Bay 12 Games Forum

Please login or register.

Login with username, password and session length
Advanced search  
Pages: [1] 2 3

Author Topic: Dwarven Research: 40.05 combat physics  (Read 5912 times)

Pirate Bob

  • Bay Watcher
  • [ETHIC: TORTURE_FOR_SCIENCE: ACCEPTABLE]
    • View Profile
Dwarven Research: 40.05 combat physics
« on: July 27, 2014, 10:31:27 pm »

First off, Huge thanks to the great Toady One for for all the bug fixes and rebalancing ranged weapons.  I plan to do detailed testing and update here, but for now some basic results.

I ran some quick tests of the frequency which bolts deflect off armor, or cause only nonserious wounds (bruises).  For each test, I used 72 arena dwarf (I'll update with raws later) targets 2 tiles away from arena dwarf shooters, separated by fortifications.  All dwaves had no skills, and were wearing helms, breastplates, greaves, gauntlets, high boots and chain mail shirts.  In the following tables, columns are bolt material and rows are armor material.

Effectiveness of bolts vs. armored dwarves, 40.03 (identical to 34.11)
Spoiler: 34.11/40.03 (click to show/hide)

Effectiveness of bolts vs. armored dwarves, 40.05
Spoiler: 40.05 (click to show/hide)

As you can see, with the new updates bolts are deflected by steel and adamantine, and not by any other armors.  There appear to be 100% deflections off the steel and adamantine plate armored portion of the dwarves, and not deflections off the chain armored parts (neck, facial features, upper arms). 

Edit - It is worth noting that these changes are largely due to internal code modifications rather than changes to the raws.  With these raws, there still would be no armor protection at all in 34.11.  In other words, Toady has changed the code for ranged attacks significantly and careful analysis will be required to fully understand the changes.

Also, with the current raws, adamantine is significantly better than steel, as adamantine chain mail reduces nearly all hits to bruises, while steel chainmail does not (98% nonserious wounds for addy, vs 92% for steel).

Overall, I would say this is a vast improvement over 34.11.  Adamantine is nearly impenetrable, and steel plate is also impenetrable, but steel-clad dwaves are still vulnerable in the areas not covered by plate armor.

It might be nice if the lower armors had a small chance to deflect bolts, and steel wasn't 100% effective, but I don't know if this is possible with the current system.  In 34.11 it was if you set the contact area to 10, as the rounding used then made different thicknesses of plate armor have different effectiveness, but this may have actually been due to one of the rounding errors Toady fixed.  I will run deflection vs. [SHOOT_FORCE] calculations today, and post results when they are available. 

Also, I would greatly welcome comments or suggestions on what should test and how I should do it or results of other people's testing.  If anyone would like to try my automated testing scripts I am more than happy to share them.  DFFD doesn't seem to be working now, so I am open to other suggestions of where to place files, or can just email them to you if you send me a PM.

Edit - here are the raws for Arena Dwarves
Spoiler: Arena Dwarves (click to show/hide)
« Last Edit: August 03, 2014, 05:31:46 pm by Pirate Bob »
Logged

Pirate Bob

  • Bay Watcher
  • [ETHIC: TORTURE_FOR_SCIENCE: ACCEPTABLE]
    • View Profile
Re: Dwarven Research: 40.05 combat physics
« Reply #1 on: July 27, 2014, 10:34:38 pm »

I will post melee combat results here soon.  My initial results suggest that melee combat follows the same rules as ranged attacks, but I need to do some more testing to confirm this.  The bug fixes in 40.05 should make this easier.
« Last Edit: July 28, 2014, 06:45:51 am by Pirate Bob »
Logged

Untelligent

  • Bay Watcher
  • I eat flesh!
    • View Profile
Re: Dwarven Research: 40.05 combat physics
« Reply #2 on: July 27, 2014, 11:48:49 pm »

Neat! Might be some rejiggering left to do, but being able to improve your soldier's ranged protection with steel/adamantine platemail is good.
Logged
The World Without Knifebear A much safer world indeed.
regardless, the slime shooter will be completed, come hell or high water, which are both entirely plausible setbacks at this point.

mnjiman

  • Bay Watcher
    • View Profile
Re: Dwarven Research: 40.05 combat physics
« Reply #3 on: July 28, 2014, 12:46:36 am »

This makes me happy... because it actually makes sense.
Logged
I was thinking more along the lines of this legendary champion, all clad in dented and dinged up steel plate, his blood-drenched axe slung over his back, a notch in the handle for every enemy that saw the swing of that blade as the last sight they ever saw, a battered shield strapped over his arm... and a fluffy, pink stuffed hippo hidden discretely in his breastplate.

