Bay 12 Games Forum

Please login or register.

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

Author Topic: Dwarven Research: A Study of the 'Archer' and 'Marksdwarf' Skills  (Read 6850 times)

Zivilin

  • Bay Watcher
  • Dwarven Science Division
    • View Profile

Dwarven Research: A Study of the 'Archer' and 'Marksdwarf' Skills

INTRODUCTION
Little information is available concerning the exact effects military skills have on dwarven battle performance. It is generally accepted that two skills are associated with ranged crossbow combat: the 'Marksdwarf' skill and 'Archer' skill. The former is a weapon skill and its general effects - increase in accuracy and rate of fire - are known [1]. The effects of the 'Archery' skill, however, are shrouded in mystery, and it is only known that skill level increases with the use of any ranged weapon [2]. The aim of this study is to elucidate the exact effects of these two skills.


METHODOLOGY
In order to address a wide range of performance criteria, the Author divided the study into three experiments:
  • Damage
  • Rate of Fire
  • Range

In all of the above Arena Dwarves (defined here) were the test subjects. Similar to the Author's previous studies, arena dwarves were divided into two teams: A (Assailant) and D (Defender). Details are given in the table below.

| Team A| Team D
Weapon:| Cedar crossbow| None
Ammo:| 100 [varies] bolts| None
Armor:| None| varies
Skills:| Archer: varies
| Marksdwarf: varies
|
|


Damage Test
The damage tests were performed in the same manner as the author's previous study i.e., by using a large number of tiny 1vs1 cells with a single fortification tile separating Assailant and Defender. In order to further improve on the accuracy of the results, 5000 such cells were used to evaluate weapon performance in each scenario.

8 scenarios were tested:

(Bolt  vs  Armor) (Skill)     (Skill)
--------------------------------------------
(Steel vs Copper) (Archer 0)  (Marksdwarf 0)
(Steel vs Copper) (Archer 14) (Marksdwarf 0)
(Steel vs Copper) (Archer 0)  (Marksdwarf 14)
(Steel vs Copper) (Archer 14) (Marksdwarf 14)
(Copper vs Steel) (Archer 0)  (Marksdwarf 0)
(Copper vs Steel) (Archer 14) (Marksdwarf 0)
(Copper vs Steel) (Archer 0)  (Marksdwarf 14)
(Copper vs Steel) (Archer 14) (Marksdwarf 14)

The purpose of these was to see how and if different skill extremes affect damage in two characteristic bolt-armor scenarios - bolts-tearing-armor and bolts-bludgeoning-armor.

The results, as if to spite the Author's 40'000 dwarf sacrifice, are utterly unimpressive. Skill level in both 'Marksdwarf' and 'Archer' made almost no difference on the type of damage sustained by Defenders. 'Marksdwarf' showed traces of influence, ever so slightly increasing the percentage of fractured (by 10% in Copper vs Steel and by 7-8% in Steel vs Copper) and shattered bones (by around 30%), but since they constitute no more than 10% and 2% of all injuries in the first place, respectively, this does not make a significant difference on combat performance. Probabilities of hitting a particular body part also remained the same throughout all scenarios, conforming to results presented on the Virtruvian Dwarf (bottom of post). This strongly coincides with the relative sizes of the body parts found in the raws, as supplied by Urist da Vinci here

The only difference found was in the Average Number of Bolts Fired in Order to Kill a Dwarf (Fig 1.)


At first glance, it seems that a counter-intuitive result is obtained: Assailants with higher 'Marksdwarf' skill used up more bolts to kill a Defender. This, however, can be easily explained by the following test.


