Detailed explanation of stats?

Every hero has a set of visible and hidden stats. In order to better understand how one’s heroes will perform, I would like to know what each of those stats do. In particular:

How are the MOD stats related to the stats they supposedly modify? Are they multipliers out of 100 or 1000 or something? Or are they growth rates? When I look at the character’s normal stats, are the MOD stats already factored into them, or is that a raw value that is combined in some way with the MOD stat later?

With regard to both attack stats, it is unclear whether they are raw damage per hit or damage per second based on some raw damage mutiplied by fire rate? If it’s a rate, is reload time and ammo factored into it?

The crit stats appear to be half of the corresponding attack stat. Is this always true? It would make the most sense if a stat like that was just the damage added to base damage when a crit occurs. If so, then what determines the probability of a crit? Is it fixed for everyone the same? Fixed for each hero? Increase with level?

Defense stats obviously reduce incoming damage, but how effectively? Is it a flat reduction per hit (e.g. if my defense is greater than your per-bullet attack, you do no damage)? Or is the reduction proportional to incoming raw bullet damage (e.g. a defense of 100 will reduce damage by 10%)? If proportional, is it calculated as a flat factor (ee.g. It can reach 100% at some ludicrously high value) or is it tiered (e.g. first 100 defense are worth 10%, next 100 worth 9%, etc so that reaching 100% is not theoretically possible? Some other mechanism like a hyperbolic curve to keep it under 100%?

With regard to the elemental stats, the conditions where they would be used are very unclear. Is it a factor in every attack? Does it add to normal attack and defense, or does it multiply? Does elemental attack only affect hits with an element advantage, and elemental defense only an elemental disadvantage?

Finally, do attack stats influence skill damage? If the skill has a non-elemental damage, is the elemental attack and defense ignored? This would be very useful information because it would heavily influence the value of skill point investments.

Thanks!

3 Likes

I was just going to ask how armor is calculated. But I’m borriwing your (great) thread.
How is armor calculated?
So Dogface’s Platinum skill allows the team allies to penetrate xxx armor. What does it mean? All attacks do xxx damage more?

Example:
Dada has damage power of 2000.
Baba has 2000 in armor.

Dada shoots Baba.

Will Dada’s attack deal 0 damage?

Dada activates a armor penetration skill. It has the value 2000.

Baba shoots Dada again.
Will Dada’s attack deal 2000 damage?

If this is how it works then I don’t see what’s so cool about Dogface Platinum skill.

3 Likes

Well, we can get an answer here in 2 ways: the devs can just tell us, or we can very, very carefully and tediously shoot each other and write down numbers over and over again to reverse engineer how it works. Last time i did something like this, i had easy access to increase or decrease one stat at a time (a different game), and the decrease part made things a lot easier. I took 167 data points there to put together the entire damage equation from 4 stats. Some of my questions alone can be answered with 2-4 data points, like when elemental stats are applied or whether the stats affect skill damage. Armor and defense are the hardest. But I do hope I don’t need to go through all that to know these things.

Well, with the upcoming duels, testing with each other should be easier, but getting a dev answer would be ideal :slight_smile:

Oh, and the reason I think something is going on with the stats that isnt obvious is the difference between expected and actual damage by my sentry vs (that mech girl with the 2 pistols? I am awful with names). If i assume both types of damage count and are all-inclusive rates, then (mech girl) should be doing at least 2x as much damage as sentry. Yet when they party together, sentry consistently does way more damage than (mech girl). There are any number of ways this might happen, but until I know exactly which way it did, stats are not very meaningful.

I was just thinking that. It will still be tough though considering that we can’t manually manipulate stats both up and down.

I think its damage per second and not damage per bullet.

So if both do 1000 Dmg, Beck (i think you meant her?) shoots 10 Bullets doing 100 Dmg each, while Sentry shoots one in 2 seconds, doing 2000 Dmg. Same damage in the end, but looks totally different while playing.

That is one reasonable way to think about it. But If the stat is intended as DPS, to calculate it only in terms of bullets per second and damage per bullet will be wrong if the reload time and frequency is not factored into it. I dont know if it is, which is why i ask.
Further, the actual DPS will be very different if defense is reducing damage at a flat rate per bullet when the figure for beck’s dps involves 20 bullets and sentry’s involves 5. Beck would suffer 4x the loss from armor in that case. Again, i dont know if thats how it works , but I’d like to, because 4x armor effect would be a big deal.

