Proc Value Calculation

M0R_Gaming

Does anyone know the formula to calculate proc set damage? Ive managed to find a formula that works decently, but it still has massive errors in it. (Take original value, divide by 5000, multiply by current weapon/spell dmg)

relentless_turnip

Surely you take the proc sets original tooltip and divide this by 5478 as they have told us this equates to 100%. Dividing it will tell you what 1% is.

Red mountain is 12840 / 5478 = 2.34392113

You can then times this number by whatever your SD/WD is. So 8000 WD would make this 18751. I think that's correct.

M0R_Gaming

relentless_turnip wrote: »

Surely you take the proc sets original tooltip and divide this by 5478 as they have told us this equates to 100%. Dividing it will tell you what 1% is.

Red mountain is 12840 / 5478 = 2.34392113

You can then times this number by whatever your SD/WD is. So 8000 WD would make this 18751. I think that's correct.

You would think that, but its slightly off. For example I used the proc value of Vateshran maul, and that division is 1.67, but the proc damage per weapon damage at 1k weapon damage is 1.80, and going all the way up to 6425 weapon damage, the proc damage is 12109, and the ratio is 1.88


When I did the same with Doylemish, it also deviated from the expected.

M0R_Gaming

I collected a bit of data regarding the vateshran maul, and plotted it. This is the results of it. At lower values it follows a linear trend, but at higher values it becomes slightly curved

Sluggy

Not sure of the formula but it is definitely linear (which makes sense as that would be the cheapest calculation to do). My guess is that it's set-specific and probably has a min value set for the baseline of the given stat and a max value for the 'effective' maximum value of a stat and then just scales by that factor. This way it could be a simple floating-point number used as the scaling factor that is stored on the db with each set. This is totally just me guessing though.

M0R_Gaming

It makes perfect sense that it would be just linear, but that's why I'm confused that it deviates from the curve at higher values. It could quite possibly be an effective maximum that is placed, but then shouldn't that curve the graph downwards and not upwards at values past 5.5k?

gariondavey

I love you M0R

dhoward5b14_ESO

Above 5478 there is a curve...

- At 5000 weapon damage ,129 weapon damage = 2.44% set damage tooltip.
- At 5779 weapon damage, 129 weapon damage = 2.05% set damage tooltip. I thought there was a curve when over 5478 (I tested two different weapon damage over 5478) but based on what others say maybe I miscalculated (they say it is linear).

Urzigurumash

Just wondering @M0R_Gaming , for your Vat 2h data, did you have the same amount of Stam for each Wpn Dmg value, or did your Stam decrease as your Wpn Dmg increased?

M0R_Gaming

Urzigurumash wrote: »

Just wondering @M0R_Gaming , for your Vat 2h data, did you have the same amount of Stam for each Wpn Dmg value, or did your Stam decrease as your Wpn Dmg increased?

I tried to keep it as consistent as possible with just weapon damage, but either way it shouldn't effect that much since I used the values displayed on the character sheet for weapon damage, not any calculated stuff.

Stam didn't decrease at all though, the pieces of armor had stam glyphs on them so that slightly affected the weapon damage, but it should be included in the data collection.

neferpitou73

dhoward5b14_ESO wrote: »

Above 5478 there is a curve...

- At 5000 weapon damage ,129 weapon damage = 2.44% set damage tooltip.
- At 5779 weapon damage, 129 weapon damage = 2.05% set damage tooltip. I thought there was a curve when over 5478 (I tested two different weapon damage over 5478) but based on what others say maybe I miscalculated (they say it is linear).

Hmm that's strange I've done it for two different sets and i get a straight line all the way through with an R^2 of 1. What class were you using for this out of curiousity? What kind of weapon?

katorga

Well, doing a sliding scale rather than a linear scale would not be "performant".