Math formulae

WoW Classic Mists of Pandaria Classic Discussion
1

Hello Everyone,

I’m a rogue player on Mirage raceway. Our guild currently is gearing up to hit T6 content. I do around 1.4k dps on tank spank bosses.

I’ve played wow for a long time and I’ve always had a problem with the general way procs and crit strikes are working in this game.

This topic has been debated many times before but it’s never been addressed properly.

The problem is that the critical strike and proccing math formulae contains a threshold number which is very biased towards the far edge number.

Imagine you have a robot that puts balls with 1s and 0s in a box. Generally we say that if you have 33% chance for something then every 3rd occurence will a 1 and the rest are 0s.

What Blizzard made is like the robot keeps putting in 0s and after a while it’s like oh damn there has not been a 1 in a while so it slams in a bunch of 1s in streak.

This is the problem.

This is a problem for rogues, because deadly poision falls off and then ticks 3 times after each other. Problem for warriors as they drop deep wounds and then they crit 3 times in streak. Streak critting also wastes ticks of flurry. Problem for fire mages and any mage with a 4 set T5. They drop ignite and then they crit 3 times in streak. Or they overwrite the t5 bonus while other times it doesn’t proc for a long time.

Now, we can’t use a simple add up and even out like how combustion works, because it would allow huge abuses if you knew exactly when you will crit.

Therefore the formula needs to be adjusted so it would not produce such extremes.

I don’t want to shiv my deadly poision with 250 hit rating and an average 0.89 offhand speed during boss fight. It’s annoying beyond reason. I miss 4,5% off my offhand attacks and 1.8% is dodged on parried on Lurker. That means from 100 attacks I land 93. Glancing attacks afaik applies deadly poision as well. So from 93 successful attacks done in around 80 seconds a 30% deadly poision DOESN’T LAND FREQUENTLY ENOUGH to upkeep a 5stack.

The formulae are constructed that they would produce the written numbers on a long run e.g. 100.000 attacks yield perfect result if you have 33% crit it will produce 33% crit.

But this is bloody terrible on raid bosses. It is one of the factors why meele is behind the DPS numbers in T5. All meeles have a proc which is at some point generates wasted ticks and attacks. As I mentioned warriors, same for cement shamans. An overwritten flurry is terribly wasted while in the next 5 seconds you are just doing normal hits and no crits.

Thank you for reading.

3 Likes
2

dont get RNG and probability mixed up!

33% crit chance is on ALL attacks white/yellow(special) and its 33% on each hit so its not a guaranteed 33% crit chance over an extended period !

you can have 100 hits 0 crits , then you can have 100 hits with 50+ crits and then you can be back to 100 hits 0 crits , its called RNG calculator, dont rogues have the same as warriors (hamstring to proc flurry since the crit coefficiancy is higher on hamstring ?

Also ur example of 100.000 hits would yield 33% crit is not correct, in the world of RNG you can NEVER get an exact number !

im at 40% crit on my warr buffed and with wep enchant proc i can reach almost 50%, my deep wounds “never” fall off unless i have to move away for aoe dmg etc

10.2% hit / 31% crit / 10% expertise(minimize glancing hits)

i can see tough why you would think this way but sir you are wrong sorry

1 Like
3

Is that the shaman tank spec I’ve heard so much about?

2 Likes
4

I think OP is talking about normalized pseudo-random crit/proc firmulas like for example DOTA2 uses. Basically it’s not pure RNG but a bell curve with highest point at the expected statistical point of next proc.
Say you have 10% crit, the formula is constructed in such a way that after you crit, you have only very small chance to crit again, then it grows with each consecutive hit and is very hight at about 10th hit after last crit, and it grows further, to be all but guaranteed 15 hits in or so and 100% guaranteed 20 hits in. Of course, it is constructed precisely so that over large sample, you still get almost exactly 10% proc.
This is an excellent system, but of course would require complete rework of core game mechanics so it has exactly 0% chance of ever happening.
Edit:
Here is the description of “randomness” in DOTA 2 for example.
https://dota2.fandom.com/wiki/Random_distribution

1 Like
5

Not to be that guy but ether your guild sucks at uploading logs or you are trying to make yourself look better when talking about this, you have hit 1287,3 dps as your highest ever number according to warcraftlogs, and from the fights I see on your highest dps log you have a 86.3% uptime on deadly and eat it fairly often with envenom, you are not soft hit capped seeing as you actually had SS miss as well, there are some of your true numbers from logs and they dont realy match how you are trying to paint this picture.

6

You need keep S&D up 100% of time, combo your CD with bloodlust (haste pot, trinket, blade fury).
Don’t use evenom for some reasons:

  • Enevom consume poison,
  • is magical damage so have a 17% magic resistance chance
  • and you loose poison for some time so the difference with evicerate/evenom is lost just by loosing the poison).
    Best is use rupture (or IEA if you have to do it), which is improved by 30% with mangle

Actually you should be using it in ImpEA, and S&D.
IEA for your guild which is +2.57% on 7k7 armor and +3.21% on 6k2 armor for all physical damage.

