Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[G-Flex]

To be fair, if you cut out the bits where I'm just going through tissues and poking at them to see if they make sense, it's... almost reasonable!

[Pandarsenic]

I think the whole thing is a very good, very helpful analysis of some of DF2010's somewhat problematic mechanics.

[Arrkhal]

The main things I noticed on my trips through the raws:

Every single metal in vanilla has an "impact strength" that was noted as being an average for stainless steel.  Gold, copper, iron, you name it.  And impact strength is compared between materials when a blunt weapon encounters some armor, so gold was considered to be just as strong as steel, leading to craziness.

Units were inconsistent, with some kind of 10fold error, or unit conversion error, or something, making impact strength even crazier.

Hardness isn't modeled, and maxedge doesn't seem to do much.  Real life wood has a lower shear strength than real life epidermis, but modeling that in game means that wooden weapons stop piercing skin unless you either buff wood or nerf skin.

Composite weapons and items still aren't in, so wooden arrows and bolts weigh less than 1/15 as much as iron ones, making them basically do nothing.

Impact elasticity has a fixed upper limit of 49,999 for stuff to behave plastically.  Right at 50,000 and up, it becomes impossible to ever tear something due to blunt trauma.  I'm still seeing if I can make muscles and organs tear and rupture when hit by a powerful blow, but it's not looking good.

[Wyrm]

I noticed that too. Impact strength should be lower than the compressive strength for all materials, not the same and certainly not higher. Impact is not the same as compression, and any realistic modeling of impact vs. compression has to take into account the timescale of the interaction, and I don't see this anywhere. The compressive strength of a material is partially due to the ability of the material to transmit stresses down the line to the end as well as across, but an impact is defined as a shock whose timescale is too short for that to happen.

Hardness isn't modeled, and maxedge doesn't seem to do much.  Real life wood has a lower shear strength than real life epidermis, but modeling that in game means that wooden weapons stop piercing skin unless you either buff wood or nerf skin.

Except with an arrow, you're pitting the compressive strength of the arrow against the shear strength of skin, so comparing shear to shear is not fair to the arrow. Most wood is stronger in compressive stress than skin is in shear, if only just barely. This explains why wooden arrows (no tip) can be used to hunt game, but are not very good at actually penetrating to do deep damage.

I must point out that materials science (including fracture, impact, and all around destruction) is a very complex one, and I wouldn't want to tackle it with anything less than a good set of materials tables, a supercomputer, and a masters in materials science. Fudging is probably going to be inevitable.

[Narmio]

Yeah, currently all materials have blunt combat attributes (impact yield and impact fracture) equal to some placeholder value 1080000 which appears to be chosen from stainless steel, according to a raws note by Toady.  I think the shear values (presumably for edged combat) might have similar placeholders.  We really can't make any comment about the formulas and processes Toady is using to do the wound calculations until we can better understand what's being caused by the placeholders.

Massively reducing impact fracture but not impact yield for armour, for example, creates a system where almost every hit from a weapon causes serious deep bruising, but never penetrates enough to break bones.  So there's modelling of some more complicated stuff than what we're seeing in fights right now (better material weapon goes through all the time, equal or lower material weapon bounces off all the time) going on.  I just think it's obscured.

We need to see if we can scare up a materials scientist from among the community, or an engineering student willing to do some library-diving to make some sense of these numbers and what they *should* be.

[G-Flex]

I'm still trying to find out why blunt weapons can "tear" through fat but not muscle. I cannot figure this out for the life of me.

[Footkerchief]

Just thickness, maybe?

[PencilinHand]

Yeah, currently all materials have blunt combat attributes (impact yield and impact fracture) equal to some placeholder value 1080000 which appears to be chosen from stainless steel, according to a raws note by Toady.  I think the shear values (presumably for edged combat) might have similar placeholders.  We really can't make any comment about the formulas and processes Toady is using to do the wound calculations until we can better understand what's being caused by the placeholders.

