If you're switching the ret pally out you're keeping the enhancement shaman in, which is about as bad if not worse use of bloodlust/heroism. But I suppose if the leader is really really fast on group swapping he could do quick swaps to make the totem loss neglicible.

That leads me to mentioning that rotating groups also generally means a couple of seconds of not having optimal group buffs, which has a minimum due to latency and human reaction time so has to be taken into account if you're going to go down to the last point of DPS from those bloodlusts/heroisms. And in reality it'll generally be at least a couple seconds in non-optimal groups, at least if it involves multiple swaps per bloodlust/heroism.

You're thinking of the weapon imbue, which has a (could be mistaken) CD of 3sec. The totem has no CD whatsoever.

No. In fact not all too long ago the Future of DPS Warrior or the Compendium demonstrated this happens.

I really wish you could make a nice macro for doing group swaps. Like for Illidan phase 2, our resto shaman actually put themselves into the same group, then split themselves because heroismed chain heal on phase 2 is crazy. Then when it wears off they do it again. They also give our coh priest both heroisms allowing him to spam coh on the grate super fast and a sp gets them both, pumping tons of mana into the group.

Mages

At 20%, a mage roughly doubles his DPS for 15 seconds (Molten Fury, Icy Veins, Combustion, Trinket, Destro Pot, Flame Cap for 20/20/15/10/5/3% multiplicative increases).
Heroism is effectively a full 30% increase on that. Mana cost increase is a valid point, but JoW and SPriest regen (if he gets heroism as well) both benefit from haste, and swapping in one mana pot or mana gem in the middle of the fight is possible without gimping your DPS.

So, during their heroism (40s), mages do an average 165% of their "normal" DPS.
(Averaged over the time with and without IV/Comb/DestPot, assuming 2 trinkets),
Assuming a base 1.8k DPS, Heroism adds 35k damage, at the cost of ~1.8k more mana, which is 3/4 of a mana pot. That requires you to drop 3/4 of a destro pot mid fight (only a trinket multiplier), which will cost you ~2.8k damage.

So, Heroism adds 35k damage and takes 3k of that away if that causes you to run into mana issues.


Warriors

As far as I know/read, warriors use two weapons at 20% to spam Execute every GCD.
Preferably a fast MH/slow OH to reliably get enough rage for Execute every 1.5s second.
Heroism increases the white DPS and the rage/sec, but not your melee GCD, and throws off your rhythm.

So, to fully benefit from Heroism, a warrior would need weapons with a 30% higher swing timer than his normal execute weapons (~2.0s MH weapon, which does not exist, and a really slow OH - OH doesn't matter that much though), to synchronise his swings with his GCD.
Then, he'd gain 30% white DPS and maybe 15-20% rage, which increases his Execute ability damage by ~5% from getting more rage per GCD.

Am I missing something there? Because the benefit to warriors really seems pretty small.


Warlocks

They cycle Life Tap/Shadow Bolt to keep their mana up, and both benefit equally from Heroism.
During Heroism, they can stack their two trinkets and try to time their Life Taps (tap to full before Heroism, end Heroism with nearly no mana, but they have no other class cooldowns.

They could probably be treated as getting 30% full haste while doing 115% of their usual damage (trinkets, and LT timing).



Don't know enough about hunters/ferals sadly, particularly shot clipping and actual cycle times, and haste effect on cats/bears.

Apparently it is possible to have a 26th person in the raid who is somewhere outside the instance, safely out of combat, do automated group switches. The API to do the swaps is locked out during combat, like many other things.

If they want to lock automated swapping they should just lock group swapping altogether of players that are in combat, forcing you to use a bad UI to do something that isn't considered an exploit is stupid - either decide it's an exploit and remove that option altogether or decide it's a part of the game and allow it to be done with any UI.

As for warriors, remember execute gains little from being used with more rage, I think the real gain is to reduce the average amount of time spent between executes which could or could not be heavily affected by bloodlust/heroism. Either way the extra damage caused by having extra rage on every execute is indeed not very big (but is still a part of the increase). And keep in mind the white DPS of an executing warrior may be a certain % of his DPS, but it should (if your warrior is good) be higher DPS than your casters to begin with. Same goes for rogues. 70% of 130% is 91% and adds rage/combat potency on top. You just can't say it affects white DPS only as an argument to not give spare bloodlusts to melee as there's more to it than that.

For mages remember the damage lost from the mana cost due to losing 3/4 a destruction potion is only true if your mana level is where you're swapping destruction potions with mana potions as a way to gain mana. While with a shadow priest that would generally be the case, if the fight is very long and/or you don't have a shadow priest you may be already using mana pots and gems on cooldown and therefore that extra mana will have to come from evocation ticks, which cause a bigger DPS loss for regening the same amount of mana even if you consider the fact you didn't spend mana during those ticks as if it was extra mana regen (meaning you gain mana of 2-8s of evo and 2-8s of not casting but lose the damage of 2-8s of casting, since if you don't add the mana you didn't use as mana gained you won't get a correct result since the damage you lost during those 2-8s wasn't completely lost - you still kept the mana).

2h Fury or Arms specced warriors sometimes will continue their rotation.
Fast/Slow MH/OH setup? This is the first time I've read anything about this. All things being equal, setups are Slow/Slow (high WW dmg), Slow/Fast (steady rage inflow), Fast/Fast (Execute spammage).
'Throws off your rhythm'? We just burn heroic strikes due to the rage inflow. A 2h continues as usually as well, haste is still valuable.
Why must we sync our swings with the GCD? We don't need 'exactly' the rage required for Execute. While anything over the max is not optimized, it is still additional damage being done. During this time, it's about damage per second, not about damage per rage, that is important.