Rate of Fire Test
Rate of Fire was obtained by counting the number of ticks between consecutive shots by Marksdwarves of equal skill. Due to the troublesome nature of counting ticks, only 10 1vs1 cells were used and only 6 volleys were fired, giving 40 to 50 samples. All arena dwarves had the same equipment:

| Team A| Team D
Weapon:| Cedar crossbow| None
Ammo:| 100 copper bolts| None
Armor:| None| All steel armor
Skills:| Archer: varies
| Marksdwarf: varies
|
|

4 scenarios were tested:

(Skill)     (Skill)
--------------------------
(Archer 0)  (Marksdwarf 0)
(Archer 14) (Marksdwarf 0)
(Archer 0)  (Marksdwarf 14)
(Archer 14) (Marksdwarf 14)

The results are presented in figure 2:

Spoiler: Figure 2. Rate of Fire (click to show/hide)

Again, only the 'Marksdwarf' skill has any effect on Rate of Fire. Grand Master Marksdwarves (skill level 14) fire roughly 1.7 times quicker then Dabblers.

Rate of Fire also explains why better marksdwarves use up more ammunition to kill a dwarf. There are a number of injuries (upper/middle spine bruising/tear, throat/heart tear) that cause death after a time delay, and only one (brain tear) that causes instant death. The probability of dealing a time-delayed death blow is the same for all skills, but the masters can get in more shots before the the victim actually suffocates/bleeds out. Thus, they are more wasteful in the ammunition department.


Range Test [+Dodging]
Three ranges were tested: 5 tiles, 10 tiles and 20 tiles. The design for the range experiments is given on Figure 3.


In order to remove the effects fortifications may have on accuracy, they were removed. The only other method to restrict movement known to the Author was to surround test subjects with pits, and thus it was done. Unlike previous experiments, Defenders can move to tiles surrounding them - this enables dodging, and thus dodging was included to the Test criteria. Unfortunately, a dodging Defender is certain to move off its prison pillar and fall into the pit below, thus changing its distance from the the Assailant and changing the Test's parameters. In order to bypass this difficulty, each assailant was given exactly ONE bolt. 4000 Experiments per scenario were performed, giving 4000 bolts fired per scenario. Arena dwarves were equipped as follows:

| Team A| Team D
Weapon:| Cedar crossbow| None
Ammo:| 1 copper bolt| None
Armor:| None| All steel armor
Skills:| Archer: varies
| Marksdwarf: varies
|
|

12 scenarios were tested:

(Range) (Skill)     (Skill)
--------------------------------------------
(5)     (Archer 0)  (Marksdwarf 0)
(5)     (Archer 14) (Marksdwarf 0)
(5)     (Archer 0)  (Marksdwarf 14)
(5)     (Archer 14) (Marksdwarf 14)
(10)    (Archer 0)  (Marksdwarf 0)
(10)    (Archer 14) (Marksdwarf 0)
(10)    (Archer 0)  (Marksdwarf 14)
(10)    (Archer 14) (Marksdwarf 14)
(20)    (Archer 0)  (Marksdwarf 0)
(20)    (Archer 14) (Marksdwarf 0)
(20)    (Archer 0)  (Marksdwarf 14)
(20)    (Archer 14) (Marksdwarf 14)

The following criteria were evaluated:
Probability of Hit (PoH) - Hits occur when the flying bolts trajectory crosses the tile occupied by the defender. A hit does not mean that the dwarf takes damage - it can lead to a deflection, glance off, dodge or miss before actually doing damage.

Probability of Dodge (PoD) - provided that a bolt hits, this criterion gives the probability that the dwarf dodges or misses it. Misses are included, since they basically seem to function similarly to dodges, with the exception that the dwarf does not move from the tile he is occupying. Note that all Defenders had absolutely no skill in dodge, so this value depend purely on Assailant skills.

Figure 4 shows the results obtained for PoH for different ranges and different skills:


Observations:
  • 'Archer' skill makes no difference.
  • Increasing 'Marksdwarf' skill provides significant boost to accuracy.
  • Accuracy seems to be (approximately) inversely proportional to distance, at least in the case of Dabbling Marksdwarfs. More generally, perhaps there is an exponential decay of accuracy with distance, whose curve depends on 'Marksdwarf' skill?

Finally, results on PoD values were obtained and presented in Figure 5. These did not vary with range, so range data is omitted.


Observations
  • 'Archer' skill is not useless! It actually has a role in the dodge skill role.
  • 'Marksdwarf' skill still has a leading role, as it decreases the PoD to a larger degree than 'Archer' skill.