Sergarr

  • Bay Watcher
    • View Profile
Re: Dwarven Research: 40.05 combat physics
« Reply #4 on: July 28, 2014, 03:46:21 pm »

Finally, a reason to get all that sweet steel armor! Praise the toads!
Logged
The real solution is to smash capitalism.
a disastrous fluke is being weaponized by shitlords

thvaz

  • Bay Watcher
    • View Profile
Re: Dwarven Research: 40.05 combat physics
« Reply #5 on: July 28, 2014, 04:02:59 pm »

About melee combat... pulp has a serious problem. I couldn't determine if it is a bug or balance issue...but horses pulping dwarves heads through bronze helmets with their hooves in one strike seems seriously broken. A serious damage through concussion,yes, it was expected, and this worked well before 0.40.xx. If anyone has a insight about this problem I think this would be the thread for it.
Logged

Kamamura

  • Bay Watcher
    • View Profile
Re: Dwarven Research: 40.05 combat physics
« Reply #6 on: July 28, 2014, 04:36:16 pm »

The notion that a steel clad warrior should be 92 percent safe against crossbow bolts is ridiculous. A direct hit with a crossbow (i.e. no glancing hit) should always result in armor penetration.
Logged
The entire content consists of senseless murder, a pile of faceless naked women and zero regard for human life in general, all in the service of the protagonist's base impulses. It is clearly a cry for help from a neglected, self absorbed and disempowered juvenile badly in need of affectionate guidance. What a sad, sad display.

Untelligent

  • Bay Watcher
  • I eat flesh!
    • View Profile
Re: Dwarven Research: 40.05 combat physics
« Reply #7 on: July 28, 2014, 04:38:10 pm »

The notion that a steel clad warrior should be 92 percent safe against crossbow bolts is ridiculous. A direct hit with a crossbow (i.e. no glancing hit) should always result in armor penetration.


Nobody is saying they should be safe. This is not the final version of ranged combat.
« Last Edit: July 28, 2014, 04:42:23 pm by Untelligent »
Logged
The World Without Knifebear A much safer world indeed.
regardless, the slime shooter will be completed, come hell or high water, which are both entirely plausible setbacks at this point.

Authority2

  • Bay Watcher
  • [ETHIC:INCREASE_FPS:REQUIRED]
    • View Profile
Re: Dwarven Research: 40.05 combat physics
« Reply #8 on: July 28, 2014, 04:49:49 pm »

Did the carrying weight of bolts get reduced too? I think someone said they massed one kilogram each and the time I played a marksdwarf in adventure mode it really weighed me down  to carry additional ammo.
Logged
:3

Misterstone

  • Bay Watcher
    • View Profile
Re: Dwarven Research: 40.05 combat physics
« Reply #9 on: July 28, 2014, 09:11:36 pm »

What is the effectiveness of chain? Or for that matter, adamantine cloth?

Also wondering what areas the full plate leaves exposed...  Face?
Logged

sv-esk

  • Bay Watcher
  • sorry for my bad english
    • View Profile
Re: Dwarven Research: 40.05 combat physics
« Reply #10 on: July 29, 2014, 07:45:33 am »

What is the effectiveness of chain? Or for that matter, adamantine cloth?

Also wondering what areas the full plate leaves exposed...  Face?
Chain converts edge damage to blunt.
Armor with [STRUCTURAL_ELASTICITY_WOVEN_THREAD] token is not armor even if made from metal. All clothes has this token.
Neck, upper arms, throat - protected only by mail shirt (body armor with [UBSTEP:1])
Cheeks, lips, eyes, eyelids, nose, ears, mouth, tongue, tooth - not protected at all(you need body armor with [UBSTEP:2] - only mods)

If you get arrow or even spear in eye your brains will not be damaged. Only eye or eyelid will be damaged. It seems in DF all creatures like snails
Spoiler: "snail" (click to show/hide)

but lips, tooth, ears, nose also on sticks
« Last Edit: July 29, 2014, 08:04:01 am by sv-esk »
Logged

Pirate Bob

  • Bay Watcher
  • [ETHIC: TORTURE_FOR_SCIENCE: ACCEPTABLE]
    • View Profile
Re: Dwarven Research: 40.05 combat physics
« Reply #11 on: July 29, 2014, 07:54:19 am »

Here are some preliminary results for deflection of copper bolts vs. [SHOOT_FORCE].  All results are for unskilled arena dwarves, using the setup described in the OP.
Spoiler (click to show/hide)
Deflection is of course when the logs actually report a bolt deflected, and "nonserious" is if a bolt either deflected or resulted in only bruises.

