|Toady One]] chose real world properties).

===Mechanical Performance Properties===
Yield: This is almost certainly 'Yield Strength', which is the amount of stress needed to cause a material to go from elastic deformation to plastic deformation. (That is, if you cease stressing the object, does it revert to its original shape or not). Since most objects only elastically deform over small distances of deformation, high Yield values generally means it takes a lot of force to noticeably 'stretch' them (but see strain at yield).

Fracture: The fracture point is the amount of stress or force necessarily to cause the material to fail, or in other words, to break.

Strain at yield (sometimes incorrectly referred to as 'elasticity'): This variable tells you how much deformation occurs to the material while it is deforming elastically. That is, as long as the force is less than the yield strength, stress * strain at yield = deformation distance. The smaller the strain at yield, the less deformation occurs under stress.

Note: Strain at yield is the inverse of the Elastic Modulus. Thus a highly elastic material has low elastic modulus, and engages in less elastic collisions.

===Modes of Applying Force===
Impact: Force applied by a sudden strike, like a hammer.

Compressive: Force applied by exerting pressure on an object, like trying to squish something between your hands.

Tensile: Force applied by pulling on something, like suspending one object via another. (e.g., if you suspend an [[elf]] from a metal pole, you are applying a tensile force to the pole).

Torsion: Force applied by twisting something. Note that you're twisting some portion of the object relative to itself to cause a torsion stress to be applied to it. (Consider trying to twist a metal rod by grasping at either end and attempting to wring it - yes, you'd have to apply a lot of force to succeed).

Shear: Force applied by pushing part of the material so it tries to slide relative to another part of it. Ie, pushing at the top of an object when the bottom part is fixed to the ground is going to primarily apply a shear stress to it (the top part will try to move in the direction you push, and the lower part will resist this shear stress).

Bending: Force applied by bending a material.

===Hypotheses based on Real World===
*High strain at yield should reduce the effectiveness of a weapon. (the more it deforms, the longer it takes to break contact with the struck surface and thus the more inelastic the collision is).
*High strain at yield should be good for armor, because it decreases the force transferred by a weapon (to a point - it can't be so inelastic as to render the armor useless!)

==Effects on Combat==
Dwarf Fortress only features a limited combat system. Item decay does not seem to be simulated properly at all, so the hypotheses are largely incorrect.

The mechanics governing how material properties impact the protection armor provides against projectiles have been [http://www.bay12forums.com/smf/index.php?topic=116151.0 studied reasonably carefully], and are described below. Armor protection from melee weapons has not been studied in detail, although it likely shares many similarities with ranged weapons. '''''Note: all of the below is based on recent experiments, and many results may be approximate or incomplete.'''''

===Projectile Properties===
The main factor determining if a projectile penetrates armor is its momentum. Material properties of the bolt/arrow (except for IMPACT_YIELD, which will be described below), as well as bolt quality and marksdwarf/archer skills do not appear to matter. Dwarven weapons launch projectiles with roughly constant momentum. However rounding of projectile masses can result in some very strange behavior. The complete calculation used to determine projectile momentum is as follows:
*Find the mass of your projectile. Mass=SOLID_DENSITY*SIZE/1,000,000.
*Floor that number to an integer.
*Divide the SHOOT_FORCE of the weapon by the integer, and round the result to the nearest integer.
*Cap the results by the SHOOT_MAXVEL for the weapon.
*Multiply the velocity by the ''unrounded'' mass of the projectile to get the momentum.
For DF34.11 values, iron, bronze, copper, silver, and steel bolts/arrows all have mass between 1 and 2, which will be floored to 1, meaning that they will have velocity equal to SHOOT_FORCE (so long as it is less than or equal to SHOOT_MAXVEL). Wood, adamantine, and bone bolts/arrows, as well as all blowdarts, have mass less than 1, and will have velocity which is always equal to SHOOT_MAXVEL and does not depend on SHOOT_FORCE.

===Interactions Between Projectiles and Armor===
When a projectile strikes armor, there are several possible outcomes:
*Conversion of edged damage to blunt damage (chain mail does this).
*The projectile absorbs the force of the collision, and is deflected.
*The armor absorbs all or part of the force of the blow.
*The armor is fractured and does nothing to stop the projectile.
For armor to be at all effective at stopping projectiles, the armor material must have SHEAR_YIELD and/or SHEAR_FRACTURE greater than or equal to the projectile material. Otherwise, the projectile just cuts through the armor like it isn’t there. Additionally, the IMPACT_FRACTURE of the armor must be large relative to the projectile momentum, or the armor fractures and the projectile passes through without slowing down significantly (this happens for metal bolts against any plate armor in the current version). [Insert Urist's flowchart here]

====Conversion of Edged Damage to Blunt Damage====
If the momentum of the projectile is not too high, then chain armor can convert the edged damage normally caused by projectiles to blunt damage, resulting in chips, fractures, jams and bruises, but no tears or cuts. Plate armor does not appear to provide this type of protection under any conditions tested so far. It is not yet known how the momentum needed for edged damage to penetrate chain armor is calculated, but for the cases examined the momentum needed is many times larger than those observed for in-game projectiles. [Insert plot here]

====The Projectile Absorbs the Force of the Collision====
If the projectile does not cut right through the armor (armor material must have SHEAR_YIELD and/or SHEAR_FRACTURE greater than or equal to the projectile material), then it must push in a chunk of the armor. If the armor is not strong enough to resist being pushed in (armor resisting the blow is covered in the next section), then the force to push in the armor is proportional to the mass of the chunk of armor. If the projectile material is not strong enough to exert this force, then it is deformed and deflects. The following algorithm determines if this type of deflection occurs:
*Calculate the volume of armor moved by the projectile. Multiply armor LAYER_SIZE by projectile CONTACT_AREA, and then round down to the nearest 100. If the result is less than 100, then instead round up to 100.
*Mass=armor SOLID_DENSITY times the volume from the last step
*If the IMPACT_YIELD of the projectile is less than mass*(800/157)/PROJECTILE_SIZE, then the projectile deforms and is reported as “deflected”.
In-game, this type of deflection is observed for wood bolts impacting metal armor, and explains why adamantine is observed to be the worst armor for deflecting wood bolts, as it has the lowest density.

====The Armor Absorbs the Force of the Collision====

This form of deflection depends only on the momentum, but not the material, of the projectile. The armor will provide 50% protection from projectiles if
momentum = (IF-IY/2)*round100(C*S)/(2400000-AU*10000-Q*30000)
where IF=IMPACT_FRACTURE and IY=IMPACT_YIELD of the armor material, C is CONTACT_AREA of the projectile (although it may be the smaller of the projectile and the body part struck), and S=LAYER_SIZE of the armor item(s) covering the body part struck, AU is the target's armor user skill (0=none, 16=legendary) and Q is the armor quality (0=none, 5=masterwork). The function round100(x) is equal to 100 if x is less than 100, and 100*floor(x/100) otherwise (rounded to the nearest 100). It unclear if or how the thicknesses of multiple armor layers covering one body part stack. This formula is approximate, and the exact values of the coefficients in the denominator may not be accurate. The amount of protection provided drops sigmoidally from 100% to zero over a fairly narrow range.

The manner in which armor offers protection is different depending on whether the bolt momentum is greater than 50000/IMPACT_STRAIN_AT_YIELD [I need to check if/how CONTACT_AREA influences this equation - this if for CONTACT_AREA=2]. If momentum is less than this threshold, then armor can completely deflect projectiles, while if it is larger, armor can significantly slow projectiles so that they cause only bruises (not tears, chips, fractures, or jams).

{{Category|Materials|*}}

v0.34:Material science

2012-11-29T23:28:11Z

Pirate Bob: /* The Armor Absorbs the Force of the Collision */

{{quality|Fine|02:27, 11 May 2012 (UTC)}}{{av}}
[[Material|Materials]] have a number of properties representing real world variables that describe how they respond to inputs. In particular, the game now has a number of variables that describe what happens to a material when it's put under stress.

==What is stress?==
In the real world, an object is stressed when a force is applied to the object. Depending on the nature of the force applied, this stress can take a number of forms, and the object can respond differently based on its material and how that material handles different stresses.

In the material raws, whenever you see 'yield', 'fracture', or '[[Strain at yield|strain at yield]]', that property is a stress-related quality.

==When does Dwarf Fortress make stress calculations?==
At present, DF seems to only apply forces during combat, and thus only stresses objects (generally armor and various body layers) at that time.

==There's a lot of stress-related properties, what do they mean?==
The first thing you'll notice is that the second word in each stress variable is one of Yield, Fracture, or strain at yield. These are mechanical performance terms.

The first set of words are things like Impact, Bending, and so forth. These describe modes of applying force.

The following explanations assumes real world physics sort of apply (since [[Main:Toady One|Toady One]] chose real world properties).

===Mechanical Performance Properties===
Yield: This is almost certainly 'Yield Strength', which is the amount of stress needed to cause a material to go from elastic deformation to plastic deformation. (That is, if you cease stressing the object, does it revert to its original shape or not). Since most objects only elastically deform over small distances of deformation, high Yield values generally means it takes a lot of force to noticeably 'stretch' them (but see strain at yield).

Fracture: The fracture point is the amount of stress or force necessarily to cause the material to fail, or in other words, to break.

Strain at yield (sometimes incorrectly referred to as 'elasticity'): This variable tells you how much deformation occurs to the material while it is deforming elastically. That is, as long as the force is less than the yield strength, stress * strain at yield = deformation distance. The smaller the strain at yield, the less deformation occurs under stress.

Note: Strain at yield is the inverse of the Elastic Modulus. Thus a highly elastic material has low elastic modulus, and engages in less elastic collisions.

===Modes of Applying Force===
Impact: Force applied by a sudden strike, like a hammer.

Compressive: Force applied by exerting pressure on an object, like trying to squish something between your hands.

Tensile: Force applied by pulling on something, like suspending one object via another. (e.g., if you suspend an [[elf]] from a metal pole, you are applying a tensile force to the pole).

