Print

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

[Pirate Bob]

The !!SCIENCE!! in this thread suggests that the primary advantage of adamantine projectiles over steel is that adamantine has higher [SHEAR_YIELD] and [SHEAR_FRACTURE].  This means that edged attacks by adamantine bolts/arrows can't be stopped at all by any armor which doesn't have equally high [SHEAR_YIELD] or [SHEAR_FRACTURE] (meaning more adamantine, and possibly slade).

Now, if we assume that melee edged attacks behave roughly the same (which circumstantial evidence supports, but is by no means proven) this raises the question of whether its worth making adamantine weapons in fort mode.  Nothing that dwarves will fight (other than berserk dwarves or perhaps a steel titan) will have steel armor, meaning that steel edged weapons should go right through all goblin, human, and certainly elven armor.  I've used a single axelord wielding a steel axe to defeat entire goblin sieges, which seems to support this.  My logs from my fort showed no cases where a hack with a steel axe was deflected.  It is possible that attacks are "slowed down" without being deflected, but this definitely does not occur for steel bolts fired at non-steel armor, which result in death just as quickly as against targets wearing no armor at all. 

If it is really true that steel cuts through non-steel armor just as well as adamantine, than it might actually be *better* than adamantine for melee weapons, as it is more massive and might have more momentum (depending on how the velocity of a melee attack is calculated).  We definitely found that having more massive bolts caused more damage, but that is partly due to some funny rounding in calculations of bolt velocity. 

Anyway, I am curious if anyone has any insights into this.  I am considering setting up some arena tests of 1 vs. 1 dwarves wearing iron armor, armed with either steel or adamantine axes.  If my conjecture is correct, the result should either be a draw or favor the steel axes.  Then my next question would be what a good test of something "hard" that dwarves might face in fortress mode would be?   I could put dwarves against dragons of course, or possibly clowns (may not be available in the arena?).  I think the problems that dwarves have scoring hits on large targets are more issues with penetration depth than weapon sharpness though.

[Oaktree]

I know that dropping items on things is different than melee strikes or bolt strikes, but can metal vs metal comparisons be done by dropping metal weapons on a target and testing for deflection/penetration that way as well?

[BoredVirulence]

If I were testing it, I would probably try making arena tests with a small group (or a single soldier) of candy / steel clad warriors, vs a large group of standard, poorly armed opponents. If a single candy warrior(whether that be candy weapon, whatever armor, or all candy) kills 150 opponents on average before dying, and a steel warrior killed 120, I think we would have an answer.

Not sure if I would issue candy / steel a standard set of armor to provide the same amount of protection, or let them use their prospective armor types. Probably do both if time allows.

Honestly no idea how this will turn out. On the one hand, I've followed you're research with bolts, and it seems as if you make a valid point. On the other hand I've never used either steel or candy (bronze is more dwarfy / more fun).

[Pirate Bob]

If I were testing it, I would probably try making arena tests with a small group (or a single soldier) of candy / steel clad warriors, vs a large group of standard, poorly armed opponents. If a single candy warrior(whether that be candy weapon, whatever armor, or all candy) kills 150 opponents on average before dying, and a steel warrior killed 120, I think we would have an answer.

That is something I had considered as well - pit a single dwarf against a horde of goblins armored like typical goblins and see how he does.  This would be significantly more difficult than 1 vs. 1 dwarf tests, as I would need to run it quite a few times to get good statistics (whereas I could just run a whole bunch of 1 vs. 1 at the same time).  If I were to clad the dwarf in full candy or even steel, then a single dwarf with high skills might be able to kill an unlimited number of goblins, so it might be more enlightening to measure how many hits it takes him to kill all the goblins.  I'm not sure, but I think results from a 1 vs. 1 test would be clearer, so I'm still leaning towards doing that, at least first.  I definitely am curious to see just how many goblins an axelord can take...

[BoredVirulence]

If I were testing it, I would probably try making arena tests with a small group (or a single soldier) of candy / steel clad warriors, vs a large group of standard, poorly armed opponents. If a single candy warrior(whether that be candy weapon, whatever armor, or all candy) kills 150 opponents on average before dying, and a steel warrior killed 120, I think we would have an answer.

If I were to clad the dwarf in full candy or even steel, then a single dwarf with high skills might be able to kill an unlimited number of goblins,

This is a good point, but it should be able to be countered by their armor. Poor armored dwarves, even with a candy battle axe will eventually die.