Your DPS is not about how crit work, even with bad Crit RNG you could do better than this especially on fight like morogrim or void walker which are really like a dummy target 100% uptime; almost not strat, just by not doing envenom.

Remember on log you play with t4 level gear, and some do it with t5-t6 so you won’t parse orange against them (not only gameplay matter, group matter too, i see you don’t have feral; and gear too matter.
And i think then banned chicken now from warcraftlog but only for new logs, old log of SSC with chicken remain so gonna make it hard to compete in this raid.

As feral i have 52% crit once buff, got some log with only 38% crit rekt by RNG but still orange parse cause not a lot of mistake in rotation.
I know it’s a bit different on rogue cause more rogue than feral to compete, but you understand the point, it’s not only about RNG, RNG can make difference between a 90 and 100 parse; but not between a 50-60 parse to 90 parse…

Anyway you don’t control RNG, the only thing you control is your gameplay, and optimisation (gear/group/consumable) so better work on it than try make it change from blizzard.

1 Like
7

Warcraftlogs don’t really show my true numbers. I mean you may believe whatever you wish.

I use envenom because even with losing the deadly ticks for a time an envenom normal hit is around 2k and I can crit up to 4500. Improved evis+not losing deadly tick doesn’t match up.

I tested this on Hydross while not putting any points to either poision dmg or improved evis. So I had 4 unused points.

Hydros has a phase while poision immune and is a suitable boss to compare evis and envenom. For my measure envenom won by a long shot. If it crits over 40% it beats rupture as well.

I NEVER miss sinister strike.

I’m not ranting for the math formula being unoptimizied because I deliberetly consume my deadly ticks; I rant because there are times I’d like for it to stay up for eg. a long dot but it doesn’t apply the tick during the entire time duration of the poision and I have to shiv which I think shouldn’t be an existing problem if the formula would be optimized for smaller parse numbers instead of the large parse number it has now.

8

Warcraft logs are true not a random addon, seeing as they literaly use the log files exported from the game.

As far as I can tell from the log diving I have done on rogues from my guild and a few others your assesment on that is wrong seeing as you dont see any rogue with high rankings ever use envenom but thats not here nor there.

Hydross is immune to poisons half the fight so thats a bad fight to look at.

Rapture is seen on every highranking rogue in the world though, your math is clearly wrong.

I have literal proof in your logs that you actually do miss sinister from time to time though.

The formal is basically that you role a x amount of sided die every time you do a attack and if you role the % of crit or of applying your poison you do it, its not optimized for having it equal out over time its a chance on each attack and if you do a infinite amount of attacks you will most likely hit the % that you have but you can just be unlucky or lucky with a system like that, for a good example of how RNG actually works go to random (dot) org and put it up to 10000 and then try to hit at 3300 or below that and see how often it happens and how many streaks you actually get. A true random number generator can spit out 9 100000000 times in a row and you cant prove its not random from that so a sample size lesser than infinite would not prove if something is truly random or not.

2 Likes
9

Thanks for understanding my text and your comment is appreciated.

10

Mage don’t really play Fire in TBC (unless they don’t care about their damage), and 4p T5 is overwritten everytime because the probability to not crit for a long period of time is very low, so you don’t pay attention to that, you assume to have 100% uptime on it even if sometime it’s between 80-95. This is probably why Arcane is so good, because you have +30% crit when you have Clearcast proc (that often mean 60% total crit), drastically reducing the RNG factor of Mage damage that you can have in Fire tree. But then, you have a new RNG dance : Clearcast proc :slight_smile: .

RNG is RNG. The ability to have 70% crit in a 2mn fight with RNG, full proc at the best moment, is what made the game pleasant for lot of raiders because even if you are undergeared or play an underdog class, you can be #1 DPS if you have enough luck.

Yeah, it suck to have bad luck and do less crit than your crit rating, but it is like this. Next time, luck is going to be with you and you are going to hammer the damage metters. Our Rogue did 2k2 DPS on Archimonde last week, #1 DPS, no glaives. When everything align, melee can deal insane damage. It’s just a bit more RNG than caster especially if you are playing a dual-wield class.

Some proc have PPM to, making them more reliable.

The more longer the fight is, the more likely you are to have the crit of your crit rating.

1 Like
11

the reason that we dont envenom actually has nothing to do with the skill or our dps. we do it to not grief the elemental shamans, because we will actively eat stormstrike charges. envenom can do fat crits, 5,5k as my highest, but truth is that if you are doing iea and snd you will barely have combat points to do any other finishers

12

maybe you misread something? it can be dodged and parrid for sure, but rogues only need 4% hit from gear for it to never miss, so i doubt that it “missed”, even if he avoided hit gear like the plague he would be yellow hit capped

1 Like
13

I’ve literally never heard of a rogue using Envenom, and with the sweaty parseboys playing nowadays they would not care whatsoever about stormstrike charges and being nice with their fellow elemental shaman players, but you do you :slight_smile:

14

Warcraftlogs don’t really show my true numbers. I mean you may believe whatever you wish.

