Let's say I'm just attacking with one weapon with Windfury. Does everyone agree that two different weapons with identical dps with 1.5 <= speed < 3.0 will generate the same amount of dps from windfury?

Edit: sorry for rephrasing this about a dozen times

Not really. Any weapon that once flurry, haste effects, berserking, mongoose, etc taking into effect where the hasted speed doesn't drop under 1.5 should have the same dps. So a 2.8 weapon compared to a 2.6 weapon shouldn't be much different at all. Of course that doesn't take into effect the second weapon and that you want to arrange things to have the MH proc as much as possible but the OH as little as possible.

No. Slower weapons will generate higher DPS, all other factors being equal.

(still true after your rephrasing)

Am I ignoring flurry ? Then look at post 36, look at the simulator combat log and tell me if you see something wrong, like WF more ofthen than 3.0 secs or Flurry not procced correctly.

About having created another thread, I didn't see only 9 threads on this forum, one for each class, so realy I don't get why creating a thread for a so specific matter is disturbing you.

Also I do believe you posted your consideration against my works without even have read the thread, I'm not simulating a DW shaman at the moment, the simulator is new and I will post it on google code for revision and or suggestion.

I'm showing what I found (that can be different from what other are claiming) that's the way mathematic models evolves.

No, this is false given current testing. If the WF cooldown did not exist, then your statement should be true.

In fact, I am wondering where this explanation is going from the OP. Perhaps it is a communication problem, but everything I am seeing seems to be saying that the OP believes his result are correct despite direct testing evidence that indicates otherwise. Testing should come before simulation, so it would seem we have a cart before the horse situation here.

Edit: My above explanation was assuming dual wield. So you're not talking about DW? So axe and shield then? But you are ignoring crits and flurry or aren't you? I'm so confused. Beyond being confused, I see this line of modeling to be fruitless. I can't imagine a serious situation in which these assumptions would be met.

By assuming 100% every fifth swing, rather 20% on individual swings, you're already wrong in not just class mechanics, but statistics in general. Having a boy for your first kid does not mean your second one is guaranteed to be a girl. Stastistically speaking, over a large sample, the numbers are roughly 50-50. But you cannot assume the first one to be a boy, and the second one to be a girl just because it fits the model. If you want to simulate things, you need to do things intended by nature of the mechanics. In this case, 20% on individual swings to proc windfury, not zero zero zero zero 100.

Is it going to make simulating harder? Yes. But it will make the model much more reasonable.

This specific matter, enhancement shaman DPS and windfury, is already covered extensively in another thread that hasn't fallen off the front page of this forum in weeks.

Then it's pretty useless because theory and testing both show that using a two-handed weapon is vastly inferior to dual wielding, even in the case of the ungodly powerful mace used during the Kael'thas encounter.

This (along with the fact that you are completely ignoring dual wield) is exactly why everyone is saying your simulator is wrong. This has not proven to be the case in testing.

That is what this graph confirms ( see the line with 0% crit )

Okay, we've gone down the road of whether or not he's checking for WF cooldown or what have you, and he says he is.

However, this doesn't negate the simple fact that if two weapons are equal DPS, and one's 3.0 and one's 1.5 speed in a perfect world, the 3.0 will do better DPS on an enhancement shaman for the simple fact that every 3.0 seconds, it hits more than twice as hard on the WF double swing. It is no contest, there is no way you can model it to not work that way, it is a simple fact that 3.0 weapons have a better damage range and benefit from AP more in increasing that damage, so the conversation really is moot.

If I'm interpretting what you're saying correctly, there's no way this can be true. While windfury would theoretically proc the same amount for weapons in the 1.5-3.0 range (under the effects of flurry), a slower weapon will still hit harder with WF procs, making it higher overall dps.

Here's my math. Take two weapons, tested independently.

Weapon A: 1.6 speed, 160 damage, 100 dps

After a successful WF proc, there's a swing within the cooldown where WF cannot proc again. The expected number of swings until you get a second Windfury proc is 5 - not including the one that can't proc - so, on average, you'll have to wait until 9.6 seconds after the first proc for it to proc again. Therefore, with this weapon, you'll average 60/9.6 = 6.25 procs per minute. The weapon strikes twice for each proc, so we have 2*160*6.25 = 2000 extra WF damage per minute.