Sample size is a bit tricky, but you should be able to get 30 tests of each with a macro, right?

Number of hits might be a more useful statistic, I figured if you were doing say 1000 tests of 1 vs 1, candy vs goblin and another 1000 of steel vs goblin, we would end up with 2000 dead goblins, and no progress would be made.

[Thormgrim]

i'm interested to see how this works out.

[Oaktree]

Might want to use dwarves or some race that you can armor better than goblins.  Goblins tend to skip using greaves, gauntlets, and boots.  Therefore a lot of the hits they take in "normal" combat can often hit spots with no metal armor.  Leading to even iron- or copper- armed axe dwarves being effective since they can chop limbs off through cloth and leather.  (Though I did once get a goblin civ that used chain shirts rather than breastplates and noted that the tendency for limb removal shifted towards lower limbs with fewer upper limb amputations.  I figured that was due to the copper or iron mail giving the goblins better coverage.)

[Zaroua]

Problem with melee weapons is that you need to take into consideration demons and forgotten beasts and other large critters. You'd need to run some sort of tests on extremely large targets to compare the ability of steel vs adamantine when it comes to severing extremely large limbs and cutting/piercing through the body to hit internal organs.

In other words, you'd need to find out if a dwarf using an adamantine battle axe is more likely to cut an elephant's leg off than a dwarf using a steel battle axe. Or piercing the heart of some random FB with a steel vs adamantine spear. You'd also probably want to run tests with varying levels of strength for your dwarves to see how much of an effect that has.

The reasoning behind the need for this test is that once you get steel weapons in vanilla, goblin/human armor stops mattering and you pretty much just cut them up as if they were naked - which is fairly obvious with a steel clad axelord, as you said. So the only thing that matters after that are very large creatures.

[Urist Da Vinci]

Try lowering MAX_EDGE of all metals to an (equal) low number, and then see if there is a difference between adamantine and steel when it comes to cutting up weaker-than-steel materials.

[Pirate Bob]

I ran a round of tests with 2304 pairs of dwarves locked in 1x3 cells.  Both dwarves had full iron armor (breastplate, chain mail, helm, greaves, gauntlets, and low boots).  One dwarf was given a steel axe, and another an adamantine axe.  1275 of the dwarves with steel axes were killed, compared to 1084 with adamantine axes, indicating that adamantine is in fact better.  Sometimes both dwarves died, so it is not a mistake that the numbers don't add to 2304.

I looked through the logs, and found that the steel battle axes were sometimes deflected by iron greaves.  I haven't run a systematic analysis yet, so there may be other contributors.  If it is in fact just the greaves, this still leaves open the question if adamantine is beneficial against goblins, and I don't believe they usually (if ever) wear greaves.  I have no idea why greaves would result in deflections, as they are the worst at deflecting bolts, but maybe it has something to do with them being large and axes having a large contact area.

Edit:  A few other details - all armor and weapons were standard quality.  Axes were modded to remove all attacks besides hacking.  I used standard unmodded dwarves (not arena dwarves), with great fighter/axedwarf and proficient armor user. 

[MrWillsauce]

Posting to learn

[Oaktree]

One place steel will probably be better than adamantine in terms of axes and spears are the alternate attack styles that a blunt attacks; pommel strikes with swords, blade flat with axes, and shaft strikes with spears.  In those cases the higher weight of the steel weapon will be superior.

[squishynoob]

(...)
I looked through the logs, and found that the steel battle axes were sometimes deflected by iron greaves.  I haven't run a systematic analysis yet, so there may be other contributors.
(...)

It could be because when you're hit in the lower body, mail shirts, breastplate and greaves all overlap. So an attack could be considered deflected by greaves (outermost layer) because it didn't punch through all layers of armor.

Were greaves deflections all from the lower body or also other bodyparts?

[Flare]

If I were testing it, I would probably try making arena tests with a small group (or a single soldier) of candy / steel clad warriors, vs a large group of standard, poorly armed opponents. If a single candy warrior(whether that be candy weapon, whatever armor, or all candy) kills 150 opponents on average before dying, and a steel warrior killed 120, I think we would have an answer.

There are a few issue with this, namely that there isn't much control over the experiment at all. Putting 1v1 battles is a way to limit the stuff that you can't account for, like criticals or a succession of unlucky or lucky instances, and there are a lot when you have a messy cluster fuck of a mob battle. Unless you test these this configuration a hundred times, issues like this will cloud the test result even if the final conclusion is somewhat accurate.

