Pierce vs Blunt vs Cut Damage

[Tydeus]

={(ran.*0.1 + 0.9)*RD} - round[ {(1.0 - 1.0 / ((ran.*0.55 + 0.45)*RF*0.014))*((ran.*0.1 + 0.9)*RD - (ran.* 0.55 + 0.45)*SF)} + {(ran.*0.55 + 0.45)*SF}]

I believe that is what was missing.

It's such a cluster fuck.

If it makes it any easier to read for the layman, you could alternatively look at this:

(click to show/hide)

Code: [Select]

        soak_factor = armor * armor_soak_factor_for_damage_type
        reduction_factor = armor * armor_reduction_factor_for_damage_type

        if weapon_type < 3:
            soak_factor *= armor_extra_penetration_soak_factor
            reduction_factor *= armor_extra_penetration_reduction_factor

        randomized_soak = (n * 0.55 + 0.45) * soak_factor
        randomized_damage = (n * 0.1 + 0.9) * raw_damage
        soaked_damage = randomized_damage - randomized_soak

        randomized_reduction = math.exp((n * 0.55 + 0.45) * reduction_factor * 0.014)
        reduced_damage = (1.0 - 1.0 / randomized_reduction) * soaked_damage

        damage_difference = round(reduced_damage + randomized_soak)
        effective_damage = randomized_damage - damage_difference


cmp posted these a while ago, probably not for the layman:
http://forums.taleworlds.com/index.php/topic,168722.msg4090900.html#msg4090900
http://forums.taleworlds.com/index.php/topic,168722.msg4090940.html#msg4090940

[CaptainQuantum]

={(ran.*0.1 + 0.9)*RD} - round[ {(1.0 - 1.0 / ((ran.*0.55 + 0.45)*RF*0.014))*((ran.*0.1 + 0.9)*RD - (ran.* 0.55 + 0.45)*SF)} + {(ran.*0.55 + 0.45)*SF}]

I believe that is what was missing.

It's such a cluster fuck.

If it makes it any easier to read for the layman, you could alternatively look at this:

(click to show/hide)

Code: [Select]

        soak_factor = armor * armor_soak_factor_for_damage_type
        reduction_factor = armor * armor_reduction_factor_for_damage_type

        if weapon_type < 3:
            soak_factor *= armor_extra_penetration_soak_factor
            reduction_factor *= armor_extra_penetration_reduction_factor

        randomized_soak = (n * 0.55 + 0.45) * soak_factor
        randomized_damage = (n * 0.1 + 0.9) * raw_damage
        soaked_damage = randomized_damage - randomized_soak

        randomized_reduction = math.exp((n * 0.55 + 0.45) * reduction_factor * 0.014)
        reduced_damage = (1.0 - 1.0 / randomized_reduction) * soaked_damage

        damage_difference = round(reduced_damage + randomized_soak)
        effective_damage = randomized_damage - damage_difference


cmp posted these a while ago, probably not for the layman:
http://forums.taleworlds.com/index.php/topic,168722.msg4090900.html#msg4090900
http://forums.taleworlds.com/index.php/topic,168722.msg4090940.html#msg4090940

Yes I got it from there, it's quite hard for the layman to follow however since it's purely in code, there are no hard terms to understand from non-coders however.

[Paul]

lern2euler

math.exp(x) = e^x

TW prolly only did this to avoid DIV0.  ->

[CaptainQuantum]

lern2euler

math.exp(x) = e^x

TW prolly only did this to avoid DIV0.  ->

/Facepalm, thanks for noticing that I omitted it, I don't know why my brain skipped over it. That is probably the reason they did it, computers don't like it when you give it a division by zero. I will edit my main post now Choice of e is rather fitting as I can tell though, because it is e^-x, which shows a decay, and all decays in Physics can be approximated to this relation.

[Paul]

1-1/e^(x*0.014) works pretty well though for reduction. For armor(x) of 0 it returns 0 reduced damage, 50 armor gives about 50% reduction and values above that reduce more while never reaching 100%. That is for ignoring armor parameter and randomness. So armor values used in Warband(0 to 70) get a pretty linear relation to the percentage of the reduced damage despite of the exponential function.

[Elmokki]

http://forum.c-rpg.net/index.php/topic,23607.msg341261.html#msg341261 now contains some analysis of damage types.

Simplified:

With about 60 base raw damage or more pierce > blunt > cut. Below that blunt > pierce > cut.

So people with low power strike should always prefer blunt over pierce if all other variables are the same. Blunt also does more damage when you get negative speed bonuses etc.

Difference between pierce and blunt is fairly small in all cases though. With very low raw damage blunt wins at considerable numbers proportionally, but the absolute damage is so very low that it hardly matters. At high raw damage pierce wins, but even at 100 raw damage it's just 7% more damage (31 average vs 29 average against 80 armor)

All in all if you could choose between equal stats weapon with blunt and pierce, blunt is slightly a better choice. Blunt weapons do have a knockdown generally though, which boosts them quite a lot.

If someone wants to take a look or play with it, here's the Excel spreadsheet I used: http://nikita.tnnet.fi/~elmokki/crpg-damage.xls