Massively reducing impact fracture but not impact yield for armour, for example, creates a system where almost every hit from a weapon causes serious deep bruising, but never penetrates enough to break bones.  So there's modelling of some more complicated stuff than what we're seeing in fights right now (better material weapon goes through all the time, equal or lower material weapon bounces off all the time) going on.  I just think it's obscured.

We need to see if we can scare up a materials scientist from among the community, or an engineering student willing to do some library-diving to make some sense of these numbers and what they *should* be.

My ears are burning....

I have an ~900 page fundamentals of material science and engineering book, a bachelors in mechanical engineering, and am unemployed(difficult job market is difficult, but it is getting better)....

What do you want to know?  The uses of yield vs fracture?  Frankly, until weapon and armor damage are in the game the concepts of material yield and fracture are largely moot.  How the concept of fatigue is implemented(if at all) will be a critical part of the weapon/armor damage.  Besides we would really need to look either at the code or at least get an oral explanation of the under-pinnings and inner-workings from Toady.  How I might expect penetration to work and how Toady may have coded it could be vastly different.

Spoiler: The methods for calculating weight are a little arbitrary. (click to show/hide)

item_armor.txt = (layer size/ material size )*material density/1010 ?? maybe a little higher
item_helm.txt = (layer size * material density)/60000 {Exception: Helm = material density / 1010 ??}
item_gloves.txt = density/100000 ??
item_pants.txt = (layer size * material density)/5500
item_shoes.txt = (layer size * material density)/75000 {Exception: High Boot = material density / 100000 ??}
item_shield.txt = density/100000 ??

 

If you want useful numbers for different materials, I can tell you that the bulk modules to COMPRESSIVE_ELASTICITY conversion isn't linear(seems to be something close to y = constant/x but I haven't bothered too do anything other than a quick plot of the data).  I can also confirm that some of the numbers(chiefly the compressive yield and fracture numbers) are bogus at best, and in fact hasn't changed sense Toady originally put the raws together ~16 months ago.

I'm still trying to find out why blunt weapons can "tear" through fat but not muscle. I cannot figure this out for the life of me.

Just thickness, maybe?

Probably, fat has the same relative thickness as skin apparently and skin is pretty regularly smashed open by warhammers(like on fingers and toes).  Although, how [THICKENS_ON_ENERGY_STORAGE] and [THICKENS_ON_STRENGTH] interact with the combat system, if at all, remains to be seen.  I think I did hear/read Toady answer a question kind of to that effect, though....

---
Frankly, when I first went looking in the armor raws originally I was expecting explicit thickness statements and was surprised to see the old [LAYER_SIZE:#] and [LAYER_PERMIT:#] tags instead.  I am still expecting the layer size to have some relevance to deflection and protection but have held off testing.  As I don't have access to a working copy of MATLAB anymore I have been writing a program to parse the gamelog for me and turn it into a more excel friendly format but have been having difficulty making my mind up on just how to go about doing that(I have maybe 40% of it done).  I am looking for a job you know. 

[G-Flex]

I believe Toady has explicitly stated that extra fat provides a bit of damage absorption in combat, so yeah, that stuff does come into play, although who knows to what degree.

Also, I invariably read your name as "Penicillinhand".

[G-Flex]

Since 0.31.06 changed some relevant creature raw values, I decided to take another look at creature-on-creature bitin'-and-fightin' combat.


Results: Better than before, but still not exceptional.

Cat vs. Cat
The cats scratched at each other a hell of a lot, sometimes not (but usually) penetrating the skin, and never penetrating more. Not too terrible. However, their bites didn't seem to do much of anything at all, except for when one decided to shake the other around. Bites always seemed to only dent the skin or glance away, even against small parts.

Cat vs. Dwarf

It crashed when I tried to spawn the dwarf.

Cat vs. Dwarf (Take Two)

The cat got in mostly-unsuccessful scratching and biting attacks (biting attacks did not penetrate, scratches rarely broke skin), while the dwarf punched the cat a lot (as well as pointless grab-limb-and-release wrestling), bruising muscle and fat and organs, but never causing much damage due to the issues with cumulative injury I've pointed out before. After quite a long time, the dwarf won, because he went Blend-Tec on the cat's eyes, gouging them out a few dozen times. The cat bled out eventually (the bruised heart might not have helped).