1 Like

Bump, and when are we getting these duels?

Anyone want to volunteer to take data?

Also we need to find heroes for testing that meet certain criteria, and find volunteers with those heroes.

  • gear sets that can separately upgrade attack stats 2 or 3 times without promoting
  • gear set that can separately upgrade defense stats 2 or 3 times without promotion
  • Easiest: 2 heroes with same clip size (but different attack stats, of course)
  • heroes with skills that have upgradeable non-elemental damage and no elemental damage
  • as an alternative to gears, we could use heroes with skills that buff the same stats at a flat rate (not by a %) that is upgradeable. In fact that would be a lot better because we could use them to get data out of everyone else, and the stat change would be reversible. Unfortunately these tend to be the high tier passive skills, of which I have none

DPS would be a valuable stat to have for each character. Calculated as if the character had no impediments and just shot at a target for one minute that had 0 armor and neutral.

True. If we figure this all out, we should be able to come up with a way to calculate that. Not certain that it will be at all simple due to possible corrections for refire and reload times. I think the worst case could be a linear equation of damage stats with some parameters you would have to look up in a table we could have to compile ourselves. Best case it’s either exactly the damage stat or sum of damage stats. Most likely, I believe, either of the two should be a good comparator between characters at present, but we won’t know for sure without a dev or data.

Okay! The duels are here, time for data collection and analytics! Volunteers are encouraged to indicate interest in some way while I start to post test protocols and maths.

Example: Dogface Platinum skill makes every hero penetrate 300 armour.
Is this the same as if the skill would improve each hero’s damage by 300?

I just think these numbers are so low. Who really cares about 300 more damage or not. I’ve been chasing Dogface Plat skill for weeks but I just can’t see what’s good about it.

Well, if it is 300 damage per shoot, there are heroes who would certainly love this skill, like Sapphyr, Matador or Baron.

Problem is, we don’t know how damage and armor work, so we can’t measure how good are skills that give armor/penetration.

What from I can tell, armor seems to be more like a flat deduction, which is obvious to tell just from a few enemies in the Gorgon Wakes Raid. That deduction is per bullet as well, so snipers are barely affected by it, but an ongoing would be heavily impacted by it. As for exact numbers, that’s why we have duels now.

Would be interesting to know how much damage each hero makes. Theoretically you need a fixed time in which you measure. (eg 30sec). How many shots can the said hero fire at this time? How often can / must he charge at this time, how long does it take to load the weapon.

The problems I see in it are the following, what do you take as an opponent? (Gorgon would be the easiest because no elemental damage.) Which hero configuration do you take? (eg Dog, Platinum [no other gear], 10 *, skills level 70, based on this, all other heroes would also need the same gear to get equal scores). With or without skill use?

It would be rather difficult to measure the difference between DPS and bullet damage, or including reload or not including reload time, without knowing how armor and elemental damage work first, or without having a reliable way to measure time with decent precision. Obviously there are good ways to measure time, but if we are crowdsourcing measurements, it wouldn’t necessarily be consistent. Tried several times to do the math on this, there is no way to extract times from shot damage alone without knowing a lot of the rest of the calculations. So let’s assume that the damage figures in the stats screen are real DPS including reload time given no armor. Reload times and fire rates will then show up as per-character constants in other calculations that will vary between characters. We can extract them at some point if we want.

There remains the question of how and when elemental damage is applied. I have seen several folks theorize (to their own satisfaction) that elemental damage is added to normal damage only when an element advantage is applied and elemental armor is added to armor only when an element disadvantage is applied. That leaves some questions. How is the extra damage added to skill damage calculated, then, when skills don’t have a reliable “fire rate”? If a skill adds “mech damage” or absorbs “mech damage” or adds “mech armor” to a character, is there any effect in absence of advantage or disadvantage?

Luckily, this is an easy theory to test in challenges. All we need is:

  • a gear G that affects elemental attack, but not attack, for a character A
    • gears include grenade belt, I think
  • a target B with the same element as A
  • a gear H that affects elemental armor but not armor for character C.
    • I want to say gears include boots? Helmets maybe? Check beforehand by long-pressing the gear icon.
  • a character D with the same element as C.