Torsion: Force applied by twisting something. Note that you're twisting some portion of the object relative to itself to cause a torsion stress to be applied to it. (Consider trying to twist a metal rod by grasping at either end and attempting to wring it - yes, you'd have to apply a lot of force to succeed).

Shear: Force applied by pushing part of the material so it tries to slide relative to another part of it. Ie, pushing at the top of an object when the bottom part is fixed to the ground is going to primarily apply a shear stress to it (the top part will try to move in the direction you push, and the lower part will resist this shear stress).

Bending: Force applied by bending a material.

===Hypotheses based on Real World===
*High strain at yield should reduce the effectiveness of a weapon. (the more it deforms, the longer it takes to break contact with the struck surface and thus the more inelastic the collision is).
*High strain at yield should be good for armor, because it decreases the force transferred by a weapon (to a point - it can't be so inelastic as to render the armor useless!)

==Effects on Combat==
Dwarf Fortress only features a limited combat system. Item decay does not seem to be simulated properly at all, so the hypotheses are largely incorrect.

The mechanics governing how material properties impact the protection armor provides against projectiles have been [http://www.bay12forums.com/smf/index.php?topic=116151.0 studied reasonably carefully], and are described below. Armor protection from melee weapons has not been studied in detail, although it likely shares many similarities with ranged weapons.

===Projectile Properties===
The main factor determining if a projectile penetrates armor is its momentum. Material properties of the bolt/arrow (except for IMPACT_YIELD, which will be described below), as well as bolt quality and marksdwarf/archer skills do not appear to matter. Dwarven weapons launch projectiles with roughly constant momentum. However rounding of projectile masses can result in some very strange behavior. The complete calculation used to determine projectile momentum is as follows:
*Find the mass of your projectile. Mass=SOLID_DENSITY*SIZE/1,000,000.
*Floor that number to an integer.
*Divide the SHOOT_FORCE of the weapon by the integer, and round the result to the nearest integer.
*Cap the results by the SHOOT_MAXVEL for the weapon.
*Multiply the velocity by the ''unrounded'' mass of the projectile to get the momentum.
For DF34.11 values, iron, bronze, copper, silver, and steel bolts/arrows all have mass between 1 and 2, which will be floored to 1, meaning that they will have velocity equal to SHOOT_FORCE (so long as it is less than or equal to SHOOT_MAXVEL). Wood, adamantine, and bone bolts/arrows, as well as all blowdarts, have mass less than 1, and will have velocity which is always equal to SHOOT_MAXVEL and does not depend on SHOOT_FORCE.

===Interactions Between Projectiles and Armor===
When a projectile strikes armor, there are several possible outcomes:
*Conversion of edged damage to blunt damage (chain mail does this).
*The projectile absorbs the force of the collision, and is deflected.
*The armor absorbs all or part of the force of the blow.
*The armor is fractured and does nothing to stop the projectile.
For armor to be at all effective at stopping projectiles, the armor material must have SHEAR_YIELD and/or SHEAR_FRACTURE greater than or equal to the projectile material. Otherwise, the projectile just cuts through the armor like it isn’t there. Additionally, the IMPACT_FRACTURE of the armor must be large relative to the projectile momentum, or the armor fractures and the projectile passes through without slowing down significantly (this happens for metal bolts against any plate armor in the current version). [Insert Urist's flowchart here]

====Conversion of Edged Damage to Blunt Damage====
If the momentum of the projectile is not too high, then chain armor can convert the edged damage normally caused by projectiles to blunt damage, resulting in chips, fractures, jams and bruises, but no tears or cuts. Plate armor does not appear to provide this type of protection under any conditions tested so far. It is not yet known how the momentum needed for edged damage to penetrate chain armor is calculated, but for the cases examined the momentum needed is many times larger than those observed for in-game projectiles. [Insert plot here]

====The Projectile Absorbs the Force of the Collision====
If the projectile does not cut right through the armor (armor material must have SHEAR_YIELD and/or SHEAR_FRACTURE greater than or equal to the projectile material), then it must push in a chunk of the armor. If the armor is not strong enough to resist being pushed in (armor resisting the blow is covered in the next section), then the force to push in the armor is proportional to the mass of the chunk of armor. If the projectile material is not strong enough to exert this force, then it is deformed and deflects. The following algorithm determines if this type of deflection occurs:
*Calculate the volume of armor moved by the projectile. Multiply armor LAYER_SIZE by projectile CONTACT_AREA, and then round down to the nearest 100. If the result is less than 100, then instead round up to 100.
*Mass=armor SOLID_DENSITY times the volume from the last step
*If the IMPACT_YIELD of the projectile is less than mass*(800/157)/PROJECTILE_SIZE, then the projectile deforms and is reported as “deflected”.
In-game, this type of deflection is observed for wood bolts impacting metal armor, and explains why adamantine is observed to be the worst armor for deflecting wood bolts, as it has the lowest density.

====The Armor Absorbs the Force of the Collision====

This form of deflection depends only on the momentum, but not the material, of the projectile. The armor will provide 50% protection from projectiles if
momentum = (IF-IY/2)*round100(C*S)/(2400000-AU*10000-Q*30000)
where IF=IMPACT_FRACTURE and IY=IMPACT_YIELD of the armor material, C is CONTACT_AREA of the projectile (although it may be the smaller of the projectile and the body part struck), and S=LAYER_SIZE of the armor item(s) covering the body part struck, AU is the target's armor user skill (0=none, 16=legendary) and Q is the armor quality (0=none, 5=masterwork). The function round100(x) is equal to 100 if x is less than 100, and 100*floor(x/100) otherwise (rounded to the nearest 100). It unclear if or how the thicknesses of multiple armor layers covering one body part stack. This formula is approximate, and the exact values of the coefficients in the denominator may not be accurate. The amount of protection provided drops sigmoidally from 100% to zero over a fairly narrow range.

The manner in which armor offers protection is different depending on whether the bolt momentum is greater than 50000/IMPACT_STRAIN_AT_YIELD [I need to check if/how CONTACT_AREA influences this equation - this if for CONTACT_AREA=2]. If momentum is less than this threshold, then armor can completely deflect projectiles, while if it is larger, armor can significantly slow projectiles so that they cause only bruises (not tears, chips, fractures, or jams).

{{Category|Materials|*}}

v0.34:Material science

2012-11-29T23:27:06Z

Pirate Bob: /* The Armor Absorbs the Force of the Collision */

{{quality|Fine|02:27, 11 May 2012 (UTC)}}{{av}}
[[Material|Materials]] have a number of properties representing real world variables that describe how they respond to inputs. In particular, the game now has a number of variables that describe what happens to a material when it's put under stress.

==What is stress?==
In the real world, an object is stressed when a force is applied to the object. Depending on the nature of the force applied, this stress can take a number of forms, and the object can respond differently based on its material and how that material handles different stresses.

In the material raws, whenever you see 'yield', 'fracture', or '[[Strain at yield|strain at yield]]', that property is a stress-related quality.

==When does Dwarf Fortress make stress calculations?==
At present, DF seems to only apply forces during combat, and thus only stresses objects (generally armor and various body layers) at that time.

==There's a lot of stress-related properties, what do they mean?==
The first thing you'll notice is that the second word in each stress variable is one of Yield, Fracture, or strain at yield. These are mechanical performance terms.

The first set of words are things like Impact, Bending, and so forth. These describe modes of applying force.

The following explanations assumes real world physics sort of apply (since [[Main:Toady One|Toady One]] chose real world properties).

===Mechanical Performance Properties===
Yield: This is almost certainly 'Yield Strength', which is the amount of stress needed to cause a material to go from elastic deformation to plastic deformation. (That is, if you cease stressing the object, does it revert to its original shape or not). Since most objects only elastically deform over small distances of deformation, high Yield values generally means it takes a lot of force to noticeably 'stretch' them (but see strain at yield).

Fracture: The fracture point is the amount of stress or force necessarily to cause the material to fail, or in other words, to break.

Strain at yield (sometimes incorrectly referred to as 'elasticity'): This variable tells you how much deformation occurs to the material while it is deforming elastically. That is, as long as the force is less than the yield strength, stress * strain at yield = deformation distance. The smaller the strain at yield, the less deformation occurs under stress.

Note: Strain at yield is the inverse of the Elastic Modulus. Thus a highly elastic material has low elastic modulus, and engages in less elastic collisions.

===Modes of Applying Force===
Impact: Force applied by a sudden strike, like a hammer.

Compressive: Force applied by exerting pressure on an object, like trying to squish something between your hands.

Tensile: Force applied by pulling on something, like suspending one object via another. (e.g., if you suspend an [[elf]] from a metal pole, you are applying a tensile force to the pole).

Torsion: Force applied by twisting something. Note that you're twisting some portion of the object relative to itself to cause a torsion stress to be applied to it. (Consider trying to twist a metal rod by grasping at either end and attempting to wring it - yes, you'd have to apply a lot of force to succeed).

Shear: Force applied by pushing part of the material so it tries to slide relative to another part of it. Ie, pushing at the top of an object when the bottom part is fixed to the ground is going to primarily apply a shear stress to it (the top part will try to move in the direction you push, and the lower part will resist this shear stress).

Bending: Force applied by bending a material.

===Hypotheses based on Real World===
*High strain at yield should reduce the effectiveness of a weapon. (the more it deforms, the longer it takes to break contact with the struck surface and thus the more inelastic the collision is).
*High strain at yield should be good for armor, because it decreases the force transferred by a weapon (to a point - it can't be so inelastic as to render the armor useless!)

==Effects on Combat==
Dwarf Fortress only features a limited combat system. Item decay does not seem to be simulated properly at all, so the hypotheses are largely incorrect.

The mechanics governing how material properties impact the protection armor provides against projectiles have been [http://www.bay12forums.com/smf/index.php?topic=116151.0 studied reasonably carefully], and are described below. Armor protection from melee weapons has not been studied in detail, although it likely shares many similarities with ranged weapons.