Alligator vs. Alligator

Bites penetrate now, but still only rarely penetrate past the scale. The "shake around" move is still devastating and deadly, of course, although it's unclear what the effects are, exactly. There's not much more to say here.

Alligator vs. Cat

Cat scratches somehow manage to tear alligator scale. Also, despite alligator bites penetrating alligator scale, they cannot seem to penetrate cat skin at all. Latching on and shaking around dispatched the cat quite quickly, though.

Alligator vs. Dwarf

Alligator bites tear the skin, and dwarf punches glance away. The alligator then shook around and tore off the dwarf's lower body. Pretty typical of a real alligator, I guess, aside from the bites not penetrating more.

Alligator vs. Dwarf in Pigtail Clothing (trousers, shirt, cap, gloves, boots, coat)

The alligator knocks the dwarf around like crazy with his biting maneuvers, but never manages to tear skin except for once on the head (presumably because the cap doesn't provide full coverage). All other attacks are deflected by the clothing, probably because alligator teeth don't punch through deeply enough. Eventually the alligator wins, by getting in a lucky shot to the head again and tearing it off by shaking it around.

Elephant vs. Donkey

Elephant kicks knock the donkey around decently, but still don't do very much damage. You'd think they would break bone or something, but they're just bruising fat and occasionally muscle. Incidentally, the donkey's kicks are doing roughly the same thing to the elephant. One donkey kick managed to tear apart the cartilage in the elephant's trunk, which is pretty weird. They bruised each other's guts a bit as well. Also, the elephant grabbed the donkey's lower body with his left rear leg and threw the donkey about five squares. So far, the elephant is the only animal I've seen doing throwing moves, and it did them several times here. The donkey managed to heal completely during the course of the battle, due to minor bruising being the only injuries he had, and seems to have the upper hand in general.

I went downstairs to do some laundry, thinking it was a stalemate. When I got back, I noticed that the elephant had placed a stranglehold on the donkey's throat using its leg, and eventually suffocated the thing. Well, I guess the elephant won.

I also noticed that approximately twenty-nine cats (see below) assaulted the elephant from elsewhere. The elephant was too over-exerted at this point, only getting in a single attack (kicking a cat's head in, shattering the skull but not killing it) before passing out. It seems like nobody's winning, though, since even when the cats shake the elephant around it doesn't do any mortal damage (or significant enough bleeding, it seems). His eyes are torn open, ears and trunk broken and torn apart, and in general he isn't having too much fun. Semper fidelis, mighty elephant.

Dwarf vs. Twenty-Nine Cats

What you'd expect from above tests. The dwarf eventually passes out from exhaustion while cats claw at him, often tearing the skin. Apparently, when given adequate opportunity, cats can shake dwarves around. This eventually caused the dwarf to bleed to death.

Anything vs. Magma

Still just as absurd. Interestingly, while live alligators will catch fire, dead ones will not. Their corpses instead rot faster (although this might have just been a coincidence) and take damage, the blood also boiling off, but only after quite a while. This dead alligator is now an "XXAlligator CorpseXX" but not on fire yet. A few turns later, it disappeared entirely, presumably due to the massive damage.



Me vs. This Data

It seems that natural attacks still don't penetrate enough. This would make sense if you consider it as a simple matter of "How long are these teeth?" vs. "How thick is this tissue?" but it's not that simple in reality. In reality, pushing down like that on tissue causes it to spread apart, compress, dent, and so forth. Even without considering that (although it should be considered!), things should penetrate deeper. This goes hand-in-hand (sorta?) with the penetration issues with things like blowguns and (to a lesser degree) bolts/arrows. After all, the fact that clothing and tissues can never really be torn through is causing issues here, providing near-perfect protection against bite wounds, even without any armor whatsoever.

