Damage modifiers and the Might attribute

[Matt71]

Hey everyone

 

I've been trying to figure out how damage modifiers are taken into account by the game.

 

I couldn't find any post about this, and the "obvious" answers (either multiplicative or additive stacking) both didn't seem to work.

 

This may not be something new (first post for me here, although I've been lurking around a bit before), but I finally found out that it works very similarly as what is explained in this thread about recovery/reload times:

 

https://forums.obsidian.net/topic/98679-mechanics-attack-speed-recovery-time-reload-time/

 

Basically:

 

The game adds all positive modifiers. Lets call P that sum.

 

Then, for each negative modifier, the game computes mod/(1+mod). Let's call that result N(mod)

 

Then the game adds P and all of the N(mod). Lets call S the sum.

 

then:

 

If S>0:

     final_coef = S+1

else:

     final_coef = 1/(1-S)

 

Real damage is obtained by multiplying the base damage by final_coef.

 

For instance, let's say you have 10 base damage, +15% from might, +55% from sneak, and two maluses, M1=-25%, M2=-50%

 

P = 0.15 + 0.55 = 0.7

N(M1) = -0.25 / (1 + -0.25) = -0.33...

N(M2)= -0.50 / (1 + -0.50) = -1

S = 0.7 + -0.33 + -1 = -0.63

final_coef = 1 / (1 - -0.63) = 1 / 1.63

Real damage = 10 * (1 / 1.63) = 10 / 1.63 = roughly 6

 

One interesting conclusion of those calcultations is that a might bonus isn't really important if you have other important bonuses from other sources.

Let's say you have a rogue with +50% from sneak and +50% from devastating blows, a might bonus on top of that won't add much to damage (especially since you prolly also have bonuses from the weapon and possibly from the skill). However, a might malus has a much bigger impact on final damage.

 

Edit:

 

If you can manage to have S > 0, then might (bonus or malus) will have little impact on your damage output. This is even more true if you can get S > 1 or, even better, S > 2. The only situation where might can have a significant impact on you damage output is if S is close to 0. Let's consider an example to illustrate this.

 

Example : Rogue attacking, base damage = 20, +55% from sneak, +45% from weapon, +50% from deathblows, +200% from devastating blow.

 

if might = 10: P = 3.5, no malus, S = 3.5, final_coef = 4.5, real damage = 90

if might = 17: P = 3.71, no malus, S = 3.71, final_coef = 4.71, real damage = 94.2 (that's only a 4.7% damage increase)

if might = 3: P = 3.5, N = -0.21/0.79 = -0.2658, S = 3.2342, final_coef = 4.2342, real damage = 84.7 (that's only 5.9% damage decrease)

Edited September 12, 2018 by Matt71

[mant2si]

Yeah, you need MIG only for Barbarians ans Caster classes because there not to much modifiers for spells and carnage

Edited September 12, 2018 by mant2si

[Haplok]

And for Amra Axe builds. 

[Elric Galad]

Mig is not very useful for Rogue indeed, as it was in PoE1. Dex is usually better because action speed modifier are more rare. Per is also better because you don't want to miss.

 

Casters have limited spells and less damage modifiers, so Mig helps them more.

[Matt71]

Just one small note about the -35% for dual wielding, since I saw someone asking about that in the "2.1 unique items" thread.

 

-0.35 / ( 1 + -0.35) = (roughly) -0.54.

 

What this means is that, if you have positive damage modifiers that total up (additively) to at least 54% and no other negative damage modifier, then the -35% from dual wielding is in fact exactly the same as a -54% modifier that would be additive with all the other modifiers.

 

As an example, if you have +30% from overpen, +30% from weapon, +12% from might and -35% from dual wielding, then your overall damage modifier is:

 

30 + 30 + 12 - 54 = 18%.

 

Quite funnily, that way of calculating things (double inversion) implies that the -35% from dual wielding is relatively insignificant for chars that can stack positive modifiers to a very high level (i.e. rogues) while it is much more detrimential to chars that don't stack many positive modifiers.

[Boeroer]

Yes, the double inversion of Deadfire is neither intuitive nor is it fun. In my opinion it's just a bad choice. Sure it prevents dmg from dropping too low once to stack too many dmg maluses - but who wants to do that anyways? A bottom cap would have done well enough. 

 

Grazes and other dmg maluses work the same way. That also makes some modals worse than they seem to be (-25% dmg for some benefical effect).

[Matt71]

That's very true.

 

As a matter of fact, I was using the word "funnily" in an ironical way , since it doesn't make much sense that a -35% malus can have a very different impact, depending on the build it's applied to.

 

I personnaly don't understand why Obs didn't go with multiplicative modifiers only. In my opinion, this is what makes the more sense. And if they were affraid that it would be possible to stack too many bonuses and reach a stellar overall bonus, a simple fix would just have been to lower some bonuses.

Edited September 18, 2018 by Matt71

[mant2si]

That's very true.

 

As a matter of fact, I was using the word "funnily" in an ironical way , since it doesn't make much sense that a -35% malus can have a very different impact, depending on the build it's applied to.

 

I personnaly don't understand why Obs didn't go with multiplicative modifiers only. In my opinion, this is what makes the more sense. And if they were affraid that it would be possible to stack to many bonuses and reach a stellar overall bonus, a simple fix would just have been to lower some bonuses.

.... or just set bottom and upper bounds ... but the real problem with their system isn't player's misunderstanding, this complication cause that even their own developers don't  understand how this should work  

 

The real example is negative crit modifier - which lead critical strikes do less damage than normal strikes

Or enchants which should balance something like +15%, -15% negative effects

Or grazes which cause -75% addictive damage instead of lowered resulted damage by half 

Or penetration +3 difference in armor lower damage by 300% and +2 difference lower them only by 75% - (I'm think that original idea was lowering incoming damage exactly by 75%, 50%, 25%, this make sense because then you will be able damage bosses with high armor )

Or .... 

 

In summary Deadfire is good example, why developers must always prefer simple things over complicated

Edited September 18, 2018 by mant2si

[mant2si]

The second funny thing is their PL system, I think only few people can explain what +2PL or +1PL mean, this is like magic, let's look on simple example

1. Monks fists DM/PEN scaling - 7PL - Mean +11 PEN, +16ACC, DM 60% and 12PL mean 14PEN, 28 ACC, 90% - 28 ACC the highest weapon accuracy bonus in the game
2. Rogue sneak attack - 7PL mean ~ 45% DM and 12 PL - 70% 
3. Any martial ability +5% to base damage per PL, any magic ability +10% DM per PL and +0.3 PEN, +5% ? to duration

Edited September 18, 2018 by mant2si

[Matt71]

I 100% agree with you, mant2si, about power levels.

 

For instance, I just found the Mask of the Grotto Deep (+2 Power Levels for all Poison keyword Abilities) and I have no way to know how that translates in terms of ability efficiency.

Guess I could know by doing some testing, but that's not the way it should work.

 

As a side note, graze and -2 pen is -100% damage, not -75%:

 

-0.50 / ( 1 + -0.50) = -1 = -100%

 

Of course this is only true if you already have positive bonuses that total up to at least 100%, otherwise computations are much more messy.