===Projectile Properties===
The main factor determining if a projectile penetrates armor is its momentum. Material properties of the bolt/arrow (except for IMPACT_YIELD, which will be described below), as well as bolt quality and marksdwarf/archer skills do not appear to matter. Dwarven weapons launch projectiles with roughly constant momentum. However rounding of projectile masses can result in some very strange behavior. The complete calculation used to determine projectile momentum is as follows:
*Find the mass of your projectile. Mass=SOLID_DENSITY*SIZE/1,000,000.
*Floor that number to an integer.
*Divide the SHOOT_FORCE of the weapon by the integer, and round the result to the nearest integer.
*Cap the results by the SHOOT_MAXVEL for the weapon.
*Multiply the velocity by the ''unrounded'' mass of the projectile to get the momentum.
For DF34.11 values, iron, bronze, copper, silver, and steel bolts/arrows all have mass between 1 and 2, which will be floored to 1, meaning that they will have velocity equal to SHOOT_FORCE (so long as it is less than or equal to SHOOT_MAXVEL). Wood, adamantine, and bone bolts/arrows, as well as all blowdarts, have mass less than 1, and will have velocity which is always equal to SHOOT_MAXVEL and does not depend on SHOOT_FORCE.

===Interactions Between Projectiles and Armor===
When a projectile strikes armor, there are several possible outcomes:
*Conversion of edged damage to blunt damage (chain mail does this).
*The projectile absorbs the force of the collision, and is deflected.
*The armor absorbs all or part of the force of the blow.
*The armor is fractured and does nothing to stop the projectile.
For armor to be at all effective at stopping projectiles, the armor material must have SHEAR_YIELD and/or SHEAR_FRACTURE greater than or equal to the projectile material. Otherwise, the projectile just cuts through the armor like it isn’t there. Additionally, the IMPACT_FRACTURE of the armor must be large relative to the projectile momentum, or the armor fractures and the projectile passes through without slowing down significantly (this happens for metal bolts against any plate armor in the current version). [Insert Urist's flowchart here]

====Conversion of Edged Damage to Blunt Damage====
If the momentum of the projectile is not too high, then chain armor can convert the edged damage normally caused by projectiles to blunt damage, resulting in chips, fractures, jams and bruises, but no tears or cuts. Plate armor does not appear to provide this type of protection under any conditions tested so far. It is not yet known how the momentum needed for edged damage to penetrate chain armor is calculated, but for the cases examined the momentum needed is many times larger than those observed for in-game projectiles. [Insert plot here]

====The Projectile Absorbs the Force of the Collision====
If the projectile does not cut right through the armor (armor material must have SHEAR_YIELD and/or SHEAR_FRACTURE greater than or equal to the projectile material), then it must push in a chunk of the armor. If the armor is not strong enough to resist being pushed in (armor resisting the blow is covered in the next section), then the force to push in the armor is proportional to the mass of the chunk of armor. If the projectile material is not strong enough to exert this force, then it is deformed and deflects. The following algorithm determines if this type of deflection occurs:
*Calculate the volume of armor moved by the projectile. Multiply armor LAYER_SIZE by projectile CONTACT_AREA, and then round down to the nearest 100. If the result is less than 100, then instead round up to 100.
*Mass=armor SOLID_DENSITY times the volume from the last step
*If the IMPACT_YIELD of the projectile is less than mass*(800/157)/PROJECTILE_SIZE, then the projectile deforms and is reported as “deflected”.
In-game, this type of deflection is observed for wood bolts impacting metal armor, and explains why adamantine is observed to be the worst armor for deflecting wood bolts, as it has the lowest density.

====The Armor Absorbs the Force of the Collision====

This form of deflection depends only on the momentum, but not the material, of the projectile. The armor will provide 50% protection from projectiles if
momentum = (IF-IY/2)*round100(C*S)/(2400000-AU*10000-Q*30000)
where IF=IMPACT_FRACTURE and IY=IMPACT_YIELD of the armor material, C is CONTACT_AREA of the projectile (although it may be the smaller of the projectile and the body part struck), and S=LAYER_SIZE of the armor item(s) covering the body part struck, AU is the target's armor user skill (0=none, 16=legendary) and Q is the armor quality (0=none, 5=masterwork). The function round100(x) is equal to 100 if x is less than 100, and 100*floor(x/100) otherwise (rounded to the nearest 100). It unclear if or how the thicknesses of multiple armor layers covering one body part stack. This formula is approximate, and the exact values of the coefficients in the denominator may not be accurate. The amount of protection provided drops sigmoidally from 100% to zero over a fairly narrow range.

The manner in which armor offers protection is different depending on whether the bolt momentum is greater than 50000/IMPACT_STRAIN_AT_YIELD (it is not yet know if/how CONTACT_AREA influences this equation - this if for CONTACT_AREA=2). If momentum is less than this threshold, then armor can completely deflect projectiles, while if it is larger, armor can significantly slow projectiles so that they cause only bruises (not tears, chips, fractures, or jams).

{{Category|Materials|*}}

v0.34 Talk:Material science

2012-11-29T23:11:03Z

Pirate Bob:

v0.34:Material science

2012-11-29T23:06:32Z

Pirate Bob: added "The Projectile Absorbs the Force of the Collision"

{{quality|Fine|02:27, 11 May 2012 (UTC)}}{{av}}
[[Material|Materials]] have a number of properties representing real world variables that describe how they respond to inputs. In particular, the game now has a number of variables that describe what happens to a material when it's put under stress.

==What is stress?==
In the real world, an object is stressed when a force is applied to the object. Depending on the nature of the force applied, this stress can take a number of forms, and the object can respond differently based on its material and how that material handles different stresses.

In the material raws, whenever you see 'yield', 'fracture', or '[[Strain at yield|strain at yield]]', that property is a stress-related quality.

==When does Dwarf Fortress make stress calculations?==
At present, DF seems to only apply forces during combat, and thus only stresses objects (generally armor and various body layers) at that time.

==There's a lot of stress-related properties, what do they mean?==
The first thing you'll notice is that the second word in each stress variable is one of Yield, Fracture, or strain at yield. These are mechanical performance terms.

The first set of words are things like Impact, Bending, and so forth. These describe modes of applying force.

The following explanations assumes real world physics sort of apply (since [[Main:Toady One|Toady One]] chose real world properties).

===Mechanical Performance Properties===
Yield: This is almost certainly 'Yield Strength', which is the amount of stress needed to cause a material to go from elastic deformation to plastic deformation. (That is, if you cease stressing the object, does it revert to its original shape or not). Since most objects only elastically deform over small distances of deformation, high Yield values generally means it takes a lot of force to noticeably 'stretch' them (but see strain at yield).

Fracture: The fracture point is the amount of stress or force necessarily to cause the material to fail, or in other words, to break.

Strain at yield (sometimes incorrectly referred to as 'elasticity'): This variable tells you how much deformation occurs to the material while it is deforming elastically. That is, as long as the force is less than the yield strength, stress * strain at yield = deformation distance. The smaller the strain at yield, the less deformation occurs under stress.

Note: Strain at yield is the inverse of the Elastic Modulus. Thus a highly elastic material has low elastic modulus, and engages in less elastic collisions.

===Modes of Applying Force===
Impact: Force applied by a sudden strike, like a hammer.

Compressive: Force applied by exerting pressure on an object, like trying to squish something between your hands.

Tensile: Force applied by pulling on something, like suspending one object via another. (e.g., if you suspend an [[elf]] from a metal pole, you are applying a tensile force to the pole).

Torsion: Force applied by twisting something. Note that you're twisting some portion of the object relative to itself to cause a torsion stress to be applied to it. (Consider trying to twist a metal rod by grasping at either end and attempting to wring it - yes, you'd have to apply a lot of force to succeed).

Shear: Force applied by pushing part of the material so it tries to slide relative to another part of it. Ie, pushing at the top of an object when the bottom part is fixed to the ground is going to primarily apply a shear stress to it (the top part will try to move in the direction you push, and the lower part will resist this shear stress).

Bending: Force applied by bending a material.

===Hypotheses based on Real World===
*High strain at yield should reduce the effectiveness of a weapon. (the more it deforms, the longer it takes to break contact with the struck surface and thus the more inelastic the collision is).
*High strain at yield should be good for armor, because it decreases the force transferred by a weapon (to a point - it can't be so inelastic as to render the armor useless!)

==Effects on Combat==
Dwarf Fortress only features a limited combat system. Item decay does not seem to be simulated properly at all, so the hypotheses are largely incorrect.

The mechanics governing how material properties impact the protection armor provides against projectiles have been [http://www.bay12forums.com/smf/index.php?topic=116151.0 studied reasonably carefully], and are described below. Armor protection from melee weapons has not been studied in detail, although it likely shares many similarities with ranged weapons.

===Projectile Properties===
The main factor determining if a projectile penetrates armor is its momentum. Material properties of the bolt/arrow (except for IMPACT_YIELD, which will be described below), as well as bolt quality and marksdwarf/archer skills do not appear to matter. Dwarven weapons launch projectiles with roughly constant momentum. However rounding of projectile masses can result in some very strange behavior. The complete calculation used to determine projectile momentum is as follows:
*Find the mass of your projectile. Mass=SOLID_DENSITY*SIZE/1,000,000.
*Floor that number to an integer.
*Divide the SHOOT_FORCE of the weapon by the integer, and round the result to the nearest integer.
*Cap the results by the SHOOT_MAXVEL for the weapon.
*Multiply the velocity by the ''unrounded'' mass of the projectile to get the momentum.
For DF34.11 values, iron, bronze, copper, silver, and steel bolts/arrows all have mass between 1 and 2, which will be floored to 1, meaning that they will have velocity equal to SHOOT_FORCE (so long as it is less than or equal to SHOOT_MAXVEL). Wood, adamantine, and bone bolts/arrows, as well as all blowdarts, have mass less than 1, and will have velocity which is always equal to SHOOT_MAXVEL and does not depend on SHOOT_FORCE.