For that matter, why does scale (at least on alligators) less effective than cat/dwarf skin? You'd think the opposite would be true. I'm not sure what's going on there.

Issues with damage not being terribly cumulative are still paramount here in damn near all the tests, but I've talked about that enough before (I tend to repeat myself enough already).

Elephant kicks not doing much is quite odd. It seems to me that creature size in general isn't doing enough, actually. A more massive creature should put more weight into a striking/wrestling move. They should also successfully absorb more force from strikes, which the elephant didn't seem to be doing very well, if at all.

I don't even have a clue what's up with the elephant wrestling the donkey, as amusing as it was. I thought wrestling for animals was disabled, but maybe I was wrong? However, I saw no other animals doing it at all. Wacky!


Magma/heat damage is still wacky, and I have some pretty serious issues, medically-speaking, with some of the raw values for animal tissue, but that'll have to wait until I talk to Djohaal/Kefka on IRC about them more (he's a med student) so I can compile a nice, well-reasoned list.



[Edit]
I almost forgot this:

Cows vs. Dogs vs. Butchery

For some reason, cows produce less fat than dogs, and slightly less meat as well. I have no freaking clue why. This perplexes me.


[Edit (again)]

From the raws, it seems like the hair template has better shear strength than skin does. Did Toady mess up and use the values for pure keratin? Is a layer of hair causing cats and other furry critters to be protected too much? Is the hair treated as a solid layer of the material instead of a stringy, fuzzy mess? I sure as hell don't know!

[Footkerchief]

Anything vs. Magma

Still just as absurd. Interestingly, while live alligators will catch fire, dead ones will not. Their corpses instead rot faster (although this might have just been a coincidence) and take damage, the blood also boiling off, but only after quite a while. This dead alligator is now an "XXAlligator CorpseXX" but not on fire yet. A few turns later, it disappeared entirely, presumably due to the massive damage.

IIRC, it's hard to get ANYTHING to ignite in the new version.  Most flammable materials seem to just evaporate from heat damage instead of catching fire.

From the raws, it seems like the hair template has better shear strength than skin does. Did Toady mess up and use the values for pure keratin? Is a layer of hair causing cats and other furry critters to be protected too much? Is the hair treated as a solid layer of the material instead of a stringy, fuzzy mess? I sure as hell don't know!

I believe that the different TISSUE_LAYER_SHAPE is meant to limit the effectiveness of hair/fur, but who knows whether that's working.

[G-Flex]

Anything vs. Magma

Still just as absurd. Interestingly, while live alligators will catch fire, dead ones will not. Their corpses instead rot faster (although this might have just been a coincidence) and take damage, the blood also boiling off, but only after quite a while. This dead alligator is now an "XXAlligator CorpseXX" but not on fire yet. A few turns later, it disappeared entirely, presumably due to the massive damage.

IIRC, it's hard to get ANYTHING to ignite in the new version.  Most flammable materials seem to just evaporate from heat damage instead of catching fire.

Yeah, the unusual part here is that the creature's tissue was fine about igniting while the thing was still alive, but not on the ground. I guess there's some issue with how temperature transfer to items is done compared to creature tissues, but the fact that the corpse got damaged means temperature was transferring, but perhaps transferring away (to the air? I'm not sure how DF handles this) almost as fast.

From the raws, it seems like the hair template has better shear strength than skin does. Did Toady mess up and use the values for pure keratin? Is a layer of hair causing cats and other furry critters to be protected too much? Is the hair treated as a solid layer of the material instead of a stringy, fuzzy mess? I sure as hell don't know!

I believe that the different TISSUE_LAYER_SHAPE is meant to limit the effectiveness of hair/fur, but who knows whether that's working.

Feh, I missed that. Thanks. I guess hair's not really an issue, then.

