[Shaman] Windfury damage still appears to be low

[honuk]

A couple tests ago there was an issue reported where offhand windfury damage appeared to be calculated incorrectly, causing shaman dps when using double windfury on two slow weapons, which was the optimal dps set up for enhancement shamans, to fall behind windury/flametongue with a fast offhand weapon. The report for this bug is here: https://github.com/ccshiro/cc-buglist/issues/145. The forum thread on this issue is here: https://www.worldofcorecraft.com/content/enh-shaman-windfury-weapon-oh-damage

This test, the values appear to be closer, but WF/WF is still behind WF/FT by around 100 dps, using the same testing method explained in that thread. While using WF/WF, windury damage accounted for approximately 35% of my total damage over a ten minute test involving only auto attacks and stormstrike usage. Using WF/FT, windfury accounted for 32% of my damage. Something has to be wrong for that to be the case.

[Barroth]

keep in mind, that WF is physical so unlike Flametongue it is affected by the 50% reduction through armor. Could be different in a raid environment with armor reduction debuffs up

[honuk]

but the primary tested difference between slow/slow WF/WF and slow/fast WF/FT is not the FT damage, it's the WF damage. I am getting nearly as much WF damage with WF on one hand as I am with it on both. hell, I could go with no enchant on the offhand and the overall dps would almost be as high as WF/WF

[honuk]

I just now tested WF/WF vs. WF/Nothing on two slow weapons. This test, like all my others, was on a level 70 target dummy, with no totems, no buffs, and no shocks. All I did was auto and stormstrike on cooldown for 10 minutes for each setup. WF/WF came in at 720 dps and WF/Nothing registered 645. And this is with my WF/N test getting a 5% lower WF crit rate than my WF/WF test. WF accounted for 38% of my damage in the first test, and 35% on the second.

[honuk]

one final test to confirm the suspicion the last one would raise: WF/FT is more dps than WF/WF even with two slow weapons. same ten minute test, same weapons, and WF/FT was at 760 dps. there is no situation in which this should be true in TBC.

[Philosophers]

You keep in mind that off hand weapons deal less damage? And as long as damage from WF is physical it is affected by damage reduction of off-hand weapon (and armor), when FT is fire - unaffected. From what I can read it can be the answer.

[honuk]

of course offhand WF does less damage than MH windfury (though the bug report I referenced above had to do with offhand WF dealing less damage than it should). that does not change the fact that in TBC WF/WF does more damage than WF/FT, especially when you are using two slow weapons. whether the problem here is proc rate, damage coefficients, or some issue with the internal cooldown on WF, I cannot say.

[Philosophers]

It is complex problem, but I can provide some explanation. In early TBC mixing WF rank "n" and WF rank "n-1" made both weapons proc on separate cooldowns (which is 3 seconds. This is time when WF superiority was out of discussion). In one patch (I believe it's 2.1.0) Blizzard changed it to have one cooldown. Since then you can proc WF maximum once per 3 seconds. In practice it is also lowered by proc chance.

Let's say we attack with 2.0 speed all time, WF proc is 100% chance, one weapon. It means that: 1st attack is WF 2nd attack is not WF 3rd attack is WF etc.

That means when you have 100% proc chance you will get only 50% in practice. With 20% chance it mean that every WF attack there will be one hit you can be 100% certain it won't be WF. That means you don't actually have 20% to proc. I count it would be circa 16%

With dual wield it looks like this: 20% on both weapons, only one at time, 3 seconds hidden CD, gives us about 36% chance to proc at a time with clear CD, what gives us about 20% of total attacks will be WF.

I know it's long and can be confusing, but my point is: when you have WF on offhand weapon it proc takes part of the main hand proc. That means main hand has lower number of WF attacks. Problem lies also in stormstrike, that is not normalized attack. Not normalized that means it benefits more from weapons with higher top damage, in simple words: from slower weapons.

In my text can be mistakes, I am too tired to re-read it and correct, sorry. Just ask questions if anything seems wrong or unclear.

[honuk]

I am very aware of TBC WF mechanics. Even after 2.1, when they fixed the off rank WF bug and kept the internal cooldown, WF/WF slow/slow was still better than WF/FT slow/fast, to say nothing of WF/FT slow/slow, which is objectively worse than FT on a fast weapon. WF on your offhand eats MH procs, but the overall proc rate on WF goes up when both weapons are imbued. while dual wielding WF weapons, any attack that occurs outside the 3 second WF internal cooldown has a 36% chance to proc WF. this number is 20% if only one weapon is imbued. double WF also leads to a higher uptime on flurry and unleashed rage. in fact, I would bet that most of the DPS difference between my WF/WF and WF/Nothing tests is because of flurry uptime and nothing else.

[honuk]

I simmed a bunch of tests using the old TBC enhancement shaman dps sim EnhSim to use as a comparison. Several things to note before I post the numbers: 1) all of these sims were against bosses, while my tests on CC were against the lvl 70 dummy. I used the 70 dummy on CC to filter out as many misses and dodges as I could to give me more numbers for the test. I don't have the stats of the level 70 dummy to plug in to this sim, so I used its default boss values, which have 7700 armor and no effective resistances. 2) These sims were set to 2000 hours each, while CC tests were 10 minutes each. Unavoidable. 3) the character sheet stats on the sim are an approximation of those on the CC test. Most were inputed directly while testing a few tests ago, so they are very close, with the exception of: 4) my "fast" offhand test used the same stats and dps as the "slow" offhand, with the speed simply changed from 2.6 to 1.8. In other words, the fast offhand is a 1.80 Netherbane, which doesn't actually exist.