===Interactions Between Projectiles and Armor===
When a projectile strikes armor, there are several possible outcomes:
*Conversion of edged damage to blunt damage (chain mail does this).
*The projectile absorbs the force of the collision, and is deflected.
*The armor absorbs all or part of the force of the blow.
*The armor is fractured and does nothing to stop the projectile.
For armor to be at all effective at stopping projectiles, the armor material must have SHEAR_YIELD and/or SHEAR_FRACTURE greater than or equal to the projectile material. Otherwise, the projectile just cuts through the armor like it isn’t there. Additionally, the IMPACT_FRACTURE of the armor must be large relative to the projectile momentum, or the armor fractures and the projectile passes through without slowing down significantly (this happens for metal bolts against any plate armor in the current version). [Insert Urist's flowchart here]

====Conversion of Edged Damage to Blunt Damage====
If the momentum of the projectile is not too high, then chain armor can convert the edged damage normally caused by projectiles to blunt damage, resulting in chips, fractures, jams and bruises, but no tears or cuts. Plate armor does not appear to provide this type of protection under any conditions tested so far. It is not yet known how the momentum needed for edged damage to penetrate chain armor is calculated, but for the cases examined the momentum needed is many times larger than those observed for in-game projectiles. [Insert plot here]

====The Projectile Absorbs the Force of the Collision====
If the projectile does not cut right through the armor (armor material must have SHEAR_YIELD and/or SHEAR_FRACTURE greater than or equal to the projectile material), then it must push in a chunk of the armor. If the armor is not strong enough to resist being pushed in (armor resisting the blow is covered in the next section), then the force to push in the armor is proportional to the mass of the chunk of armor. If the projectile material is not strong enough to exert this force, then it is deformed and deflects. The following algorithm determines if this type of deflection occurs:
*Calculate the volume of armor moved by the projectile. Multiply armor LAYER_SIZE by projectile CONTACT_AREA, and then round down to the nearest 100. If the result is less than 100, then instead round up to 100.
*Mass=armor SOLID_DENSITY times the volume from the last step
*If the IMPACT_YIELD of the projectile is less than mass*(800/157)/PROJECTILE_SIZE, then the projectile deforms and is reported as “deflected”.
In-game, this type of deflection is observed for wood bolts impacting metal armor, and explains why adamantine is observed to be the worst armor for deflecting wood bolts, as it has the lowest density.
{{Category|Materials|*}}

v0.34:Material science

2012-11-29T22:52:10Z

Pirate Bob: /* Effects on Combat */

v0.34:Material science

2012-11-29T22:50:28Z

Pirate Bob: /* Effects on Combat */ - adding in results of arena testing for ballistics

v0.34 Talk:Material science

2012-11-29T22:49:33Z

Pirate Bob:

v0.34 Talk:Material science

2012-11-29T22:20:20Z

Pirate Bob:

v0.34 Talk:Material science

2012-11-29T22:19:56Z

Pirate Bob:

I don't understand what the first paragraph in the "Effects on Combat" section mean, so I am leaving them alone for now. I am starting to add the results of our ballistics testing to the combat section. --[[User:Pirate Bob|Pirate Bob]] 22:19, 29 November 2012 (UTC)

v0.34 Talk:Bolt

2012-09-10T15:55:05Z

Pirate Bob: /* This armor is useless! */

== Inconsistency? ==
"Tests show that where even masterwork wooden bolts bounce harmlessly off copper armor, iron bolts pierce without any issue."

"Tests show that even wooden bolts are effective against armored targets with the new release. Copper bolts have been known to pierce adamantine consistently."

These two statements flatly contradict each other, is either true? [[User:OrangePikmin|OrangePikmin]] 21:31, 7 April 2012 (UTC)

The problem is the assumption that adamantine is the best armor against bolts. I have done arena testing which clearly shows it is not. Steel and copper armor will both deflect 98% of pine bolts, while adamantine only deflects 58%. This is quite strange, and I am not sure why this is happening. My best guess is that, since adamantine is completely rigid and very light, it essentially converts the edged attack of a bolt into a blunt attack with the same force and contact area.

Summary of Arena Testing:
* Dwarves armored with breastplate, helm, greaves, 2 gauntlets, and 2 low boots of indicated material.
* Armed with adamantine crossbow, and 100 bolts of indicated material.
* Great archer/marksdwarf, grand master armor user.
* Usually 20 dwaves from each side positioned on either side of the 3 tile wide trenches in the standard arena.
* Used a perl script to count the indicated words within gamelog.txt
{| class="wikitable"
|-
! bolt material
! armor material
! hits ("flying")
! deflections ("deflect")
! through armor ("through")
! headshots ("head")
! fatal headshots ("killed")
|-
| pine
| steel
| 2011
| 1814
| 29
| 93
| 0
|-
| copper
| steel
| 856
| 0
| 772
| 38
| 26
|-
| iron
| steel
| 302
| 0
| 267
| 22
| 13
|-
| steel
| steel
| 319
| 0
| 295
| 21
| 13
|-
| pine
| adamantine
| 996
| 533
| 386
| 51
| 10
|-
| copper
| adamantine
| 708
| 0
| 649
| 36
| 26
|-
| bronze
| adamantine
| 569
| 0
| 522
| 31
| 23
|-
| iron
| adamantine
| 609
| 0
| 551
| 32
| 20
|-
| steel
| adamantine
| 663
| 0
| 594
| 36
| 17
|-
| pine
| copper
| 1390
| 1241
| 25
| 61
| 0
|-
| copper
| copper
| 579
| 0
| 522
| 26
| 11
|-
| bronze
| copper
| 552
| 0
| 510
| 25
| 18
|-
| iron
| copper
| 496
| 0
| 438
| 23
| 14
|-
| steel
| copper
| 512
| 0
| 459
| 22
| 11
|-
| pine
| pine
| 739
| 0
| 684
| 43
| 31
|-
| pine
| leather
| 740
| 0
| 695
| 37
| 27
|}
I am continuing testing, and will try to add a more complete table of results to the bolt wiki page when I am finished.--[[User:Pirate Bob|Pirate Bob]] 12:37, 15 August 2012 (UTC)

I now have a fairly large sample of results, and it doesn't look like any armor does anything to stop metal bolts. I kind of feel like going forward is a waste of time, since none of the armor actually, you know, does anything... --[[User:Pirate Bob|Pirate Bob]] 03:11, 17 August 2012 (UTC)

While it would be interesting, I unfortunately cannot test bone bolts in the arena, as they are not available. I did test wood on wood and wood on leather (both useless). I might also try testing other types (i.e. featherwood and whatever wood is densest) to get some idea of how bolt effectiveness varies with density, but doing more metal bolt tests when nothing stops them is not worth the several hours it would take.

I think instead I will redesign my test: On one side I will place a fixed number of unarmed, but armored, dwarves, and on the other side of the moat dwarves with the appropriate bolts for ammo. I will then record the average number of hits to kill the dwarf targets. It would be nice to also record the number of hits until each dwarf first falls unconscious, since this is usually what really matters in combat, but this will require a much fancier script to analyze. I am saving all of my logs, so I can always go back and re-analyze them for unconscious dwarves.

I will of course also get the same data already in my table from these tests. However, collecting data that says no armor deflects metal bolts is not interesting. I want to be able to grade the relative (in)effectiveness of each armor, and also compare survival with armor to without armor. If metal armors really do nothing to help survival compared to no armor, I would say that qualifies as a moderately serious bug, and I would proceed with trying to figure out how to modify bolt and/or crossbow raws to get more sensible behavior. --[[User:Pirate Bob|Pirate Bob]] 12:19, 17 August 2012 (UTC)

:Side note: it's "adamanti'''ne'''", not "adamanti'''um'''" - this is Dwarf Fortress, not Marvel Comics. --[[User:Quietust|Quietust]] 12:50, 15 August 2012 (UTC)

: This is great research, Bob. Please continue testing. Ideally test every combination of arrow material and every armor. try to include wooden, leather and bone armor. And Bone arrows, too, please. If I find the time, I may try to verify your results. I did not know that you have a gamelog.txt of the fights in the arena. I can write a Python script analyzing the figthts, then. --[[User:Nagidal|Nagidal]] 13:16, 15 August 2012 (UTC)

: I found that I need to slightly improve my script so that it only counts "deflect" and "through" in lines that also contain "flying *** bolt". I thought that by putting the dorfs on opposite sides of a moat so there can't be melee combat I would prevent all non-bolt hits, but I forgot to account for them dodging into the moat and falling to their deaths. Also, in case it's not obvious to anyone, the number of deflections plus the number of armor penetrations does not add up to the number of hits because some parts of the dwarves (upper arms, neck) are not armored. I could add chainmail to fix this, but I was more interested in the effectiveness of bolts against plate armor. I also was considering some of the tests with leather clothing underneath to see if this has any effect. Especially for adamantine armor. Theoretically it should help cushion the impact (real knights would always wear padding under metal armor), but I don't know if DF models that. --[[User:Pirate Bob|Pirate Bob]] 00:14, 17 August 2012 (UTC)
: I have updated my perl script to only count statistics for lines containing "flying", and I am working on filling out the table. I wrote some macros to help with arena setup. --[[User:Pirate Bob|Pirate Bob]] 01:45, 17 August 2012 (UTC)

== Decorated bolts ==

There used to be an exploit wherein encrusting bolts with gems gave them enormous value boost because decoration value is counted for every bolt in stack. As I get it, Toady fixed it by setting value multiplier for bolt decorations to 1/3 instead of 10. Should be this mentioned somewhere? Also, if you manage to get a stack of more than 30 bolts, this is still an issue - e.g., for bolts[60] decorations value is doubled. --04:17, 21 May 2012 (UTC)

== This armor is useless! ==

I did some arena testing of the number of hits required to kill unarmed dwarf "targets". In all tests, I placed 224 targets, filling all squares adjacent to the 3-wide moat, and the opposite sides I place 242 shooters (I placed a few extra shooters by the "bridge" on the right side). I then let the arena run until all targets were dead, and counted up the total number of hits (counted the word "flying" in gamelog.txt).