To test,

  • challenge A vs B without applying G
    • collect stats for A,
      • attack I
      • elemental attack J1
    • collect stats for B
      • armor K
      • elemental armor L
  • note single shot damage of A to B, E1
  • apply G to A, challenge A vs B
    • collect stats for A,
      • elemental attack J2
  • note single shot damage of A to B, E2
  • challenge C vs D without applying H
    • collect stats for C,
      • armor M
      • elemental armor N1
    • collect stats for D,
      • attack O
      • elemental attack P
  • note single shot damage of D to C, F1
  • apply H to C
    • collect stats for C,
      • elemental armor N2
  • challenge C vs D
  • note single shot damage of D to C, F2
  • post all data A through P

Obviously we want to know if E1 = E2 and F1 = F2, which would indicate that elemental attack and armor apply only to element advantages or disadvantages. The rest is gravy that can go into a larger dataset. Conducting only the A vs B or only the C vs D parts is useful as well.

Go.

Alright I have a Cinder with appropriate gears for the full elemental attack ramp test procedure. That is:

  • Cinder has gear slots for a grenade belt and goggles, both ready to equip.
    • Her stats:
      • Attack (Damage) An
      • Elemental Attack (Damage):
        • Ae0 before either gear
        • Ae1 with grenade belt only
        • Ae2 with both gears equipped
    • Bronze skill stated (bio-chem) damage Sb
    • Silver skill stated (bio-chem) damage Ss
  • Need 3 volunteer targets of each element, where the target is T
    • Requirements:
      • Enough HP and defense to survive a trial
      • No passive skills that have bonus armor, elemental armor, or damage reduction
      • No passive skills that may produce a buff for armor, elemental armor, or damage reduction.
      • No passive skills that may produce a debuff for damage or elemental damage.
    • During test period (all trials over all volunteer targets) volunteer targets must not be upgraded in any way.
    • Stats collected:
      • Defense (Armor) DnT
      • Elemental Defense (Armor) DeT
    • 1 trial for each value of Ae (0, 1, and 2) conducted across targets first.
      • During trial, volunteer target stands still, does literally nothing.
      • Collect for each trial N:
        • Hit damage for 1 bullet HaTN
        • Hit damage for bronze skill HbTN
        • Hit damage for silver skill HsTN

That produces a data set (3 elements)x(3 skills)x(3 Dn & De)x(3 Ae). Each point group will undergo linear regression, forming a new group (3 slope + 3 intercept). The process is applied recursively until intercepts are 0 or R2 indicates noise. The final set reconstructs a damage equation specific to Cinder with the same An, but the structure should answer all remaining questions about how elemental damage/armor and element advantage work, what affects skill damage other than skill level, and some general information about how armor works.

Some mysteries will remain that will be resolved by subsequent tests using different characters and/or different ramps.

Volunteer targets generously donated by LEGION364. THANK YOU!

Full results are pending some analysis, but the highlights in the abstract:

  • Elemental damage is part of every bullet’s damage, not just when there is an element advantage.
  • The elemental attack stat does not affect skill damage, even when there is an element advantage.
  • It appears as if elemental defense stat does not affect skill damage regardless of element.
  • In the case of element advantage, damage or elemental damage is doubled relative to neutral. In the case of element disadvantage, damage or elemental damage is halved relative to neutral.
2 Likes

Unsure of notation used but, for below,
An = Attack(normal)
Ae = Attack(elemental)
Ac = Attack(cumulative)
EM = elemental multiplier
It sounds like the elemental multiplier is applied to both An and Ae the same way. This sounds strange to me as it seems to indicate there is no reason the two forms of A should be separate at all. Given Ac = An + Ae, if EM is to be a coefficient for both An and Ae why would they bother having them separate vs just having EM affect Ac directly?

Addendum: Perhaps it’s because Ae does not affect shooting of barriers? Speaking of, this is low priority but I’m curious if there is a multiplier ‘BM’ for how Ac (or An) affects barriers. Eg, panzer and baron seem to chew up barriers like nothing so perhaps this is done via a higher BM value

No, only Ae is doubled. Perhaps I should have been more clear about that. That makes two differences, though. Ae is not affected by armor, An is not affected by elemental armor or element.

I haven’t investigated shields yet, but yes, they do affect damage.