Anyway, numbers. I ran 12 tests. The first 6 are cover WF oh, FT oh, and no enchant oh with both slow and fast weapons all against a boss with zero buffs/debuffs, including no totems and no shock usage. The last 6 are the same setups against a boss with every possible buff/debuff that the sim offers, including shock usage and totems:

No buffs:

wf/wf slow/slow: 611.89 (wf 37.36% of dps) wf/ft slow/slow: 582.59 (wf 22.55% of dps) wf/n slow/slow: 508.10 (wf 25.91% of dps)

wf/wf slow/fast: 576.79 (wf 35.11% of dps) wf/ft slow/fast: 580.01 (wf 22.58% of dps) wf/n slow/fast: 498.21 (wf 26.23% of dps)

Full Buffs:

wf/wf slow/slow: 1833.64 (wf 30.28% of dps) wf/ft slow/slow: 1703.49 (wf 19.43% of dps) wf/n slow/slow: 1592.80 (wf 20.76% of dps)

wf/wf slow/fast: 1748.07 (wf 28.07% of dps) wf/ft slow/fast: 1700.59 (wf 19.35% of dps) wf/n slow/fast: 1569.52 (wf 21.02% of dps)

You'll note that the only time WF/FT is ever better than WF/WF is with a fast weapon and no buffs whatsoever against a boss, and even that advantage is small enough to be within the margin of error. The second thing worth poiting out is that WF counts for a substantially larger pecentage of total damage done with WF/WF than it does with any other combination, and that this is true in every scenario.

[Ickler]

I made a T5 geared shaman, and I was keeping 850 DPS on dummies. There is no way that's under what it should be, remember that physical classes benefit insanely much from a raid setting, with both raid debuffs and buffs... Armor reductions and AP modifiers.

And by doing so btw, WF was nearly 40% of my damage done.

[honuk]

and if you used FT, you would have done more damage. what's your point? there is also no way you were getting 40% of your damage as WF if you were also shocking, and if you were maintaining 850 dps you were either using shocks, or your test was way too short.

[Ickler]

I was using shocks, but as you say my test was not very prolonged. I ended just before the third Shamanistic Rage. I'll retest next test, and make it much longer + focus more on the proper rotations. But by the DPS numbers, I would say Enh is more than fine. If FT does more, perhaps FT scales too well?

[honuk]

If FT scaled too well, tests would show FT making up an abnormally large percentage of overall dps. Instead, WF/FT had a much higher percent of damage coming from WF on CC than one would expect, which leads me to believe that either 1) something is bugged with WF OH damage, causing it to eat too many WF procs and/or still hit for less damage than it should, or 2) WF is proccing too often when it is only imbued on the main hand. The latter possibility is very easy to test, but alas, I did not take a screenshot of the number of WF hits in any of my tests.

[ccshiro]

@honuk Can you share that simulator, please?

It would also be nice if next test some of y'all can do some testing and see if we can figure out what exactly might happen to make it go wrong (is the off-hand eating more procs than intended, for example?)

[honuk]

Here's a link to the sim I used: https://code.google.com/p/enhsim/downloads/detail?name=EnhSim0.2.7.rar&can=2&q=

This is the last version of that sim released before 3.0. It doesn't have a GUI. You just edit the numbers directly in the config file and then pass that file into the sim.

I'll certainly be messing around with WF tests next test. I'll see if I can scrounge up some other shaman support too.

[honuk]

whelp, I've been testing for a few minutes on this latest test and if I haven't found the problem, I've certainly found a problem! windfury (and stormstrike!) don't have a crit coefficient. they "crit" as they should, and their crits appear to proc flurry, again as they should, but they don't do any more damage than a regular hit. I have some recount screenshots demonstrating this that I'll post later. just wanted to get this out there.

[honuk]

turns out the no crit thing only happened early on in the test. when I logged on later WF and SS were critting normally. if you guys didn't hotfix, I have no idea what happened.

[ccnim]

We did a hotfix for the issue.

[Amarok-bt]

maybe some of these informations will be valueable: "Windfury Weapon

• Scales with weapon DPS, attack power, melee haste, melee crit, melee tohit.
• Mitigated by armor and block, avoided by dodge, parry, and miss.
• Has a 5% special attack type melee miss chance.
• Benefits from weapon mastery for 10% more damage.
• The best enchant for enhancement shamans, on both hands.
• The two extra attacks can proc flurry, unleashed rage, and shamanistic rage. They do not consume flurry charges, and don't glance.
• The two offhand attacks are affected by the standard offhand 50% damage reduction. The WF proc's attack power bonus is not. "

[Escape243]

edit: found the boss dummies

In the gear character: Napp test account: TEST2716 is currently equipped, simulated DPS was considerably higher than what I was seeing on the test server. My sim config file: https://drive.google.com/open?id=0B6pi9KPNdc_AckRXX2ptaXZDMEE (config file is set for the first sim, meaning stormstrike is turned on)

First Sim, elected not to shock, used stormstrike on CD, WF enchant MH&OH Simulated DPS: 606.29 Recorded DPS: 524.8 (aprox 10 mins) http://i.imgur.com/M7R4Z3s.png (shows simulated data and recount data)

Second Sim, only white hits + WF enchant MH&OH (I couldn't turn of SS completely in the Sim so I set the cooldown to 10 billion seconds which means we will see no stormstrikes after the first) Simulated DPS: 506.90 Recorded DPS: 435.7 (aprox 10 mins) http://i.imgur.com/mT3gThM.png

My guess is the disparity has something to do with WF's damage but it is hard to say.