Shaman: Windfury
 

This article has not been updated for 3.0 but most of the group shenanigans are gone as both Windfury Totem and Wrath of Air totem are raid wide, and Grace of Air has been rolled in to Strength of Earth. The remainder of the article that deals with Windfury Weapon remains relevant and unchanged.


This article describes the effects of the Shaman buff "Windfury". (Tables in this article are messed up due to a plugin)

 

Contents

1Windfury Totem vs. Windfury Weapon 2Windfury Weapon 3Windfury Totem 3.1Windfury vs. Grace of Air 3.1.1Damage 3.1.2Tanking 3.2Twisting with Grace of Air or Tranquil Air

[top]Windfury Totem vs. Windfury Weapon

Shamans have two different versions of Windfury. One is a totem that provides a main hand weapon buff to the entire group, the other is a self-only buff that can be applied to the Shaman's own weapons only. There are some critical differences between these two buffs:

	Windfury Totem	Windfury Weapon
Damage Type	White Damage	Yellow Damage
Number of Attacks per Proc	1	2
3 Second Hidden Cooldown	No	Yes
Can proc from Yellow damage	No	Yes

[top]Windfury Weapon

Please note that the mechanics for Windfury Weapon do not apply to the buff applied by Windfury Totem. This section of the article relates to the Shaman's self-buff only. If you are confused by this, please unsubscribe because I cannot make it any more clear.

Though no one is really sure how long it has existed, the 3.0 second cooldown that is associated with Windfury Weapon was brought to light around the 2.0 release when Dual Wield became available as a talent and it became common to wield weapons with a speed faster than 3.0. Proving this is very easy -- you will never, ever receive back to back Windfury Weapon procs while autoattacking with a single weapon whose speed is less than 3.0. (Slow effects such as Ice Armor that artificially inflate your weapon speed can cause this to happen.) Blizzard has never stated a specific reason for implementing this cooldown, though it is probable that some combination of latency and freak WF proc chains would be sufficient to kill other players far too quickly. It also acts as something of a limiter on PVE DPS, as an unlimited proc rate for Windfury would almost certainly produce unreasonably high amounts of damage.

The implication of this cooldown is very straightforward: it is in the best interests of the Shaman to wield weapons whose speed is as slow as possible while still dual wielding. The reasoning for this is also straightforward - the white damage done by a weapon of any speed that has the same DPS as any other weapon of a different speed with the same DPS, so you cannot modify your white damage contribution from your weapon. You must maximize your extra attacks - in this case, maximizing the damage that each Windfury proc and Stormstrike attack will cause. Since neither WF procs nor Stormstrike are normalized (using a flat AP value regardless of weapon speed for a given class of weapon) they both hit much harder when using slow weapons. Under a perfect set of circumstance the maximum proc rate for Windfury Weapon is 20ppm, though it is typically around 15 ppm when wielding slow weapons -- this is far from a trivial contribution to your damage.

The concept here is the same as the reason that Warriors and Rogues use slow weapons in their main hands-- when you are making a limited number of attacks per minute (Mortal Strike, Sinister Strike, Backstab) it makes sense to have those attacks hit as hard as possible. Windfury works in much the same way that these attacks do, only without the player interaction.

A quick example: a 1.5 speed weapon that does 100 dps versus a 2.6 weapon that does 100 DPS wielded by a shaman with 1400 AP. We'll ignore target AC, avoidance, wielder crit rate and flurry for the purposes of this argument, as they are scalars that would affect both equally. Similarly, the difference between main hand and off hand strikes is 50% so this also affects both weapon speeds.

White Damage (autoattack) Average DPS after AP contribution
1.5 speed: 100 + (1400/14) = 200 DPS
2.6 speed: 100 + (1400/14) = 200 DPS

We see by this example that DPS does exactly what it advertises it does -- both weapons do an equal amount of damage over a given time span. The 1.5 does not hit as hard as the 2.6, but hits more often. However this does not hold true for the WF procs:

Windfury Proc Damage
Speed * (weapon DPS + DPS from AP + DPS from Windfury AP bonus) = Damage
1.5 speed weapon: 1.5 * (100 + 100 + 31) = 347 damage
2.6 speed weapon: 2.6 * (100 + 100 + 31) = 600 damage

As you can see from this example, the slower weapon's Windfury Proc is nearly double the damage of its faster companion even though their white damage contributions are equal. Unlike your normal attacks, the 1.5 speed weapon does not proc Windfury Weapon significantly more often than its 2.6 counterpart -- at most you will see 2 to 3 additional procs per minute, and this is not nearly enough to make up for the tremendous damage difference. Even at a nigh-unattainable 20 PPM versus a 2.6 speed weapon with a conservative 15 PPM, a 1.5 speed weapon cannot match up.