{| class="wikitable"
|-
! Ammo
! Armor
! Dwarf Targets
! Hits
! Hits/kill
|-
| copper
| adamantine
| 224
| 3985
| 17.8
|-
| copper
| copper
| 224
| 3819
| 17.1
|-
| copper
| none
| 224
| 3894
| 17.4
|}

The numbers speak for themselves - armor is completely useless against even copper bolts. There is no significant difference between any of the results - I think the small differences are just due to random statistics. I will try to test this with a few more combinations, and maybe repeat a few of the tests for more statistics, but my preliminary results have pretty much confirmed my suspicion that armor does nothing against metal bolts. This just leaves the question, [http://www.youtube.com/watch?v=aqoDL4KlxvY why do we even wear it] :P ? --[[User:Pirate Bob|Pirate Bob]] 20:40, 17 August 2012 (UTC)
For more science you could mod a little and test again. For example mod the bolts to blunt attack and see if it has any differences in piercing adamantine. Or make the contact area larger. Also I recall an old [http://www.bay12forums.com/smf/index.php?topic=89997.msg2488474#msg2488474 experiment]about weight affecting bolts, the result was no significant difference. So there is a paradox: if we accept that weight has no effect, the pine bolts show meterial sharpness has some effect, while the variant metal bolts show that sharpness has little to no effect. So I advice do the experiment again with silver and adamantine bolts. --[[User:Lcy03406|Lcy03406]] 04:45, 18 August 2012 (UTC)

I have tried silver and adamantine bolts, and neither are ever deflected by copper armor. I will post details if I have time, but posting to the wiki is rather cumbersome - I think I may start a forum post for more involved discussion, as this is proving more complicated that I initially expected. I believe that the main thing that determines armor penetration and damage is kinetic energy. A bow/crossbow applies a force of 1000 (units?), over some fixed distance (which is not indicated in the raws), such that the energy applied does not depend on mass. However, there is also a maximum bolt speed of 1000 (units?), so if a bolt is very light (i.e. wood) its kinetic energy will just be its mass times 1000^2. I also found that copper blowdarts are never deflected by steel armor, and was unable to find any manipulation of the raws which would allow projectiles to be defected regularly but still do modest damage to unarmored targets. In particular, increasing the contact area to 20 just makes bolts *more* deadly, as they start to sever limbs etc, and are still never deflected by armor. I am unsure if what Toady changed about bolts has anything to do with the raws, but I noticed that on the DF2010 wiki page is says the raws changed in 0.31.09, and it gives different values for the raws than used currently, so I'll have to try the old values. Anyway, I have lots of data, but still far less than I would like, and I need to figure out a better way to discuss it, and perhaps get more people involved in collecting/analyzing it. --[[User:Pirate Bob|Pirate Bob]] 17:45, 28 August 2012 (UTC)
At the suggestion of a friend, I wrote a script which automatically tests every armor/ammo combination, so after I finish analyzing the results of that I will post complete results. Only surprise was that adamantine deflects adamantine 100% of the time, but no other armor/ammo combination produced any deflections. --[[User:Pirate Bob|Pirate Bob]] 12:23, 29 August 2012 (UTC)

My results have proven somewhat complex, and I felt a forum thread might be a better place for detailed discussion. I was also pleased to find that Zivilin had just posted similar results, so I added mine to his thread: [http://www.bay12forums.com/smf/index.php?topic=116151.0]--[[User:Pirate Bob|Pirate Bob]] 15:55, 10 September 2012 (UTC)

v0.34 Talk:Bolt

2012-08-29T12:23:52Z

Pirate Bob: /* This armor is useless! */

v0.34 Talk:Bolt

2012-08-28T17:50:27Z

Pirate Bob: /* This armor is useless! */

v0.34 Talk:Bolt

2012-08-28T17:45:25Z

Pirate Bob: /* This armor is useless! */

== Inconsistency? ==
"Tests show that where even masterwork wooden bolts bounce harmlessly off copper armor, iron bolts pierce without any issue."

"Tests show that even wooden bolts are effective against armored targets with the new release. Copper bolts have been known to pierce adamantine consistently."

These two statements flatly contradict each other, is either true? [[User:OrangePikmin|OrangePikmin]] 21:31, 7 April 2012 (UTC)

The problem is the assumption that adamantine is the best armor against bolts. I have done arena testing which clearly shows it is not. Steel and copper armor will both deflect 98% of pine bolts, while adamantine only deflects 58%. This is quite strange, and I am not sure why this is happening. My best guess is that, since adamantine is completely rigid and very light, it essentially converts the edged attack of a bolt into a blunt attack with the same force and contact area.

Summary of Arena Testing:
* Dwarves armored with breastplate, helm, greaves, 2 gauntlets, and 2 low boots of indicated material.
* Armed with adamantine crossbow, and 100 bolts of indicated material.
* Great archer/marksdwarf, grand master armor user.
* Usually 20 dwaves from each side positioned on either side of the 3 tile wide trenches in the standard arena.
* Used a perl script to count the indicated words within gamelog.txt
{| class="wikitable"
|-
! bolt material
! armor material
! hits ("flying")
! deflections ("deflect")
! through armor ("through")
! headshots ("head")
! fatal headshots ("killed")
|-
| pine
| steel
| 2011
| 1814
| 29
| 93
| 0
|-
| copper
| steel
| 856
| 0
| 772
| 38
| 26
|-
| iron
| steel
| 302
| 0
| 267
| 22
| 13
|-
| steel
| steel
| 319
| 0
| 295
| 21
| 13
|-
| pine
| adamantine
| 996
| 533
| 386
| 51
| 10
|-
| copper
| adamantine
| 708
| 0
| 649
| 36
| 26
|-
| bronze
| adamantine
| 569
| 0
| 522
| 31
| 23
|-
| iron
| adamantine
| 609
| 0
| 551
| 32
| 20
|-
| steel
| adamantine
| 663
| 0
| 594
| 36
| 17
|-
| pine
| copper
| 1390
| 1241
| 25
| 61
| 0
|-
| copper
| copper
| 579
| 0
| 522
| 26
| 11
|-
| bronze
| copper
| 552
| 0
| 510
| 25
| 18
|-
| iron
| copper
| 496
| 0
| 438
| 23
| 14
|-
| steel
| copper
| 512
| 0
| 459
| 22
| 11
|-
| pine
| pine
| 739
| 0
| 684
| 43
| 31
|-
| pine
| leather
| 740
| 0
| 695
| 37
| 27
|}
I am continuing testing, and will try to add a more complete table of results to the bolt wiki page when I am finished.--[[User:Pirate Bob|Pirate Bob]] 12:37, 15 August 2012 (UTC)

I now have a fairly large sample of results, and it doesn't look like any armor does anything to stop metal bolts. I kind of feel like going forward is a waste of time, since none of the armor actually, you know, does anything... --[[User:Pirate Bob|Pirate Bob]] 03:11, 17 August 2012 (UTC)

While it would be interesting, I unfortunately cannot test bone bolts in the arena, as they are not available. I did test wood on wood and wood on leather (both useless). I might also try testing other types (i.e. featherwood and whatever wood is densest) to get some idea of how bolt effectiveness varies with density, but doing more metal bolt tests when nothing stops them is not worth the several hours it would take.

I think instead I will redesign my test: On one side I will place a fixed number of unarmed, but armored, dwarves, and on the other side of the moat dwarves with the appropriate bolts for ammo. I will then record the average number of hits to kill the dwarf targets. It would be nice to also record the number of hits until each dwarf first falls unconscious, since this is usually what really matters in combat, but this will require a much fancier script to analyze. I am saving all of my logs, so I can always go back and re-analyze them for unconscious dwarves.

I will of course also get the same data already in my table from these tests. However, collecting data that says no armor deflects metal bolts is not interesting. I want to be able to grade the relative (in)effectiveness of each armor, and also compare survival with armor to without armor. If metal armors really do nothing to help survival compared to no armor, I would say that qualifies as a moderately serious bug, and I would proceed with trying to figure out how to modify bolt and/or crossbow raws to get more sensible behavior. --[[User:Pirate Bob|Pirate Bob]] 12:19, 17 August 2012 (UTC)

:Side note: it's "adamanti'''ne'''", not "adamanti'''um'''" - this is Dwarf Fortress, not Marvel Comics. --[[User:Quietust|Quietust]] 12:50, 15 August 2012 (UTC)

: This is great research, Bob. Please continue testing. Ideally test every combination of arrow material and every armor. try to include wooden, leather and bone armor. And Bone arrows, too, please. If I find the time, I may try to verify your results. I did not know that you have a gamelog.txt of the fights in the arena. I can write a Python script analyzing the figthts, then. --[[User:Nagidal|Nagidal]] 13:16, 15 August 2012 (UTC)

: I found that I need to slightly improve my script so that it only counts "deflect" and "through" in lines that also contain "flying *** bolt". I thought that by putting the dorfs on opposite sides of a moat so there can't be melee combat I would prevent all non-bolt hits, but I forgot to account for them dodging into the moat and falling to their deaths. Also, in case it's not obvious to anyone, the number of deflections plus the number of armor penetrations does not add up to the number of hits because some parts of the dwarves (upper arms, neck) are not armored. I could add chainmail to fix this, but I was more interested in the effectiveness of bolts against plate armor. I also was considering some of the tests with leather clothing underneath to see if this has any effect. Especially for adamantine armor. Theoretically it should help cushion the impact (real knights would always wear padding under metal armor), but I don't know if DF models that. --[[User:Pirate Bob|Pirate Bob]] 00:14, 17 August 2012 (UTC)
: I have updated my perl script to only count statistics for lines containing "flying", and I am working on filling out the table. I wrote some macros to help with arena setup. --[[User:Pirate Bob|Pirate Bob]] 01:45, 17 August 2012 (UTC)

== Decorated bolts ==

There used to be an exploit wherein encrusting bolts with gems gave them enormous value boost because decoration value is counted for every bolt in stack. As I get it, Toady fixed it by setting value multiplier for bolt decorations to 1/3 instead of 10. Should be this mentioned somewhere? Also, if you manage to get a stack of more than 30 bolts, this is still an issue - e.g., for bolts[60] decorations value is doubled. --04:17, 21 May 2012 (UTC)

== This armor is useless! ==

I did some arena testing of the number of hits required to kill unarmed dwarf "targets". In all tests, I placed 224 targets, filling all squares adjacent to the 3-wide moat, and the opposite sides I place 242 shooters (I placed a few extra shooters by the "bridge" on the right side). I then let the arena run until all targets were dead, and counted up the total number of hits (counted the word "flying" in gamelog.txt).