Conclusions
  • The 'Marksdwarf' skill is confirmed to affect weapon accuracy and rate of fire. It also decreases enemy dodge chance. It has negligible effect on the damage type.
  • The 'Archer' skill is has almost no impact on crossbow ranged performance. The sole exception found in this study was dodge chance, which decreased significantly with an increase in skill (but still not as significantly as with 'Marksdwarf'). There was absolutely no significant correlation between 'Archer' skill level and damage type.

[Edited for fallacious conclusions]
« Last Edit: October 05, 2012, 01:45:54 am by Zivilin »
Logged
No signature

Gavakis

  • Bay Watcher
  • WAHAHAHAH
    • View Profile
Re: Dwarven Research: A Study of the 'Archer' and 'Marksdwarf' Skills
« Reply #1 on: October 04, 2012, 06:45:43 pm »

wow.
Logged

Roraborialisforealis

  • Bay Watcher
  • [HATEBILLION:UNTHINKABLE]
    • View Profile
Re: Dwarven Research: A Study of the 'Archer' and 'Marksdwarf' Skills
« Reply #2 on: October 04, 2012, 07:34:38 pm »

Awesome
Logged
Spoiler (click to show/hide)

misko27

  • Bay Watcher
  • "Legitimate Businessman"
    • View Profile
Re: Dwarven Research: A Study of the 'Archer' and 'Marksdwarf' Skills
« Reply #3 on: October 04, 2012, 07:58:51 pm »

Yay!Another one of these!

Interesting. So Previous skill with a cossbow does not nescisitate a skill increase in bows.
Logged
Quote from: Legitimate Businessman
There were warning vis a vis your that your legs would be very broken should you not come back alone.
Occam's Razor: "Among competing hypotheses, the hypothesis with the fewest assumptions should be selected."
Hanlon's Razor-attachment: "Never attribute to malice that which is adequately explained by stupidity."

Urist McEngraver

  • Bay Watcher
  • I mod as elves. Wonder what that says about me.
    • View Profile
Re: Dwarven Research: A Study of the 'Archer' and 'Marksdwarf' Skills
« Reply #4 on: October 04, 2012, 07:58:53 pm »

Nice work.  8)
Logged
Also, if you've got undead wooden furniture in the dining hall, you have some big issues anyways.
This is Dwarf Fortress, not the Fluffy Wambler Express.

Jahsg

  • Bay Watcher
    • View Profile
Re: Dwarven Research: A Study of the 'Archer' and 'Marksdwarf' Skills
« Reply #5 on: October 04, 2012, 08:12:25 pm »

damn
Logged

arkhometha

  • Bay Watcher
  • [PREFSTRING: ARCHAEOLOGY]
    • View Profile
Re: Dwarven Research: A Study of the 'Archer' and 'Marksdwarf' Skills
« Reply #6 on: October 04, 2012, 08:17:59 pm »

Interesting work! Thanks a lot!
Logged

Zivilin

  • Bay Watcher
  • Dwarven Science Division
    • View Profile
Re: Dwarven Research: A Study of the 'Archer' and 'Marksdwarf' Skills
« Reply #7 on: October 05, 2012, 01:37:59 am »

Yay!Another one of these!

Interesting. So Previous skill with a cossbow does not nescisitate a skill increase in bows.

Pretty much. If 'Marksdwarf' and 'Bowdwarf' operate in a similar way, then dwarves with High 'Archer' and low 'Bowdwarf' are woefully inaccurate at long distances, but if they actually hit then there's less of a chance that the enemy will dodge. Still, higher weapon skill does the same, and does it far better.
Logged
No signature

Dutchling

  • Bay Watcher
  • I can smell fear on you.
    • View Profile
Re: Dwarven Research: A Study of the 'Archer' and 'Marksdwarf' Skills
« Reply #8 on: October 05, 2012, 06:06:56 am »

Well that gives me a lot of free points to spend for my ranged adventurers...

Has there ever been a similar test regarding the fighter skill?
Logged
is literally computer
    beep boop feelz

Zivilin

  • Bay Watcher
  • Dwarven Science Division
    • View Profile
Re: Dwarven Research: A Study of the 'Archer' and 'Marksdwarf' Skills
« Reply #9 on: October 05, 2012, 01:40:43 pm »

Well that gives me a lot of free points to spend for my ranged adventurers...

