If the 3 second ICD is present on windfury weapon in TBC classic, is there an attack speed breakpoint where totem is actually better than the straight weapon buff? My understanding is the ICD only applied to the weapon buff and not the totem. If your AS is 1.5s with your mainhand after haste and flurry, a 3s ICD would disqualify 50% of your swings from benefitting from the weapon buff.
Napkin Math:
Windfury Weapon
AS: 1.5s
Attacks per minute: 40
Eligible Attacks per minute: 20
Average WW Attacks per minute: 4
Windfury Totem
AS: 1.5s
Attacks per minute: 40
Eligible Attacks per minute: 40
Average WT Attacks per minute: 8
It looks like WW is really only about 35-40% more damage than WT. If the average plays out over time, isn’t WT more damage?
No. The fact that Windfury Weapon gives 2 extra attacks when the totem just gives 1 kills that idea.
It would disqualify 50% of your swings if Windfury had a 100% procrate. Swings can only get disqualified if Windfury actually procs first. Due to this fact, your napkin math has a big hole in it. The percentage of swings disqualified is going to be much lower.
It might be 2 vs 1 on the tooltip, but it really is 3 vs 2. You have to account for the initial swing damage that is applied to both procs which is why WW is only 35-40% better and not 100% better.
It disqualifies 50% of your swings because with an ICD of 3 seconds, the maximum number of procs is 20 per minute. There’s no way to proc more than that, whereas the totem can proc off every swing.
Yes, it can proc 20 times per minute. Just because it can happen doesn’t mean that it’s the optimal DPS strategy.
Maximizing DPS is almost always about maximizing your average DPS, not your highest theoretical DPS. If you pull 600 DPS half the time and 1000 DPS the other half, you average 800 DPS. Doing a consistent 900 DPS every time is generally better, even though the high end of 1000 in the prior example is a bigger number.
Look into your example closer. You average 4 procs per minute, but you’re estimating 20 disqualified attacks per minute. You know that each Windfury Weapon proc disqualifies 1 subsequent attack, yet your example has one Windfury Weapon proc for each 5 disqualified attacks.
Half of the attacks are disqualified because you have 40 attacks per minutes at 1.5s AS and the maximum number of procs is always 20. The average number is assuming the 20% proc rate plays out over time. So only 1 in 5 eligible attacks turn out to be windfury procs making the average being 4 procs per minute.
Let’s look at an extreme (and admittedly absurd) example to show what I mean. What happens if you reduce AS to .1s?
Windfury Weapon
AS: .1s
Attacks per minute: 600
Eligible Attacks per minute: 20
Average WW Attacks per minute: 4
Windfury Totem
AS: .1s
Attacks per minute: 600
Eligible Attacks per minute: 600
Average WT Attacks per minute: 120
In the extreme case it’s clear that WT will produce far greater average damage than WW. My question really is where that breakpoint is based on AS. It exists, but might not be realistically achievable.
No. It’s one combined ICD. Your OH proccing WF disqualifies your MH attacks from proccing it for those 3 seconds.
You’re calculating it in the wrong order. You’re disqualifying attacks and THEN calculating the proc occurrence. But in reality, an attack cannot be disqualified without WW first proccing.
In the 1.5s AS example, each WW proc can only disqualify a maximum of 1 attack. The fact that your number of disqualified attacks is 5 times greater than the number of WW procs should be a giant red flag for you. You need to figure out what’s wrong with your logic
Here’s an example with 1.5s AS. Let’s use the following key:
EA - elligible attack
DQ - disqualified attack
WW - Windfury Weapon proc
If WW were to proc on every attack, the attack sequence would look like this:
WW - DQ - WW - DQ - WW - DQ - WW - DQ - WW - DQ
So you correctly calculate 50% of your attacks would be DQs, which is correct, but it’s only true if the other 50% were WW procs. This is your fallacy.
You then realize that WW only has a 20% procrate, so you take out 80% of the procs. That makes the attack sequence look like this:
WW - DQ - EA - DQ - EA - DQ - EA - DQ - EA - DQ
You haven’t realized that all the DQs that don’t immediately follow the WW are invalid. Only the first attack after a WW should be a DQ. That correction makes the attack sequence look like this:
WW - DQ - EA - EA - EA - EA - EA - EA - EA - EA
Of the 10 attacks, one is a WW, one is a DQ, and the other 8 are EAs. This is what your first example had - 1/10th of your attacks are WW procs. But here, there’s only 1 WW proc for every 8 EAs. So in this example, WW has an 11% procrate, and you’re comparing that 11% procrate to the totem variant. You’re roughly halving the effectiveness of WW and then comparing it to the totem.
For awhile there that was the case, from memory it was fixed around 2.1 or 2.2 patch. If we go in with 2.4.3 itemisation and talents, it will likely also be fixed, hence 2x WF Rank 5 using 2 2.6 (or slower) weapons will be the best.