{| class="wikitable"
|-
! Ammo
! Armor
! Dwarf Targets
! Hits
! Hits/kill
|-
| copper
| adamantine
| 224
| 3985
| 17.8
|-
| copper
| copper
| 224
| 3819
| 17.1
|-
| copper
| none
| 224
| 3894
| 17.4
|}

The numbers speak for themselves - armor is completely useless against even copper bolts. There is no significant difference between any of the results - I think the small differences are just due to random statistics. I will try to test this with a few more combinations, and maybe repeat a few of the tests for more statistics, but my preliminary results have pretty much confirmed my suspicion that armor does nothing against metal bolts. This just leaves the question, [http://www.youtube.com/watch?v=aqoDL4KlxvY why do we even wear it] :P ? --[[User:Pirate Bob|Pirate Bob]] 20:40, 17 August 2012 (UTC)
For more science you could mod a little and test again. For example mod the bolts to blunt attack and see if it has any differences in piercing adamantine. Or make the contact area larger. Also I recall an old [http://www.bay12forums.com/smf/index.php?topic=89997.msg2488474#msg2488474 experiment]about weight affecting bolts, the result was no significant difference. So there is a paradox: if we accept that weight has no effect, the pine bolts show meterial sharpness has some effect, while the variant metal bolts show that sharpness has little to no effect. So I advice do the experiment again with silver and adamantine bolts. --[[User:Lcy03406|Lcy03406]] 04:45, 18 August 2012 (UTC)

I have tried silver and adamantine bolts, and neither are ever deflected by copper armor. I will post details if I have time, but posting to the wiki is rather cumbersome - I think I may start a forum post for more involved discussion, as this is proving more complicated that I initially expected. I believe that the main thing that determines armor penetration and damage is kinetic energy. A bow/crossbow applies a force of 1000 (units?), over some fixed distance (which is not indicated in the raws), such that the energy applied does not depend on mass. However, there is also a maximum bolt speed of 1000 (units?), so if a bolt is very light (i.e. wood) its kinetic energy will just be its mass times 1000^2. I also found that copper blowdarts are never deflected by steel armor, and was unable to find any manipulation of the raws which would allow projectiles to be defected regularly but still do modest damage to unarmored targets. In particular, increasing the contact area to 20 just makes bolts *more* deadly, as they start to sever limbs etc, and are still never deflected by armor. I also checked the raws from DF 0.31 and found the contact area was still 2 (not 20 as reported on the wiki) in DF2010, so I don't think what Toady changed with projectiles has anything to do with the raws (unless the change happened before DF2010?). Anyway, I have lots of data, but still far less than I would like, and I need to figure out a better way to discuss it, and perhaps get more people involved in collecting/analyzing it. --[[User:Pirate Bob|Pirate Bob]] 17:45, 28 August 2012 (UTC)

v0.34 Talk:Bolt

2012-08-17T20:50:25Z

Pirate Bob: /* This armor is useless! */

v0.34 Talk:Bolt

2012-08-17T20:42:49Z

Pirate Bob: /* This armor is useless! */

== Inconsistency? ==
"Tests show that where even masterwork wooden bolts bounce harmlessly off copper armor, iron bolts pierce without any issue."

"Tests show that even wooden bolts are effective against armored targets with the new release. Copper bolts have been known to pierce adamantine consistently."

These two statements flatly contradict each other, is either true? [[User:OrangePikmin|OrangePikmin]] 21:31, 7 April 2012 (UTC)

The problem is the assumption that adamantine is the best armor against bolts. I have done arena testing which clearly shows it is not. Steel and copper armor will both deflect 98% of pine bolts, while adamantine only deflects 58%. This is quite strange, and I am not sure why this is happening. My best guess is that, since adamantine is completely rigid and very light, it essentially converts the edged attack of a bolt into a blunt attack with the same force and contact area.

Summary of Arena Testing:
* Dwarves armored with breastplate, helm, greaves, 2 gauntlets, and 2 low boots of indicated material.
* Armed with adamantine crossbow, and 100 bolts of indicated material.
* Great archer/marksdwarf, grand master armor user.
* Usually 20 dwaves from each side positioned on either side of the 3 tile wide trenches in the standard arena.
* Used a perl script to count the indicated words within gamelog.txt
{| class="wikitable"
|-
! bolt material
! armor material
! hits ("flying")
! deflections ("deflect")
! through armor ("through")
! headshots ("head")
! fatal headshots ("killed")
|-
| pine
| steel
| 2011
| 1814
| 29
| 93
| 0
|-
| copper
| steel
| 856
| 0
| 772
| 38
| 26
|-
| iron
| steel
| 302
| 0
| 267
| 22
| 13
|-
| steel
| steel
| 319
| 0
| 295
| 21
| 13
|-
| pine
| adamantine
| 996
| 533
| 386
| 51
| 10
|-
| copper
| adamantine
| 708
| 0
| 649
| 36
| 26
|-
| bronze
| adamantine
| 569
| 0
| 522
| 31
| 23
|-
| iron
| adamantine
| 609
| 0
| 551
| 32
| 20
|-
| steel
| adamantine
| 663
| 0
| 594
| 36
| 17
|-
| pine
| copper
| 1390
| 1241
| 25
| 61
| 0
|-
| copper
| copper
| 579
| 0
| 522
| 26
| 11
|-
| bronze
| copper
| 552
| 0
| 510
| 25
| 18
|-
| iron
| copper
| 496
| 0
| 438
| 23
| 14
|-
| steel
| copper
| 512
| 0
| 459
| 22
| 11
|-
| pine
| pine
| 739
| 0
| 684
| 43
| 31
|-
| pine
| leather
| 740
| 0
| 695
| 37
| 27
|}
I am continuing testing, and will try to add a more complete table of results to the bolt wiki page when I am finished.--[[User:Pirate Bob|Pirate Bob]] 12:37, 15 August 2012 (UTC)

I now have a fairly large sample of results, and it doesn't look like any armor does anything to stop metal bolts. I kind of feel like going forward is a waste of time, since none of the armor actually, you know, does anything... --[[User:Pirate Bob|Pirate Bob]] 03:11, 17 August 2012 (UTC)

While it would be interesting, I unfortunately cannot test bone bolts in the arena, as they are not available. I did test wood on wood and wood on leather (both useless). I might also try testing other types (i.e. featherwood and whatever wood is densest) to get some idea of how bolt effectiveness varies with density, but doing more metal bolt tests when nothing stops them is not worth the several hours it would take.

I think instead I will redesign my test: On one side I will place a fixed number of unarmed, but armored, dwarves, and on the other side of the moat dwarves with the appropriate bolts for ammo. I will then record the average number of hits to kill the dwarf targets. It would be nice to also record the number of hits until each dwarf first falls unconscious, since this is usually what really matters in combat, but this will require a much fancier script to analyze. I am saving all of my logs, so I can always go back and re-analyze them for unconscious dwarves.

I will of course also get the same data already in my table from these tests. However, collecting data that says no armor deflects metal bolts is not interesting. I want to be able to grade the relative (in)effectiveness of each armor, and also compare survival with armor to without armor. If metal armors really do nothing to help survival compared to no armor, I would say that qualifies as a moderately serious bug, and I would proceed with trying to figure out how to modify bolt and/or crossbow raws to get more sensible behavior. --[[User:Pirate Bob|Pirate Bob]] 12:19, 17 August 2012 (UTC)

:Side note: it's "adamanti'''ne'''", not "adamanti'''um'''" - this is Dwarf Fortress, not Marvel Comics. --[[User:Quietust|Quietust]] 12:50, 15 August 2012 (UTC)

: This is great research, Bob. Please continue testing. Ideally test every combination of arrow material and every armor. try to include wooden, leather and bone armor. And Bone arrows, too, please. If I find the time, I may try to verify your results. I did not know that you have a gamelog.txt of the fights in the arena. I can write a Python script analyzing the figthts, then. --[[User:Nagidal|Nagidal]] 13:16, 15 August 2012 (UTC)

: I found that I need to slightly improve my script so that it only counts "deflect" and "through" in lines that also contain "flying *** bolt". I thought that by putting the dorfs on opposite sides of a moat so there can't be melee combat I would prevent all non-bolt hits, but I forgot to account for them dodging into the moat and falling to their deaths. Also, in case it's not obvious to anyone, the number of deflections plus the number of armor penetrations does not add up to the number of hits because some parts of the dwarves (upper arms, neck) are not armored. I could add chainmail to fix this, but I was more interested in the effectiveness of bolts against plate armor. I also was considering some of the tests with leather clothing underneath to see if this has any effect. Especially for adamantine armor. Theoretically it should help cushion the impact (real knights would always wear padding under metal armor), but I don't know if DF models that. --[[User:Pirate Bob|Pirate Bob]] 00:14, 17 August 2012 (UTC)
: I have updated my perl script to only count statistics for lines containing "flying", and I am working on filling out the table. I wrote some macros to help with arena setup. --[[User:Pirate Bob|Pirate Bob]] 01:45, 17 August 2012 (UTC)

== Decorated bolts ==

There used to be an exploit wherein encrusting bolts with gems gave them enormous value boost because decoration value is counted for every bolt in stack. As I get it, Toady fixed it by setting value multiplier for bolt decorations to 1/3 instead of 10. Should be this mentioned somewhere? Also, if you manage to get a stack of more than 30 bolts, this is still an issue - e.g., for bolts[60] decorations value is doubled. --04:17, 21 May 2012 (UTC)

== This armor is useless! ==

I did some arena testing of the number of hits required to kill unarmed dwarf "targets". In all tests, I placed 224 targets, filling all squares adjacent to the 3-wide moat, and the opposite sides I place 242 shooters (I placed a few extra shooters by the "bridge" on the right side). I then let the arena run until all targets were dead, and counted up the total number of hits (counted the word "flying" in gamelog.txt).