Damage from Windfury procs in 1 minute:
Quantity 20 WF procs from a 1.5 speed weapon: 13880 damage (231 DPS)
Quantity 15 WF procs from a 2.6 speed weapon: 18000 damage (300 DPS)

Adding 3/3 Elemental Weapons improves the overall damage done by your Windfury Weapon procs by 40%, resulting in:
20 WF procs from 1.5 w/ 3/3 EW: 19432 (323 dps)
15 WF procs from 2.6 w/ 3/3 EW: 25200 (420 dps)


Total DPS from auto-attacking:
1.5 speed weapon: 200 + 323 = 523 DPS
2.6 speed weapon: 200 + 420 = 620 DPS

Difference: nearly 15%. Simply by wielding a slower weapon, a shaman can noticeably improve their melee damage. The benefits of slow weaponry are not limited to this example. This particular example uses an extremely conservative AP value -- the gap widens the higher the Shaman's AP climbs. The example further ignores additional benefits of slow weapons including better burst damage in PVP, improved Flurry uptime, and harder Stormstrike hits. The uptime of Unleashed Rage is generous enough that it is very easy to keep up even with dual slow weapons. There is simply no good reason to use a faster weapon in conjunction with Windfury. For these reasons it is best to ensure that your procs hit as hard as possible by using slow weapons (this is also one of the reasons that stacking Strength/AP and Crit is more beneficial than stacking hit).

When using haste effects such as Bloodlust or Haste Potions or Unholy Frenzy (or all of the above), the amount of autoattack DPS gained offsets any penalties for a faster weapon speed -- Enhancement Shamans scale very well with haste and the slight WF proc dropoff should not scare you away from looting useful items with passive haste on them. Flurry is absolutely still worth taking, as the 30% haste vastly outweighs any potential lost procs.

For a period of time after 2.0's release, it was possible to mix ranks of Windfury Weapon and cause the main hand to have a separate timer from the offhand, but this loophole was closed in a subsequent patch. All ranks of Windfury weapon share a hidden internal proc cooldown.

[top]Windfury Totem

Windfury Totem is an air totem spell that Shamans receive at level 32. Every five seconds, this totem will pulse a 10 second duration main hand weapon buff to all members of the group within range. This weapon buff will be not applied if any other weapon buff is present. Using things like Sharpening Stones, Mana Oil, or Poisons will prevent you from gaining the use of Windfury.

[top]Windfury vs. Grace of Air

Which air totem a Shaman should use is a constant source of argument. Typically, the most heated competition will be between Grace of Air and Windfury totem. The totem that will provide your group with the most benefit is the one that you should use, and the makeup of your group dictates what the "right" answer is.

[top]Damage

Grace of Air and Windfury generally affect your groupmates' damage in the following fashion:

	Windfury Totem	Grace of Air
Warrior (Arms)	Very High	Minor (w/Sharpening Stone)
Warrior (Fury)	High	Minor (w/Sharpening Stones)
Rogue	High	Medium (w/Poisons)
Hunter	None	Medium
Feral Druid (Cat)	None	High
Feral Druid (Bear)	None	High
Retribution Paladin	Very High	Minor
Protection Paladin	Medium	Minor

Though the post-2.1 Windfury nerf has greatly reduced the power of the totem, it is still the single best buff that a Warrior or Rogue can have. Windfury damage is now very easy to estimate:
Where WF Atk Bonus Damage is 41.2 for 2/2 Enhanced Weapon Totems, 36.55 for 1/2, and 31.8 for 0/2.

This value will be slightly higher for Fury Warriors since Windfury crits can generate Flurry charges, but it is still a very good estimate. This does not factor in that Windfury hits will generate additional Rage for Warriors, which can in turn be used to do even more damage.

At very low levels of itemization (level 70 blues and greens), it is possible for Grace of Air to provide a Warrior or Rogue with greater benefit than what Windfury provides. However from Karazahn-level itemization and onward, Windfury will conclusively beat any other possible weapon buff in combination with Grace of Air for either class. For this reason, it is advantageous to place as many Warriors and Rogues in groups with Shamans as possible.

As a general rule of thumb, if there are three or four Warriors or Rogues in any combination in the group with the Shaman, it is better to use Windfury. Other group combinations require a much more in-depth analysis to determine the "best" answer.

It is important to note that in no situation will Flametongue Totem ever outperform Poisons or Sharpening Stones/Weightstones, or even Wizard Oil for Paladins that are tanking. Wrath of Air is superior for generating Paladin threat than either Graceof Air or Windfury.

[top]Tanking

When a Shaman is assigned to the tank's group, the question of which totem to use takes on a different context. Instead of determining which provides better DPS, the Shaman should determine which is needed more: threat generation or tank avoidance. For a Warrior or Paladin, Windfury can provide a great deal of additional threat generation. For Warriors, Paladins and especially Druids, Grace of Air provides a great deal of dodge. The right answer in this situation is often dependent on whether or not your DPS classes are threat-capped. If they are, the Paladin or Warrior should probably be using Windfury for additional threat generation to allow the boss to be killed faster. Though this will cause the tank to take more damage per second, the total amount of time spent in the fight will be lessened due to higher DPS and the total amount of healing will almost always be reduced.