Has there ever been a similar test regarding the fighter skill?

Not to my knowledge. I'm trying to branch out into melee testing, so hopefully I'll get around to it eventually.

On that note, perhaps someone knows how to disable wrestling skills from a dwarf without disabling his ability to equip armor? I edited punches, kicks etc. out of the dwarf raw files, but wrestling skills seem to be tied to the [GRASP] and [LIMB] tags in the body_default.txt. When I edited [GRASP] out from hands, this caused my arena dwarves to become unable to don gauntlets. I'd prefer they had their gauntlets, as long as they don't try to strangle anyone with them.
Logged
No signature

Urist Da Vinci

  • Bay Watcher
  • [NATURAL_SKILL: ENGINEER:4]
    • View Profile
Re: Dwarven Research: A Study of the 'Archer' and 'Marksdwarf' Skills
« Reply #10 on: October 05, 2012, 02:56:47 pm »

How about repeating the 1-bolt accuracy test vs shields? At high skill levels, and using only one shield (no exploits), who wins - archer/marksdwarves or shield-users?

Zivilin

  • Bay Watcher
  • Dwarven Science Division
    • View Profile
Re: Dwarven Research: A Study of the 'Archer' and 'Marksdwarf' Skills
« Reply #11 on: October 05, 2012, 05:57:53 pm »

How about repeating the 1-bolt accuracy test vs shields? At high skill levels, and using only one shield (no exploits), who wins - archer/marksdwarves or shield-users?

It is done. I reduced the range to two tiles, so as to cram as many experiment cells as possible into the arena and finish it with one run.

3920 Grand Master Marksdwarves/Archers with cedar crossbows and one copper bolt each
vs
3920 Grand Master Shield Users with steel shields.

The result is an overwhelming victory for shield users.
Out of 3920 shots, 3651 were blocked. That's an insane 93% block rate. Very impressive. This might merit a shield study.
Logged
No signature

Shoes

  • Escaped Lunatic
    • View Profile
Re: Dwarven Research: A Study of the 'Archer' and 'Marksdwarf' Skills
« Reply #12 on: October 06, 2012, 05:45:48 am »

What about bows and arrows? Wouldn't that be the weapon of choice of archers?
Logged

Delioth

  • Bay Watcher
  • Mass Gauntlet Wizard
    • View Profile
Re: Dwarven Research: A Study of the 'Archer' and 'Marksdwarf' Skills
« Reply #13 on: October 06, 2012, 08:48:48 am »

What about bows and arrows? Wouldn't that be the weapon of choice of archers?

As of now, there is NO functional difference between bows and crossbows. Also, the 'Archer' skill relates to general ability to use ranged weapons.
Logged
Products of my boredom (RAW format)
The same products in a pretty mod!
Quote from: Shawtay
Delioth, you sir, are a wonderful, wonderful person.
When in doubt, Apply Magma= Significant other problems-Apply Magma. Issues with car-Apply Magma (Collect insurance). Troubles at school/work-Apply Magma. Hard Bed-Apply Magma (Collect insurance)

misko27

  • Bay Watcher
  • "Legitimate Businessman"
    • View Profile
Re: Dwarven Research: A Study of the 'Archer' and 'Marksdwarf' Skills
« Reply #14 on: October 06, 2012, 01:11:19 pm »

What about bows and arrows? Wouldn't that be the weapon of choice of archers?

As of now, there is NO functional difference between bows and crossbows. Also, the 'Archer' skill relates to general ability to use ranged weapons.
What about the arrow and bolt? I've noticed differences in the RAWs between them. And the Dart.
Logged
Quote from: Legitimate Businessman
There were warning vis a vis your that your legs would be very broken should you not come back alone.
Occam's Razor: "Among competing hypotheses, the hypothesis with the fewest assumptions should be selected."
Hanlon's Razor-attachment: "Never attribute to malice that which is adequately explained by stupidity."
Pages: [1] 2