{| class="wikitable"
|-
! Ammo
! Armor
! Dwarf Targets
! Hits
! Hits/kill
|-
| copper
| adamantine
| 224
| 3985
| 17.8 +/- 1.2
|-
| copper
| copper
| 224
| 3819
| 17.1 +/- 1.1
|-
| copper
| none
| 224
| 3894
| 17.4 +/- 1.2
|}

The numbers speak for themselves - armor is completely useless against even copper bolts. I will try to test this with a few more combinations, and maybe repeat a few of the tests for more statistics, but my preliminary results have pretty much confirmed my suspicion that armor does nothing against metal bolts. This just leaves the question, [http://www.youtube.com/watch?v=aqoDL4KlxvY why do we even wear it] :P ? --[[User:Pirate Bob|Pirate Bob]] 20:40, 17 August 2012 (UTC)

v0.34 Talk:Bolt

2012-08-17T20:40:21Z

Pirate Bob: /* This armor is useless! */

== Inconsistency? ==
"Tests show that where even masterwork wooden bolts bounce harmlessly off copper armor, iron bolts pierce without any issue."

"Tests show that even wooden bolts are effective against armored targets with the new release. Copper bolts have been known to pierce adamantine consistently."

These two statements flatly contradict each other, is either true? [[User:OrangePikmin|OrangePikmin]] 21:31, 7 April 2012 (UTC)

The problem is the assumption that adamantine is the best armor against bolts. I have done arena testing which clearly shows it is not. Steel and copper armor will both deflect 98% of pine bolts, while adamantine only deflects 58%. This is quite strange, and I am not sure why this is happening. My best guess is that, since adamantine is completely rigid and very light, it essentially converts the edged attack of a bolt into a blunt attack with the same force and contact area.

Summary of Arena Testing:
* Dwarves armored with breastplate, helm, greaves, 2 gauntlets, and 2 low boots of indicated material.
* Armed with adamantine crossbow, and 100 bolts of indicated material.
* Great archer/marksdwarf, grand master armor user.
* Usually 20 dwaves from each side positioned on either side of the 3 tile wide trenches in the standard arena.
* Used a perl script to count the indicated words within gamelog.txt
{| class="wikitable"
|-
! bolt material
! armor material
! hits ("flying")
! deflections ("deflect")
! through armor ("through")
! headshots ("head")
! fatal headshots ("killed")
|-
| pine
| steel
| 2011
| 1814
| 29
| 93
| 0
|-
| copper
| steel
| 856
| 0
| 772
| 38
| 26
|-
| iron
| steel
| 302
| 0
| 267
| 22
| 13
|-
| steel
| steel
| 319
| 0
| 295
| 21
| 13
|-
| pine
| adamantine
| 996
| 533
| 386
| 51
| 10
|-
| copper
| adamantine
| 708
| 0
| 649
| 36
| 26
|-
| bronze
| adamantine
| 569
| 0
| 522
| 31
| 23
|-
| iron
| adamantine
| 609
| 0
| 551
| 32
| 20
|-
| steel
| adamantine
| 663
| 0
| 594
| 36
| 17
|-
| pine
| copper
| 1390
| 1241
| 25
| 61
| 0
|-
| copper
| copper
| 579
| 0
| 522
| 26
| 11
|-
| bronze
| copper
| 552
| 0
| 510
| 25
| 18
|-
| iron
| copper
| 496
| 0
| 438
| 23
| 14
|-
| steel
| copper
| 512
| 0
| 459
| 22
| 11
|-
| pine
| pine
| 739
| 0
| 684
| 43
| 31
|-
| pine
| leather
| 740
| 0
| 695
| 37
| 27
|}
I am continuing testing, and will try to add a more complete table of results to the bolt wiki page when I am finished.--[[User:Pirate Bob|Pirate Bob]] 12:37, 15 August 2012 (UTC)

I now have a fairly large sample of results, and it doesn't look like any armor does anything to stop metal bolts. I kind of feel like going forward is a waste of time, since none of the armor actually, you know, does anything... --[[User:Pirate Bob|Pirate Bob]] 03:11, 17 August 2012 (UTC)

While it would be interesting, I unfortunately cannot test bone bolts in the arena, as they are not available. I did test wood on wood and wood on leather (both useless). I might also try testing other types (i.e. featherwood and whatever wood is densest) to get some idea of how bolt effectiveness varies with density, but doing more metal bolt tests when nothing stops them is not worth the several hours it would take.

I think instead I will redesign my test: On one side I will place a fixed number of unarmed, but armored, dwarves, and on the other side of the moat dwarves with the appropriate bolts for ammo. I will then record the average number of hits to kill the dwarf targets. It would be nice to also record the number of hits until each dwarf first falls unconscious, since this is usually what really matters in combat, but this will require a much fancier script to analyze. I am saving all of my logs, so I can always go back and re-analyze them for unconscious dwarves.

I will of course also get the same data already in my table from these tests. However, collecting data that says no armor deflects metal bolts is not interesting. I want to be able to grade the relative (in)effectiveness of each armor, and also compare survival with armor to without armor. If metal armors really do nothing to help survival compared to no armor, I would say that qualifies as a moderately serious bug, and I would proceed with trying to figure out how to modify bolt and/or crossbow raws to get more sensible behavior. --[[User:Pirate Bob|Pirate Bob]] 12:19, 17 August 2012 (UTC)

:Side note: it's "adamanti'''ne'''", not "adamanti'''um'''" - this is Dwarf Fortress, not Marvel Comics. --[[User:Quietust|Quietust]] 12:50, 15 August 2012 (UTC)

: This is great research, Bob. Please continue testing. Ideally test every combination of arrow material and every armor. try to include wooden, leather and bone armor. And Bone arrows, too, please. If I find the time, I may try to verify your results. I did not know that you have a gamelog.txt of the fights in the arena. I can write a Python script analyzing the figthts, then. --[[User:Nagidal|Nagidal]] 13:16, 15 August 2012 (UTC)

: I found that I need to slightly improve my script so that it only counts "deflect" and "through" in lines that also contain "flying *** bolt". I thought that by putting the dorfs on opposite sides of a moat so there can't be melee combat I would prevent all non-bolt hits, but I forgot to account for them dodging into the moat and falling to their deaths. Also, in case it's not obvious to anyone, the number of deflections plus the number of armor penetrations does not add up to the number of hits because some parts of the dwarves (upper arms, neck) are not armored. I could add chainmail to fix this, but I was more interested in the effectiveness of bolts against plate armor. I also was considering some of the tests with leather clothing underneath to see if this has any effect. Especially for adamantine armor. Theoretically it should help cushion the impact (real knights would always wear padding under metal armor), but I don't know if DF models that. --[[User:Pirate Bob|Pirate Bob]] 00:14, 17 August 2012 (UTC)
: I have updated my perl script to only count statistics for lines containing "flying", and I am working on filling out the table. I wrote some macros to help with arena setup. --[[User:Pirate Bob|Pirate Bob]] 01:45, 17 August 2012 (UTC)

== Decorated bolts ==

There used to be an exploit wherein encrusting bolts with gems gave them enormous value boost because decoration value is counted for every bolt in stack. As I get it, Toady fixed it by setting value multiplier for bolt decorations to 1/3 instead of 10. Should be this mentioned somewhere? Also, if you manage to get a stack of more than 30 bolts, this is still an issue - e.g., for bolts[60] decorations value is doubled. --04:17, 21 May 2012 (UTC)

== This armor is useless! ==

I did some arena testing of the number of hits required to kill unarmed dwarf "targets" In all tests, I placed 224 targets on opposite sides of the 3-wide arena moat from 242 shooters (I placed a few extra shooters by the "bridge" on the right side). I then let the arena run until all targets were dead, and counted up the total number of hits (counted the word "flying" in gamelog.txt).

{| class="wikitable"
|-
! Ammo
! Armor
! Dwarf Targets
! Hits
! Hits/kill
|-
| copper
| adamantine
| 224
| 3985
| 17.8 +/- 1.2
|-
| copper
| copper
| 224
| 3819
| 17.1 +/- 1.1
|-
| copper
| none
| 224
| 3894
| 17.4 +/- 1.2
|}

The numbers speak for themselves - armor is completely useless against even copper bolts. I will try to test this with a few more combinations, and maybe repeat a few of the tests for more statistics, but my preliminary results have pretty much confirmed my suspicion that armor does nothing against metal bolts. This just leaves the question, [http://www.youtube.com/watch?v=aqoDL4KlxvY why do we even wear it] :P ? --[[User:Pirate Bob|Pirate Bob]] 20:40, 17 August 2012 (UTC)

v0.34 Talk:Bolt

2012-08-17T20:39:48Z

Pirate Bob:

== Inconsistency? ==
"Tests show that where even masterwork wooden bolts bounce harmlessly off copper armor, iron bolts pierce without any issue."

"Tests show that even wooden bolts are effective against armored targets with the new release. Copper bolts have been known to pierce adamantine consistently."

These two statements flatly contradict each other, is either true? [[User:OrangePikmin|OrangePikmin]] 21:31, 7 April 2012 (UTC)

The problem is the assumption that adamantine is the best armor against bolts. I have done arena testing which clearly shows it is not. Steel and copper armor will both deflect 98% of pine bolts, while adamantine only deflects 58%. This is quite strange, and I am not sure why this is happening. My best guess is that, since adamantine is completely rigid and very light, it essentially converts the edged attack of a bolt into a blunt attack with the same force and contact area.