I'd venture to say you're missing a lot. Assuming an Enh Sham (the avg guild seems to use a 3x rogue 1enh sham 1warr melee setup), Lust will influence the amount of WF procs (due to it being % based). Why is this important? WF procs are our biggest external influence on dps.
A recent post on one of our 2 dps threads here suggests that haste does not [negatively] influence rage generation, but that instead, rage gain is based on weapon damage. Thinking about this, a hastened OH will be producing more rage per second than normal, which in turn leads to more successive MH rage burnoff with Heroic Strikes. Furthering this madness is the aforementioned WF procs. What does this mean? Far less glancing, which is a direct 25% boost to said white damage (glances), on top of the additional dmg modifier from Heroic Strike itself.

It's complicated, obviously. I'm not advocating our class makes the best use of Lust (I'd assume melee do, but this thread will hopefully examine who does exactly), but our 'benefit being small' is, simply put, dead wrong.

Edit: Clarity

I didn't say it's better, just that it's not just "he increases 30% of 50% and therefore only gets 1/2 the effect" but he actually gets more than that. Wether it's better or not depends on wether it's really 50% or not as well as how much actual DPS he's doing (which would depend on gear, stage in the fight, cooldowns used, etc).

I'm not advocating giving it to one class over another, just pointing out things that shouldn't be overlooked as they could (or could not, depending on what the actual numbers that hadn't been brought up yet are) be enough to turn things around. In fact I'm pretty sure an executing mage group will turn out to be either the best or 2nd best use of bloodlust once actually calculated, but the data brought up so far is far from enough to be definite about anything. And even if you know mages and melee should get the <20% bloodlust there's still the harder descision of who gets the rest of the bloodlusts and when.

Remember windfury attacks are simply a % increase to white DPS, if WWS counts them as yellow (which according to the post below it seems to count them as white) you should just recalculate and counnt them as white in terms of bloodlust effectiveness, then it won't really be an argument you need to bring up regarding who gets more from bloodlust. Same goes for heroic strikes - while they count as yellow they're actually benefit from haste just like white damage would - in fact they benefit more due to having more rage to use them more often. So another thing you need to avoid being mislead by is the %white dps meters/wws give you, as they don't include all the DPS that gets boosted by bloodlust for melee classes.

While mages have IV as an extra cooldown (aside from item cooldowns which melee classes also have access to), melee has deathwish/blade flurry. While IV is probably a bigger DPS increase, the other classes also do more damage. then again their trinkets may not have a "use" ability due to having better options.

At the end there's a lot more to calculate than just showing how well you scale with bloodlust, as there are too many other factors that are just as significant if not more significant and were ignored.

The only thing that is 100% sure is that you shouldn't bloodlust (aside from non-dps classes...) shamans (other than the one casting it of course), paladins, nor any druids, as I've yet to see anything showing they can keep up with the DPS nor do they have any special cooldowns to boost it, so using it on any other class will probably be better (which makes sense, since those classes also boost DPS of other classes...).

Taking the white damage contribution of a warrior's dps from a WWS report is grossly skewed because Heroic Strike is effectively white damage that can't glance and has a small additional damage component. If you were to eliminate Heroic Strike from the list of dps abilities a warrior has, their contribution from white damage would probably shoot up to something like 60%. Also to give a rough estimate of mage dps with cooldowns at execute range and not do similarly with a warrior is a bit one-sided. The warrior would gain something in the order of 85% dps increase just from using Recklessness. You can then add in Haste Pot, Trinket, Racials (if applicable) and perhaps Deathwish. I personally have seen my threat per second maintained at roughly 2200 for 15 seconds during execute range, that's with Salvation and Improved Berserker Stance. That's roughly 46.2% threat reduction, putting my dps at about 4089.2 for 15 seconds. Heroism/Bloodlust directly affects the number of attacks we can make in those 15 seconds of Recklessness. Not only that but with Recklessness and Haste Pot and Heroism running we can generally get close to 80+ rage executes every gcd for those 15 seconds. 10 Executes at 80 rage, all guaranteed to crit and assuming no dodges/parries, would be 2885 dps on their own. Again, this isn't counting Deathwish which most raiding dps warriors actually will have considering Arms is "the" raiding spec.

I still won't try and definitively state that warriors are the god-given choice for sub20% Heroisms, but we do seem to be somewhat under-represented/estimated at the moment here.

WWS doesn't count WF totem hits as yellow, so it doesn't need to be calculated separately. And yes, I accounted for heroic strike, because actual "white" damage from a warrior is more like 40% of their overall DPS, whereas I was considering their "white" contribution to be 60% of their non-execute-range DPS (rough estimate, it's probably lower after removing the HS damage component and subjecting the swings to potential misses). I assume warriors generally aren't spamming HS in execute range.

[e] Pots (Haste vs. Destruction), racials, and trinkets apply to everyone, not just warriors, which is why I left them out. The question is not how much rage you have on each execute, the question is how much of that rage is a result of Heroism. If Heroism is the difference between 80 rage executes per GCD vs. 70 rage executes, it's not that significant of a contribution to execute spam. If it's an 80 vs. 50 rage difference, however, it is.

Besides considerations on how each class profits from Heroisms... How often do you see a group with 3 DPS warriors? Never, I'd say. How often will you have a group with 3 mages to benefit from Heroism? In many guilds, most of the time...

Since it's been pretty much determined that other classes will get less benefit than either a mage or warrior in Execute range, I'd say <20% Hero's should go to a mage group in priority...