Actually, it might be! Think about this: That only covers the tissue. If an item, hypothetically, gets made out of hair, it would function like solid hair-leather. This doesn't happen right now, but consider, say, wool, or what implications this might have for other things. The basic weirdness here is that the composition of the material is handled totally differently for armor, for tissues, and for other things, since it's all kind of cludged about, primarily so special cases can be handled. Ideally, you should be able to say "This [THING] is made out of [STRANDS] of [HAIR]" no matter the identity of [THING], instead of having completely different ways to say "[STRANDS]" for each type of [THING] (or no way at all, in some cases). I'd say that the material system really should have another tier to its hierarchy, so that it's MATERIAL -> FORM -> ITEM instead of simply MATERIAL -> ITEM with some FORM cludges in place for certain types of FORMs/ITEMs. This way, you could have, say, armor made out of scaled/barred/woven/linked material, or material in any other arbitrary form, simply depending on where the material came from, in a natural manner, without it having to be handled specially.


Interestingly enough, the scales tissue template has TISSUE_LAYER_SHAPE:SCALES and not TISSUE_LAYER_SHAPE:LAYER. I have no idea how this affects things.


For fun, I changed hair to have TISSUE_SHAPE:LAYER and had two cats fight each other, resulting in stuff like this:

Cat 2 scratches Cat 1 in the upper body, tearing the hair!
Cat 2 bites Cat 1 in the left rear paw, denting the hair!

Amusing.




A note regarding weight encumbrance due to equipment
We know that encumbrance due to hauled items works in DF (I've seen a dwarf haul around a slade stone, and it takes... a while), and I know from previous tests that DF doesn't take weight of a weapon into account (or at least not enough) when determining swing/attack speed (a dwarf with a slade weapon weighing as much as a small vehicle probably shouldn't be able to lift it; dwarves with heavy weapons should attack more slowly, clumsily, and with less velocity, but this doesn't appear to be the case).

What I didn't know is that encumbrance in general appears to be sort of... not there, at least when it comes to worn equipment.

I spawned an elephant wearing a slade mail shirt. The shirt weighed 33,278 kg (that's about 33 metric tons), but the elephant danced around his opponent and ran around just fine. I also spawned two dwarves on opposing teams, one wearing nothing and the other wearing a full suit of slade armor (about 1982 kg or, in other words, a hell of a lot more than the car I drive), and saw them rush each other. They seemed to move at virtually the same pace, even though a dwarf hauling a slade stone does so at a reasonably slow pace.

[Greiger]

In my own experiments in arena with weight I noticed that it can eventually have an effect.  I can't say I did quite as professional testing as you did, but in arena I weilded a lion corpse, and elephant corpse and a whale corpse and beat up unarmed elves with the weapons.  it was with the same creature weilding the corpses at least. 

the lion corpse wasn't very effective, but it could break a bone now and then, which surpassed my expectations.  Elf died when a chunk of skull lodged in the brain.

The elephant corpse was remarkably effective.  I would say very nearly as effective as an actual iron maul.  It regularly broke bones, almost every swing.  I don't quite recall how that elf died though, think it bled to death or something from a compound fracture messing up an artery.

The whale corpse however was absolutely worthless.  I assume it was too heavy.  The most damaging strike was a dent to the fat.  The elf eventually strangled me to death.

[G-Flex]

Keep in mind that a whale corpse weighs a lot more than any weapon, though.

I just spawned four whales and let them air-drown, and they weighed between 12,400 and 35,000 (quite a range there) kilograms. I also spawned some elephants, getting weights between 4,400 and 7,500 kilograms.

For comparison, even a slade maul is "only" 260 kilograms.


I'm glad you brought your experience up, because if you're right, it means that weapon weight is taken into consideration in this manner, but only at truly absurd weights.

Of course, it might still be true that you were still able to attack quickly with the whale corpse, or that it didn't slow down your movement, or that things might play out differently for armor.

[Footkerchief]

I'm glad you brought your experience up, because if you're right, it means that weapon weight is taken into consideration in this manner, but only at truly absurd weights.

I suspect he balanced that part of the calculations before the weight units got changed.  Similar oversights would explain related bugs like armor items weighing ridiculously little and the problem you mentioned above with heavy armor not burdening its wearers.