The formula is correct because it calculate the average damage after X seconds, and being WF having 20% chance proc it means that in average you get 1 WF each 5 swings, and that is what the formula consider.

Look again at the formula and tell me if that states 2 WF inside a 3 sec window.

Look, noble effort and all, but you can check the enhance shaman thread for 2 very nice sims that are extremely accurate from theory-->in-game in their modeling of DW windfury mechanics. You've been registered on the forum since April, I'm astounded that you haven't seen these.
Enhance Shaman: The Collected Works of Theorycraft, Vol I

You're correct, slow weapons are better than fast weapons. However this is news to absolutely no one at this point, and in any case you're disregarding any potential benefits that fast weapons might have for purposes of Shamanistic Rage.

Parsing shows that Windfury is less than 20% of all hits for any achievable weapon speed that's dual wieldable. That amount varies slightly from night to night but is highly dependent upon weapon speed. If you're dual wielding 1.3 daggers there's no way in hell you'll see a 20% proc rate. Your formula is wrong for this reason. You aren't showing people something brand new here. There are two or three colossal threads on enhancement shamans, I'd really suggest reading them over a little more carefully than you have.

Try using a 0.1 weapon speed with internal cooldown of 3 seconds. You'll be wasting almost all of your WF procs because of the cooldown. I'm not math guru and I usually step back from these kind of threads but this is so obviously wrong that I can't help myself.

As noted, the issue here is the 3 second cooldown. To take it to an extreme to magnify the issue, imagine you have a .1 second weapon. You swing 30 times in 3 seconds, but only one of them can proc WF. That means your WF proc rate is not 20%, but 1/30 = 3.3%. A 3.3% WF proc rate means it only increases our DPS by (very roughly) ~6.6% by giving us two extra swings (32 instead of 30). If it were procing 20% of the time (as it would in theory if our swing speed were 3.0 or slower) then we get two extra swings per 5, or 7 swings per 5, a DPS increase of (roughly) 40%.

Edit: haha, beaten with the .1 second example.

That means that all was written already on that subject so no one else can write anything else on it? I hope that is not your idea or we were still with Newton idea of the universe.

See this is where your logic is wrong. WF has a 20% chance to proc when not in the 3 second cooldown. This will chance your formula. You base your 1 in 5 swings on the 20%, but the 1.5 speed weapon will have 2 swings with 0% chance to proc. You need to add those swings into your formula. Once the 1.5 speed procs once, then your average would need to be 1 in 7.

I see your point, now see my point (that can be wrong ofc).

Your: WF has 20% chance to proc when not in the 3 sec CD
Mine: WF has 20% chance to proc at each swing but if less than 3 sec are last from last proc then you don't get it.

I thought a bit about this. The reason that you want slow weapons is that slow weapons hit harder with windfury.

But this 0.1 mainhand would not waste a lot of windfury time off cd. Seeing as as soon as WF is off the cd, with a fast hitting weapon you will proc a new wf faster.

This is why 2.9 weapons aren't that good. If you Proc wf, you'll "waste" one swing with wf on cd, meaning the first possible new wf will be at 2*2.9s = 5.8s. (Not dualwielding/dualwield same-speed weaps.)

Equally, if you only have flurry for haste, couldn't your mainhand be as fast as 2.2 speed as long as it keeps a high dps, as this will ensure tight timing close after WF cd cools down?

2.2 * 0.7 = 1.54

(This is from a "without impeding dps greatly" viewpoint)

This would tie in with the DPS drop people see when hasting their mainhand past 1.5, and capitalise on always staying in the WF cd "Sweet spot" for your mainhand. My only concern is how the offhand will affect this, and I can't theorycraft too well right now, being at work and all.

I'm not very smart. I freely admit to my normal level of intelligence. That said, I know enough about myself to know when to ask for help. Smarter people than me have done the math on WF and no matter how they do it, it comes out the same.

You have a 20% chance to proc WF when not in the 3 second cooldown. That 3 second cooldown skews the actual proc rate. You're taking the 20% as gospel, when it's just not.

dragkhar, just because you want something to be true doesn't mean it is. WF cannot proc for 3 seconds after the previous windfury. Attacks don't 'queue' up so that as soon as you get out of the 3s cooldown, you immediately get another. That's why people save stormstrike for right as WF ends, because it causes more chances right at the 3s mark to try and get more procs. This is why your above statement of 'not using stormstrike' is kinda bunk. We creatively use game mechanics, and slow weapons, to maximize the AP contribution of WF weapon. Fast weapons were sexy when you could quad proc and had just about 0 internal cooldown. Those days don't exist anymore, and fast weapons just plain suck. They suck up WF cooldowns that don't hit as hard. So, when you're swinging some 1.4 speed weapon that's WF critting for.. 400, the guy next to you, using 2 2.6 speed weapons with the same DPS is critting for near 1000 or more. These differences add up, because every 3 seconds, he's getting more out of WF than you are.

And if I'm wrong, I'm sure Nite is going to kill me, but I'm reasonably sure that my brief stint as an enhancement shaman taught me this, at least