Summary of Arena Testing:
* Dwarves armored with breastplate, helm, greaves, 2 gauntlets, and 2 low boots of indicated material.
* Armed with adamantine crossbow, and 100 bolts of indicated material.
* Great archer/marksdwarf, grand master armor user.
* Usually 20 dwaves from each side positioned on either side of the 3 tile wide trenches in the standard arena.
* Used a perl script to count the indicated words within gamelog.txt
{| class="wikitable"
|-
! bolt material
! armor material
! hits ("flying")
! deflections ("deflect")
! through armor ("through")
! headshots ("head")
! fatal headshots ("killed")
|-
| pine
| steel
| 2011
| 1814
| 29
| 93
| 0
|-
| copper
| steel
| 856
| 0
| 772
| 38
| 26
|-
| iron
| steel
| 302
| 0
| 267
| 22
| 13
|-
| steel
| steel
| 319
| 0
| 295
| 21
| 13
|-
| pine
| adamantine
| 996
| 533
| 386
| 51
| 10
|-
| copper
| adamantine
| 708
| 0
| 649
| 36
| 26
|-
| bronze
| adamantine
| 569
| 0
| 522
| 31
| 23
|-
| iron
| adamantine
| 609
| 0
| 551
| 32
| 20
|-
| steel
| adamantine
| 663
| 0
| 594
| 36
| 17
|-
| pine
| copper
| 1390
| 1241
| 25
| 61
| 0
|-
| copper
| copper
| 579
| 0
| 522
| 26
| 11
|-
| bronze
| copper
| 552
| 0
| 510
| 25
| 18
|-
| iron
| copper
| 496
| 0
| 438
| 23
| 14
|-
| steel
| copper
| 512
| 0
| 459
| 22
| 11
|-
| pine
| pine
| 739
| 0
| 684
| 43
| 31
|-
| pine
| leather
| 740
| 0
| 695
| 37
| 27
|}
I am continuing testing, and will try to add a more complete table of results to the bolt wiki page when I am finished.--[[User:Pirate Bob|Pirate Bob]] 12:37, 15 August 2012 (UTC)

I now have a fairly large sample of results, and it doesn't look like any armor does anything to stop metal bolts. I kind of feel like going forward is a waste of time, since none of the armor actually, you know, does anything... --[[User:Pirate Bob|Pirate Bob]] 03:11, 17 August 2012 (UTC)

While it would be interesting, I unfortunately cannot test bone bolts in the arena, as they are not available. I did test wood on wood and wood on leather (both useless). I might also try testing other types (i.e. featherwood and whatever wood is densest) to get some idea of how bolt effectiveness varies with density, but doing more metal bolt tests when nothing stops them is not worth the several hours it would take.

I think instead I will redesign my test: On one side I will place a fixed number of unarmed, but armored, dwarves, and on the other side of the moat dwarves with the appropriate bolts for ammo. I will then record the average number of hits to kill the dwarf targets. It would be nice to also record the number of hits until each dwarf first falls unconscious, since this is usually what really matters in combat, but this will require a much fancier script to analyze. I am saving all of my logs, so I can always go back and re-analyze them for unconscious dwarves.

I will of course also get the same data already in my table from these tests. However, collecting data that says no armor deflects metal bolts is not interesting. I want to be able to grade the relative (in)effectiveness of each armor, and also compare survival with armor to without armor. If metal armors really do nothing to help survival compared to no armor, I would say that qualifies as a moderately serious bug, and I would proceed with trying to figure out how to modify bolt and/or crossbow raws to get more sensible behavior. --[[User:Pirate Bob|Pirate Bob]] 12:19, 17 August 2012 (UTC)

:Side note: it's "adamanti'''ne'''", not "adamanti'''um'''" - this is Dwarf Fortress, not Marvel Comics. --[[User:Quietust|Quietust]] 12:50, 15 August 2012 (UTC)

: This is great research, Bob. Please continue testing. Ideally test every combination of arrow material and every armor. try to include wooden, leather and bone armor. And Bone arrows, too, please. If I find the time, I may try to verify your results. I did not know that you have a gamelog.txt of the fights in the arena. I can write a Python script analyzing the figthts, then. --[[User:Nagidal|Nagidal]] 13:16, 15 August 2012 (UTC)

: I found that I need to slightly improve my script so that it only counts "deflect" and "through" in lines that also contain "flying *** bolt". I thought that by putting the dorfs on opposite sides of a moat so there can't be melee combat I would prevent all non-bolt hits, but I forgot to account for them dodging into the moat and falling to their deaths. Also, in case it's not obvious to anyone, the number of deflections plus the number of armor penetrations does not add up to the number of hits because some parts of the dwarves (upper arms, neck) are not armored. I could add chainmail to fix this, but I was more interested in the effectiveness of bolts against plate armor. I also was considering some of the tests with leather clothing underneath to see if this has any effect. Especially for adamantine armor. Theoretically it should help cushion the impact (real knights would always wear padding under metal armor), but I don't know if DF models that. --[[User:Pirate Bob|Pirate Bob]] 00:14, 17 August 2012 (UTC)
: I have updated my perl script to only count statistics for lines containing "flying", and I am working on filling out the table. I wrote some macros to help with arena setup. --[[User:Pirate Bob|Pirate Bob]] 01:45, 17 August 2012 (UTC)

== Decorated bolts ==

There used to be an exploit wherein encrusting bolts with gems gave them enormous value boost because decoration value is counted for every bolt in stack. As I get it, Toady fixed it by setting value multiplier for bolt decorations to 1/3 instead of 10. Should be this mentioned somewhere? Also, if you manage to get a stack of more than 30 bolts, this is still an issue - e.g., for bolts[60] decorations value is doubled. --04:17, 21 May 2012 (UTC)

== This armor is useless! ==

I did some arena testing of the number of hits required to kill unarmed dwarf "targets" In all tests, I placed 224 targets on opposite sides of the 3-wide arena moat from 242 shooters (I placed a few extra shooters by the "bridge" on the right side). I then let the arena run until all targets were dead, and counted up the total number of hits (counted the word "flying" in gamelog.txt).

{| class="wikitable"
|-
! Ammo
! Armor
! Dwarf Targets
! Hits
! Hits/kill
|-
| copper
| adamantine
| 224
| 3985
| 17.8 +/- 1.2
|-
| copper
| copper
| 224
| 3819
| 17.1 +/- 1.1
|-
| copper
| none
| 224
| 3894
| 17.4 +/- 1.2
|}

The numbers speak for themselves - armor is completely useless against even copper bolts. I will try to test this with a few more combinations, and maybe repeat a few of the tests for more statistics, but my preliminary results have pretty much confirmed my suspicion that armor does nothing against metal bolts. This just leaves the question, [http://www.youtube.com/watch?v=aqoDL4KlxvY why do we even wear it] :P ?

v0.34 Talk:Bolt

2012-08-17T12:41:12Z

Pirate Bob: /* Inconsistency? */

v0.34 Talk:Bolt

2012-08-17T12:19:42Z

Pirate Bob: /* Inconsistency? */

v0.34 Talk:Bolt

2012-08-17T03:11:24Z

Pirate Bob: /* Inconsistency? */

v0.34 Talk:Bolt

2012-08-17T03:00:01Z

Pirate Bob: /* Inconsistency? */

v0.34 Talk:Bolt

2012-08-17T01:45:01Z

Pirate Bob: /* Inconsistency? */

v0.34 Talk:Bolt

2012-08-17T00:14:16Z

Pirate Bob: /* Inconsistency? */

v0.34:Bolt

2012-08-15T12:43:22Z

Pirate Bob: /* Bugs */

{{quality|Fine|11:53, 26 December 2010 (UTC)}}{{buggy}}{{av}}

Bolts are a type of [[ammunition]] used by [[crossbow]]s. They can be made from either [[wood]], [[bone]], or weapons-grade [[metal]]s. As with all non-siege ammo, hunting and military dwarves must be equipped with [[quiver]]s to be able to carry bolts around for shooting.

Wooden and bone bolts are produced in a [[Craftsdwarf's workshop]] by a [[Wood crafter|Wood Crafter]] or a [[Bone carver|Bone Carver]], respectively. Each job will produce a stack of 25 wooden bolts from a single log, or a stack of 5 bone bolts taking a single bone from a stack. Metal bolts require [[Weaponsmith|Weapon''smithing'']] (not Metal''crafting'') and a [[Metalsmith's forge]] or [[Magma forge]], creating a stack of 25 bolts from a single metal bar.

When fired, bolts have a chance of shattering on impact or bouncing off onto the ground as free-laying items. However, they are immediately [[forbidden]] after being fired (to prevent <s>suicidal</s> hauling dwarves from running into active battlefields), and have to be [[reclaim]]ed by the player, either directly or from the [[stocks]] screen.

Given that bolts are a projectile weapon, their weight influences their effectiveness (higher weight means higher inertia, making them harder to deflect). Tests show that where even masterwork wooden bolts bounce harmlessly off copper armor, iron bolts pierce without any issue. Bone bolts are significantly better than wood, but metal bolts are still recommended for military use. However, this would make [[adamantine]] a very bad choice for bolt making due to its ridiculously low density (see [[Weapon#Material|weapon materials]]).

==Masterwork Bolts==
As with other items, high-skill dwarves can make masterwork-[[Item quality|quality]] bolts. While they don't care if the bolts are fired and shatter on impact, they do get [[thought|upset]] if a bolt gets carried off the map - say by a retreating enemy with the bolt stuck in him.

The simplest remedy for this is to simply '''make more bolts''' - a dwarf who has produced hundreds of masterwork bolts will be more annoyed by [[fly|flies]] than the loss of a masterwork. This is because the happiness penalty for the loss of a masterwork is divided by the number of remaining masterwork items made by that dwarf (see [[List of Dwarven Thoughts]] for more information).

==Bugs==
Adamantine armor seems to be less effective than even copper against bolts in this release. Copper and steel armor both deflect 98% of pine bolts, while adamantium only deflects 58%. Metal bolts appear not to ever be deflected by any armor, but testing is ongoing. {{bug|5516}}

Bone bolts are not advisable for weapon traps as only a stack of 5 will be stored per crossbow, making frequent restocking necessary.

It is largely unclear what of the above constitutes a bug and what is intended behaviour.

*If squad is assigned multiple ammo types, dwarves with "individual choice ranged" carry wrong ammo{{bug|1374}}.
*Masterwork bolts cause unhappy thoughts when they get carried off the map or melted down{{bug|3169}}.

{{gamedata|[ITEM_AMMO:ITEM_AMMO_BOLTS]
[NAME:bolt:bolts]
[CLASS:BOLT]
[SIZE:150]
[ATTACK:EDGE:2:2000:stab:stabs:NO_SUB:1000]
}}

v0.34 Talk:Bolt

2012-08-15T12:38:29Z

Pirate Bob: /* Inconsistency? */

v0.34 Talk:Bolt

2012-08-15T12:37:22Z

Pirate Bob: /* Inconsistency? */