[top]Twisting with Grace of Air or Tranquil Air

Since both Grace of Air and Windfury Totem are both short-duration buffs, it is possible to 'twist' both totems to be active at the same time. This is mana intensive and requires a lot of Global Cooldown time, and may not be appropriate for all fights. This has been recognized by Blizzard as an unintentional mechanic and will likely be changed in the future.
WoW Forums -> Totem Twisting

A macro for twisting two totems might look like:
Disquette has written an addon that will keep track of the refreshing windfury buff that will assist in twisting.

Pros:

• Party gains 88 Agility or extra threat reduction
• Overall DPS of the party is increased

Cons:

• The shaman is placed on constant global cooldown, impairing his ability to perform other activities such as refreshing other totems, casting Purge, or healing
• The mana cost of refreshing 2 totems every 9 seconds is enormous. Raids that do not have paladins using Judgment of Wisdom will be unable to sustain this method. The shaman will have very little mana available for shocks and will likely only be able to sporadically use flame shock.
• Until you are very practiced with the technique you will find your attention devoted to watching the timers which means you will be less attentive to changes in the game environment
• Fights that have any sort of significant movement will make this very difficult.
• Latency or simply being late to refresh the Windfury totem will cause the Windfury buff to run off your party's weapons, which means they are losing opportunities for a Windfury proc. The potential DPS gain from a GoA totem might need to be offset by the potential for a loss in Windfury procs.

It is up to the player to decide whether or not twisting is appropriate or feasible. There is no hard or fast 'right' answer.

WFT's AP boost is almost certainly implemented as a modification to the player's AP for enough time for a swing. I've never seen any specific testing on it, but on a laggy server you can easily get multiple instants per fight that are about 445 AP (or 578.5) above where they should be that happen to be very close to WF procs. This is true for that WWS, too:
Wow Web Stats

Windfury Attack - Spells - World of Warcraft

The "Duration 1.5 seconds" field seems relevant here. Clearly it isn't always lasting 1.5 seconds, but there's probably a reason why it's so long. In this log, there's a lot of cases where there's gaps of up to .5 seconds between the WF proc appearing and the hit appearing, and this is on a stable server (I assume). Given that, 1.5 seconds seems like a reasonable length to ensure that unless the server is dying the WF proc will consume the buff, but without letting the buff sit around forever. However, the buff sitting around forever is only a problem if it actually does something to non-WF swings.


None of this applies to the self-enchant, which is implemented as a server-side dummy effect. It's somewhat strange how different they are -- surely it would have been easier to just make the WFW effect generate rage as an additional effect rather than have two completly different implementations of the same spell.

For Protection Paladins Wrath of Air is actually superior single target threat up to a very high level of spell damage. An important thing to note is that it also opens up the ability to use [Superior Wizard Oil] which effectively means you get 143 more spell damage with Wrath of Air. As each 2 points of spell damage is worth 1 threat per second, that means the combination of Wrath of Air with Wizard Oil gives you 71.5 extra threat per second.

Windfury Totem on the other hand provides a 20% increase to the threat of Seal and Auto-Attack, as well as an additional flat extra 6.3 threat per second from the attack power (Before armor reduction). This means your Seal and Auto-Attack threat needs to be at least 326 per second, a value which requires 1066 spell damage, which even in completely optimized Paladin tanking gear is currently unrealistic.

The base TPS value of Seal of Righteousness is 51.4, while the base auto-attack TPS will be 76.0 (Before armor, this is for a Blood Elf), meaning we need to get an additional 198.6 TPS from spell damage for Seal of Righteousness, which requires 1066.6 spell damage.

All numbers above were without taking into account armor, misses, parrying and dodging. All of which would serve to further increase the amount of spell damage needed to make Windfury better, on the other side I also did not take into account spell resists.

As always whether or not to drop Windfury totem is very much dependent on the setup of the rest of the group however. While Wrath of Air along with Wizard Oil is certainly better than Windfury for a Protection Paladin, the actual difference in between the two in practice is small enough that if the rest of the group would fare better with Windfury, Windfury should still be dropped.

Should mention that for tanking warriors that are not expertise capped, which is still most of them, Windfury Totem can mean a random increase in damage taken due to the additional parry-able attack.

Thus, it's probably a good idea to check with your warrior before each fight.

Guess someone should flag this for revision, as section 3 is completely outdated and is now moot.

Where x is weapon speed in swing per second, t is the windfury cooldown in seconds, and P(A) is the probability that a hit will trigger windfury with windfury's cooldown taken into account.

Is this equation correct? (I didn't actually symbolically solve this problem, since that kind of math goes way over my head. I simply looked at the results of my simulator and came up with an equation that seemed to fit the output)

Edit: I've graphed it and compared the output to the expected output for a couple of values, and it does seem to be correct. Although if you put in 3 seconds as the cooldown a 3.0 speed weapon will hit inside the cooldown, so if you want it to act differently you could just put in 2.999 for the cooldown.

Also, the probability is only correct for auto attacks, which makes it less useful.