My initial guess is that deflection can occur if ([IMPACT_FRACTURE]-[IMPACT_YIELD]/2)*round100([CONTACT_AREA]*[LAYER_SIZE])/100000 is greater than [SHOOT_FORCE].   Round100(x) is equal to x rounded down to the nearest 100, with a minimum of 100 (this is what was used in 34.11, and the rough results suggest it is unchanged as I do not see a "shoulder" in the deflection curves which would be expected if the different layer thicknesses behaved differently).
Here are the deflection results plotted vs. normalized force = [SHOOT_FORCE]/([IMPACT_FRACTURE]-[IMPACT_YIELD]/2)/round100([CONTACT_AREA]*[LAYER_SIZE])
Spoiler (click to show/hide)
As you can see, deflection starts to occur when the normalized force is less than 100,000, and increases smoothly until it reaches full deflection at about 80,000.  Statistical errors are around 2% due to my sample size, which causes the small-scale jaggedness of the curves.  Note that the plateau at deflection =0.9 is because about 10% of the dwarf is covered only by chain mail (upper arms, face, neck).  Chain mail appears to reduce most wounds to nonserious at normalize force about equal to 40,000, and deflect at around 15,000.

I believe the weird kink in deflections off steel armor is because steel has relatively low IMPACT_STRAIN_AT_YIELD.  In 34.11, deflection could only occur if bolt momentum was less than 50,000/IMPACT_STRAIN_AT_YIELD.  Otherwise, the bolts would not deflect, but would be significantly slowed and cause only bruises.  In this version, either the constant used in this calculation is no longer 50,000 or the bolt momentum is calculated differently (probably the latter).

I will add a summary of these results to the OP once I have verified them better.
« Last Edit: July 29, 2014, 07:57:08 am by Pirate Bob »
Logged

Footkerchief

  • Bay Watcher
  • The Juffo-Wup is strong in this place.
    • View Profile
Re: Dwarven Research: 40.05 combat physics
« Reply #12 on: July 29, 2014, 09:39:28 am »

Neck, upper arms, throat - protected only by mail shirt (body armor with [UBSTEP:1])
Cheeks, lips, eyes, eyelids, nose, ears, mouth, tongue, tooth - not protected at all(you need body armor with [UBSTEP:2] - only mods)
but lips, tooth, ears, nose also on sticks

Are you sure about that?  If so, this bug may have been fixed: http://www.bay12games.com/dwarves/mantisbt/view.php?id=1821
Logged

Pirate Bob

  • Bay Watcher
  • [ETHIC: TORTURE_FOR_SCIENCE: ACCEPTABLE]
    • View Profile
Re: Dwarven Research: 40.05 combat physics
« Reply #13 on: July 29, 2014, 10:39:11 am »

Neck, upper arms, throat - protected only by mail shirt (body armor with [UBSTEP:1])
Cheeks, lips, eyes, eyelids, nose, ears, mouth, tongue, tooth - not protected at all(you need body armor with [UBSTEP:2] - only mods)
but lips, tooth, ears, nose also on sticks

Are you sure about that?  If so, this bug may have been fixed: http://www.bay12games.com/dwarves/mantisbt/view.php?id=1821
My results suggest that chain mail shirts protect the facial features, as deflection goes to 100% at low forces, but I am not totally sure about this.  It is possible that at those low forces bolts deflect off unarmored body parts, but this seems unlikely.  Is there a specific test that I should run to confirm if this bug still exists?  Maybe just mod a breastplate to have [UBSTEP:MAX] and [LBSTEP:MAX] and see which body parts are protected?   I believe I already have scripts somewhere which parse the logs to determine deflection rate off each body part.

sv-esk

  • Bay Watcher
  • sorry for my bad english
    • View Profile
Re: Dwarven Research: 40.05 combat physics
« Reply #14 on: July 29, 2014, 11:45:24 am »

Quote
My results suggest that chain mail shirts protect the facial features,
How?
Spoiler: no protection (click to show/hide)
Sometimes all attacks are deflected by helm, not sure, why. Move around a bit and try again.
If so, this bug may have been fixed: http://www.bay12games.com/dwarves/mantisbt/view.php?id=1821
Not fixed
Spoiler: not fixed (click to show/hide)
Spoiler: panzer (click to show/hide)
« Last Edit: July 29, 2014, 12:09:44 pm by sv-esk »
Logged
Pages: [1] 2 3