In any case, this is measuring one aspect of the metals only, namely if the metal in question was turned into a type of weapon, how many poorly armed opponents can it kill. It'd probably be even worse if you don't used standardized uniforms for their opponents. What it says about its effectiveness against other types of metal is left unsaid, or at least you cannot conclude anything of that sort from the data you've created.

[Pirate Bob]

One place steel will probably be better than adamantine in terms of axes and spears are the alternate attack styles that a blunt attacks; pommel strikes with swords, blade flat with axes, and shaft strikes with spears.  In those cases the higher weight of the steel weapon will be superior.

I forgot to mention that I eliminated all the alternate attacks for axes, so that they can only hack.  Good thought though.

(...)
I looked through the logs, and found that the steel battle axes were sometimes deflected by iron greaves.  I haven't run a systematic analysis yet, so there may be other contributors.
(...)

It could be because when you're hit in the lower body, mail shirts, breastplate and greaves all overlap. So an attack could be considered deflected by greaves (outermost layer) because it didn't punch through all layers of armor.

Were greaves deflections all from the lower body or also other bodyparts?

There were definitely deflections from all parts covered by greaves.  Also a good suggestion.

Try lowering MAX_EDGE of all metals to an (equal) low number, and then see if there is a difference between adamantine and steel when it comes to cutting up weaker-than-steel materials.

I tried this, and the addy axes still win 643 to 541 (I only ran one round of 1152 this time).  This ratio is consistent with the original test with addy's MAX_EDGE=100,000.  I would have really liked this to work, as that would say that MAX_EDGE (and hence weapon quality) actually does something to impact damage, but it seems not.  Here's to hoping we eventually find a case where it matters in the range normally used by weapons...

Since deflections seemed to be coming primarily from the greaves, I tried another test where I removed the greaves.  I kept all other armor components (iron breastplate, mail shirt, helm, gauntlets and low boots) and set MAX_EDGE back to 100,000 for addy.  As mentioned above, dwaves wielded axes modded to only have hacking attacks.  These tests used standard dwarves, not arena dwarves (as I forgot to check that until just now and for some reason had switched back to standard dwarves the last time I ran tests).  Anyway, with bare-assed dwarves, the dwarves wielding steel axes win 1238 to 1167.  This is a fairly close, but the difference is just over two standard deviations, so it appears to be significant. 

Urist - I believe you said at one point you have a formula for calculating the contact area of body parts, but I wasn't able to find it.  My current theory is that the chance of deflection goes up of the contact area of the weapon is similar to the size of the body part, or something like that (as axes deflect more off the large dwarven asses, and bolt deflect more off hands, feet, and heads).  Anyway, it would be interesting to plot deflection data vs. body part size.

For these tests, I plan to next pit dwarves vs. elephants, for both practical and historical reasons.  I will probably give the dwarves shields, and max out shield, dodge, and armor use skills, as I'd prefer to measure how many hits it takes a dwarf to kill an elephant, not the other way around...

Edit:
And the results are in.  Addy does seem to have a clear advantage for killing elephants.  I gave dwarves full adamantine armor with adamantine shields.  Dwarves were great fighter/axedwarf, grand master armor user, shield, and dodger.  All gear in this and previous tests was standard quality.  When given steel axes, dwarves killed 949 elephants in 29807 hacks, with 164 dwarves dead.   With adamantine axes, 1032 elephants were killed in only 4828 hacks, with 90 dwarves dead.  I will have to try running again with MAX_EDGE lowered for addy, but I really should go to work now...

Edit #2:
Work is overrated anyway.  I ran again with addy's MAX_EDGE set to 10,000, and this time I found that dwarves killed 978 elephants in 7036 hacks, with 160 dwarves dead.  This appears to be a significant difference compared to MAX_EDGE=100,000, so I'd say it does matter here.   However, it's also not the same as steel, so other things seem to matter as well. 

Summary of my conclusions so far:
(1) Steel is as good as (and maybe a little better than) adamantine for fighting armored humanoids with at maximum iron armor. 
(2) Adamantine has a clear advantage for fighting large creatures (elephants).
Given that steel-clad dwarven axelords can easily decimate goblin invaders, gaining a slight advantage in killing them using steel axes is not very important.  This is clearly outweighed by the advantage adamantine gives in killing large creatures, which actually pose a threat to a skilled military.  Time to go make my dwarves some adamantine axes...