Warcraftlog show true number.

I use envenom because even with losing the deadly ticks for a time an envenom normal hit is around 2k and I can crit up to 4500. Improved evis+not losing deadly tick doesn’t match up.

I tested this on Hydross while not putting any points to either poision dmg or improved evis. So I had 4 unused points.

Hydros has a phase while poision immune and is a suitable boss to compare evis and envenom. For my measure envenom won by a long shot. If it crits over 40% it beats rupture as well.

You know poison is magic, it can be resisted by boss since you don’t have spell hit you have 17% miss chance on it,
Same happen with envenom finisher 17% miss chance, and that not cause you don’t see it on a full fight (cause you don’t send enough envenom to play with RNG) that it should never happen.

But there is one thing you don’t check it’s the damage per energy ratio, envenom show nice big damage but cost 35 energy while rupture cost 25.

So you need to compare 4 envenom to 4 rupture + 1 sinister strike.
And i don’t even count the dps lost from removing poison in this equation but poison do 277 per 3s according to your log, so at 5 stack if you take 10s to refresh to 5 poison you loose 3 tick of 277 damage which is a lot.

And i only speak about raw damage, idon’t count the extra point this extra Sinister strike give you cause with 5 siniter strike you do extra rupture too, and we have to count it too.
And i don’t count the extra energy from talent “rentless strike” that come from this extra rupture that come from extra sinister strike you do by not using envenom.

And you can check on your log 4 rupture + SS is better than 4 envenom just with raw damage :

https://zupimages.net/viewer.php?id=22/07/xwpq.png

So now you understand why the top rogue don’t use envenom
And for formula coefficient, envenom use 15% Ap at 5 pt, while rupture use 25% at 5pt.

I NEVER miss sinister strike.

You don’t miss SS if you have 9% hit, but still can be dodged, the final talent of combat spec only work for finisher (rupture, KS, expose armor, envenom, eviscerate).
Mean you can use a KS on a rogue under evasion even if he face you, but Sinister strike will be dodge if he face you.

I’m not ranting for the math formula being unoptimizied because I deliberetly consume my deadly ticks; I rant because there are times I’d like for it to stay up for eg. a long dot but it doesn’t apply the tick during the entire time duration of the poision and I have to shiv which I think shouldn’t be an existing problem if the formula would be optimized for smaller parse numbers instead of the large parse number it has now.

As i told it for envenom, poison is magic and can be resisted, on apply and on tick.
You can check on your log you often have deadly poison resist tick.
And when resist happen for application, if do it too much in a row can happen you loose those 5 stacks.

For exemple on the log of the screenshot up there, you did a total of 586 melee, looking for you weapon one if 1.4s an one 2.6s and poison is only left hand.
So with 1.4s off hand and S&D you should have a 1.07 attack speed offhand and a 2s main hand attack speed so a ratio of almost 2 offhand hit for 1 main hand hit.

so 2/3 of 586 is 390 auto from off hand (or something close), you got 109 (if i counted well) proc poison, which is 27% proc slightly under 30% but not that big of a problem), and 9 of it was from shiv.

You can check it here:

https://fr.classic.warcraftlogs.com/reports/9v4G3btM1p2R6fXy#fight=44&type=auras&source=26&view=events&spells=debuffs&by=target

refresh debuff is when you apply it, and remove debuff when it fade, anyway never let it fade better shiv instead sinister strike if about to fade (look like you do it already).

2 Likes
15

Hmm you are right I never considered the damage / energy ratio, I only checked raw damage. But most of the time I do upkeep rupture and only envenom if I can keep rupture and snd up steady.

I also have to admit I didn’t consider the shaman stormstrike.

So my envenom is not to replace rupture but evis.
If I lose 3x277 (831) damage on deadly ticks then it comes down if envenom crits or no. If it did crit then it was worth it, if not it is damage loss.

I also appreciate how you analyzed my stats so yea, thank you.

16

Only time when you can use envenom is when boss is about to die and you have 5 cps to spend.

17

Yeah man, I know how you feel. Warcraftlogs get my DPS wrong all the time too, it’s actually 400k on all bosses.
I’m actually parsing 101% on all bosses because I’m so good the logs just break.
Sad life for us god players man :frowning:

Or, y’know, maybe the low DPS is from the fact that you’ve got 60%-80% uptime on SnD, don’t combo CD’s with Lust/Hero and things like that.

1 Like
18

I believe the game will only allow 33% crit so any crit stats you have above 33% are more or less wasted stats.

19

no its not

20

I don’t mean that the logs are wrong but not all my raids are uploaded. So it’s not consistent.

According to details my snd uptime is normally between 83 and 86%.

I don’t use my on-use trinket or my cds on hero, I use it at the very beginng, it notmally allows me to use them at least 1 more time throughout a fight compared to players who use them first only on heroism.

Your ironic tone is not required.

Thanks for your thoughts though.