Weapon B: 2.8 speed, 280 damage, 100 dps

After a successful WF proc, there's a swing within the cooldown where WF cannot proc again. The expected number of swings until you get a second Windfury proc is 5 - not including the one that can't proc - so, on average, you'll have to wait until 16.8 seconds after the first proc for it to proc again. Therefore, with this weapon, you'll average 60/16.8 = (25/7)procs per minute. The weapon strikes twice for each proc, so we have 2*280*(25/7) = 2000 extra WF damage per minute.

Of course, I have not factored in the bonus attack power . . . naturally, the faster weapon procs more times per minute, which translates to more bonus attack power hits per minute, which translates into somewhat higher dps (I think?). If you ignore the bonus attack power, though, any speed such that 1.5 <= X < 3 should generate precisely the same dps from WF.

What's my error?

You're also ignoring the fact that the base WF damage takes AP value based on the weapon speed as opposed to being a flat value like Fiery, and it isn't normalized like MS or SS. The 2.8 speed has 1.2*AP more contribution per hit than the 1.6 does, which results in higher damage per WF proc.

The bonus attack power from WFW is a wash between the two since they have the same proc rate.

Think about that again. A slower weapon makes harder WF but slower weapons has more WF proc (and yes I'm aware I can not have 2 WF inside a 3 seconds window). See my very first post and check that 2 cycles with 2 different weapons.

How is this thread 3 pages?

The mods are sleeping in.

Yeah, you're right. Just checked, and my last paragraph is wrong. So - forget about the bonus attack power. Why isn't every weapon between 1.5 and 3.0 exactly the same?

I editied my post, I apologize for not being clear initially. The short answer is that WF procs are not normalized, so in a situation where you have a similar proc rate (1.6 vs 2.8 both assumed to be 5 PPM) the slower procs are going to have a higher damage contribution from AP and thus hit harder, which results in higher overall DPS.

In contrast an ability that is normalized like Mortal Strike that is, essentially, 10 PPM will recieve the same contribution from AP whether the attack speed is a 1.6 or a 2.8.

Nice catch.

In the formula I wrote the WF bonus takes care of weapon speed but still it doesn't matter take a look at it again. (Consider the formula valid with 0% crits ).

In the simulator because I was in hurry I supposed the WF damage not increased in function of weapon speed and that can make the difference. So when the target gets a WF event it doesn't increase the base weapon damage.

For today I can not fix that, but I will for sure tomorrow and see how it changes.

A question I have: how much big is a time tick ? What I need to know is:

Speed 2.7 with -30% due to flurry becomes: 1.89, is blizzard using a 0.01 tick time so the weapon swing exactly at 1.89?

Yeah I forgot about the lack of normalization. That means that yeah slower is better. Now of course its more complicated than that because faster does mean its possible to waste less time after you hit the 3 second mark, but then again offhand gets thrown into that as well. All in all it is extremely complex. But still slower harder hitting weapon means more dps.

Please re-read my post. WF is a flat 20% proc chance per hit on a three-second cooldown. Extrapolating from this, you don't get "1.6 vs 2.8 both assumed to be 5 PPM" - you get that the 1.6 speed weapon is at 6.25 PPM, and the 2.8 speed weapon is at 3.57 PPM. The faster weapon hits more often, for less; the slower weapon hits less often, for more - and they both generate the same dps from WF. Where's my math wrong?

Is not wrong indeed, that is inline with my tough

Here's a more generic example: Take a weapon with speed S and dps D, where 1.5 <= S < 3.

-Expected lapse between WF procs is exactly 6S
-Expected procs per minute is exactly 10/S
-Expected WF damage per minute (ignoring bonus attack power) is exactly (SD)(10/S) = 10D

So, when wielding a single weapon of speed 1.5 <= S < 3, WF dps is a linear function of weapon dps.

Edit: Please note that I am making no claims or suppositions about dual wielding. I just want to make sure that everyone is in complete agreement with the simplest case (one weapon, no flurry) before I move on to that. Again, I'd like to see a criticism of my math on this.