Elemental chances test results

So there was this little conversation started on the Nisha Top Gear thread about elemental effects proc chances when you matched elements.

Basically some people (Mainly @Aite and @akdere1) said that if a fire gun for example, had 20% chances to ignite it’s target, the chances would rise to 40% on flesh (Because fire does twice the damage to flesh in UVHM) and would fall to about 5% on shield and Armor (because fire on those is only 40%)

I initially laughed it off (Like the arrogant SOB that I am) but they insisted I at least give it a try…which I reluctantly did, if only to prove them wrong so they would shut up about it and once again prove my superiority through witty rhetoric…

Turns out that my initial testings revealed that there might be something to it as my results seemed in line with their claims. So I immediately started screaming about it like an amphetamine-crazed Isaac Newton wannabe. After @Sljm wisely recommended that I calm my tits and reminded me of the fickle nature of RNG (And of @HeyCarNut’s brillant post on the subject) I set out to find the truth using much more reliable (and tedious) testing methods.

This is a thread about them.

##Methodology

  • Strip naked
  • Turn on my playstation

…wait, that’s not right.

  • Turn on my playstation
  • Strip my CHARACTER naked (removing all gear except for the guns I’m testing)
  • Turn BAR off
  • Reset skills
  • Only spec into snap shot, Bona Fide Grit and Fistful of bullets (so that I won’t have to reload as often… it’s already long as it is.) They shouldn’t affect the results in any way.
  • Get a hold of one gun of each element: Snipers, non-Maliwan (because they have splash) Snipers were chosen because they have high elemental chance, so the results would be more significant. (I could have used railers too but I found 4 snipers first)
  • Shoot at the flesh dummy’s torso with the shock sniper
  • Register each time an elemental effect proc’ed
  • repeat until all 132 bullets are gone
  • save/quit then reload character (to reset RNG seed)
  • repeat twice (so 396 shots in total)
  • shoot 4 more time to satisfy my inner OCD
  • Repeat whole process with the shielded and armored dummies
  • Complain that it’s very long
  • Repeat WHOLE process with a fire, corrosive and cryo sniper
  • Post my numbers here so we can discuss them

I will edit this post with new results as I generate them (I’m not doing this in a single night… I have friends and responsibilities)

##The actual numbers

Number of procs:

Shock 30%
On flesh: 86/400 = 21.5%
On armor: 83/400 = 20.75%
On shield: … 400/400 = 100% …I’m as puzzled as you are :flushed:

Fire 30%
on flesh: 167/400 = 41.75%
on shield: 98/400 = 24.5%
on armor: 103/400 = 25.75%

Corrosive 30%
on flesh: 96/400 = 24%
on shield: 80/400 = 20%
on Armor: 168/400 = 42%

Cryo 34.5%
on flesh: 108/400 = 27%
on Shield: 31/400 = 7.75%
on Armor: 98/400 = 24.5%

it does look like there is substance to the idea, but so far, I’m getting well below 30% on both flesh and armor, and WAAYY more than 30% on shield. In fact, even if it was linked directly to elemental modifiers, 100% procs is above the 90% chances we should be seeing if it was the rule (shock is 3x damage on shields in UVHM).
Edit: Shock is x2.75, not x3.

Discuss :slight_smile:

8 Likes

Just to make sure, did you stop shooting after you see the DOT proc and resume shooting just when the DOT ends?

Kidding :stuck_out_tongue:

4 Likes

Yes. I also paused between each shots to let recoil correct.

Which makes me dread testing corrosive…
I think I’ll do that last :wink:

Shock is 2.75 on shield in UVHM, not 3x

Expected rates | Found rates
.3 | .215
.3 | .2075
.825 | 1

The Shock application is pretty wild. The idea that the multiplier is working on the 30% application rate, boosting it to 82.5%, does seem to fit, but… 400/400 is a pretty f*cking lucky streak. Incendiary and Corrosive testing will help that make sense. If they’re higher than 60%… Well, let’s not get ahead of ourselves.

The ~21% application rates are pretty low, but it’s on the bottom end of what you could expect in an “unlucky” random set of that size at 30% rate. I’d say that’s close enough to be confident that the card’s not explicitly lying. It wouldn’t surprise me at all if your Cryo rates ended up +/- 10% on the 1x multipliers. It’s pretty interesting that they were so close to each-other and both consistently ~-9%, though. But that might be coincidence. The Shock numbers being that absurdly high sort of rule out “DOT resistance” or something.

If they’re low for some other reason, we’d need more data anyhow. I can’t think of a reason. If both Cryo tests against 1x multipliers ended up ~10% low like the Shock ones, I’d think something else might be up.

Keep the numbers coming, this picture can only get clearer.

Quick reference for readers:

1 Like

Good one, Chuck. This makes me see the cryo long nail in a different light for applying cryo!

1 Like

Good job, @Chuck80! I’m looking forward to see the results of your cryo testings :wink:

What about slapping on a +elemental chance COM and see if the numbers scale, btw?

Good idea
Though I’ll save that for last.

One idea that could explain the 9-10% loss in procs could be a sort of “soft cap” like for fire rate.

That only exists because of fps limitations.

I know… but I have no other decent explanation to propose yet other than that

can you put the percentages behind those numbers?

86/400 = 21,5%
83/400 = 20.75%

Ok, Fire is up

30% chances here also

Fire on flesh: 167/400 = 41.75%
Fire on shield: 98/400 = 24.5%
Fire on armor: 103/400 = 25.75%

It seems to me that while a good element match brings more chances to cause an effect, a bad match seems to have no effect.

There may be some other thing at work underneath, something more simple than the elemental table values… something like “good match = double chances (with shock on shield considered “doubly good”)”

Also, the slightly lowered chances compared to what’s expected might be due to the fact that both my shock and fire snipers were level 69… I just picked the first ones I saw and didn’t pay attention to that, that was careless. I don’t want to redo the tests (and since it took me 4 freaking hours to find a non-maliwan shock sniper, I don’t want to search again)

My corrosive and Cryo snipers are both level 70, so if there is a slight difference upwards in the next tests, it might be the cause.

Moving on to cryo now… Back in a few hours.

I doubt the gun’s level has anything to do with it.

Cryo is done! (this is boring)

My above theory has been squashed

34.5% chance on the card

Cryo on flesh: 108/400 = 27%

Cryo on Shield: 31/400 = 7.75%

Cryo on Armor: 98/400 = 24.5%

The slight lowered values compared to what was expected seems to be a constant.

and it’s always 30% proc chance on the card?

Yes so far.

I think my corrosive one has 34.5% so we’ll see if it makes a big difference…later :stuck_out_tongue:

After many pints, eh?

1 Like

Goddamnit!
Big edit: the CRYO one is the one with 34.5%

My corrosive one is 30% also…

Bleah! I’m done with corrosive (and with testing for a bit)

30% chance on the card

Corrosive on flesh: 96/400 = 24%

Corrosive on shield: 80/400 = 20%

Corrosive on Armor: 168/400 = 42%

Seems like roughly the same numbers as the fire one, but reversed for flesh and armor.

##Preliminary conclusions:

1: I’m getting systematically less chances than written on the card, except when a few odd birds are concerned.

2: Most matches, even some of the normally bad ones, don’t seem to be affected and have normal chances (apart from the difference I mentioned above)

3: Fire on flesh and corrosive on armor have better chances, about twice as much

4: Cryo on shield has massively decreased chances, something like 1/3 as much

5: Shock on shield is MASSIVELY better, at least 4x as much… might be more but it’s hard to draw conclusions from a perfect score…

##I will have to do a few more things:

1: re-test shock on shield with a gun with less chances to cause effect, so we can see the ceiling.

2: pick one of the average matches (one of the many in the 20ish%, like shock on flesh or cryo on armor) and test it a lot more (like maybe a thousand times) to get a more precise percentage of the “base” chance. It’s probably somewhere between 20 and 25% for a gun with 30% written on the card…

3: Pick the same match and test it with a different %chance gun to see if the difference between the card and the results is fixed or proportional.