I'm a little confused as to how haste will screw up an ABar>FB rotation. GCD from Abar + 3s on Fireball totals to 4.5 seconds before any haste effects, for a full 1.5 seconds over the cooldown of ABar. Since the goal of the rotation is to contain it within 3 seconds to always cast Abar on cooldown (correct me if this is the wrong idea) then it seems like a fire subspec can handle substantially more haste before the cycle is disrupted than frost subspec can. Someone discussed this in a post upthread concerning the Frostbolt rotation, since under BL/IV you drop under the 3s timer and could more profitably use Frostfire bolt. One of the main advantages of fire over frost, from what I've read, is that it doesn't run into those issues.

That said, I think you're correct when you say frost will have preferable mana control. The point spread required for fire is going to take some tweaking, I think.

Doesn't this mean though, seeing as how Fireball is a 3 second cast +GCD = 4.5 seconds, that for this spec, you'd need a hell of a lot of haste just to break even and be able to spam Arcane Barrage every three seconds? What about a 51/10/10 spec for Frostfire Bolt? (51/9/11?) Moving into the realm of conjecture, but seeing as Roy says that FFB scales better than all the other spells, would it be the spam spell of choice in the rotation? Or just during Icy Veins?

Keep in mind that as Arcane, Barrage is going to be the highest DPS spell you have. Using Fireball or Frostfire Bolt may increase your 'filler DPS', but this is very likely to be offset by a decrease in the amount of time you spend casting Arcane Barrage. Using a 3-second filler seems like a bad idea to me.

ABar isnt the highest DPS, FFB is.

I did some napkin math about Abr with fire/frost rotations, and was bit suprised to find out that at 1500 dmg and 25% crit, fire lost only some ~10dps when compared to frost rotation. During bloodlust it even surpassed frost by ~100dps, although frost was 250 more dps when it used IV alongside that BL (and I didnt even account the synergy between IV+AP in this math).

Considering that fireball scales better than frostbolt, I'd assume that it will eventually surpass frost although I'm yet to do math for 1800/2000 sdmg setting. Besides, frost is "hastecapped" at around 15% haste, which is really easy to get, while fire can take around 25% haste before losing benefit.

In the end, it's a matter of preference which build you want to take, since the damage between the specs being only few % apart. Roywyn had also similiar results in his math somewhere earlier.

Yes, this was precisely my point. Fireball becomes a preferable filler to Frostbolt with more haste, not 0, because it avoids problems with dropping below the 3 second timer during cooldowns. My initial thought as far as Frostfire goes is that a 51/10/10 spec might not make it as powerful as Fireball with a full fire subspec even with the crazy crits, but I don't have the math to back that up. Roywyn might have a better answer (excellent work, by the way, it was good to have solid numbers).

One other point in favor of using fire as the filler is that you can take Imp. Scorch, which depending on how the new Shaman lava burst thingies end up, and how good fire destro is, could potentially be an important debuff for more than just yourself. I'm thinking 10-mans with a shammy or lock friend here.

Roywyn, did your calculations for Abar/Fball filler cycle include keeping the scorch debuff up? Is it even worth putting up when less than half your damage is fire?

Edit: I'm going to try to do some actual calculations on haste for fireball vs. frostbolt tomorrow. This has me all curious now. I have a friend in beta who will hopefully be able to test whether the new wrath of air and the new ret aura will stack with each other, but if anyone knows the answer already it'd be helpful. I'm assuming nothing will stack with BL but it seems impractical to have the ret aura affect melee and spell speed if Wrath of Air and the new WF just overwrite it.

That's true, but the differing haste buffs that exist will cause problems. Given certain group compositions differing amounts of haste will be preferred. For instance, if you have WoA and NWP, you don't want more than 14.4% haste from gear if you plan to use Frostbolt as a filler. However, if you also have a moonkin in the group, you won't want more than 4.8% haste, so that your cycle doesn't drop below 3s when you crit and get the extra haste. That is assuming that the haste buffs don't stack. If they do (I've heard rumors that do, but I'd like to think it isn't true), then you'll have to use FB or FFB as a filler when you crit and get the extra haste, and you wouldn't want more than 4.1% haste from gear.

Alternatively, if you have a a ret paladin and NWP, but no WoA, you'll want up to 22.1% haste from gear, or 8.9% haste and switching to FFB or FB if you get a Heroism, assuming Ret Aura and Heroism don't stack. You'd want up to 5.7% haste if they do stack.

So, depending on your haste from gear, using a 3s filler spell is quite possible. And if all the haste buffs stack, I am curious if ABar will be worth casting, since your GCD would be hard capped at 1 second, while your frostbolts would be casting in 1.11 seconds.

Unless there's something I missed, they specifically stated that they don't want haste buffs obtained from group or party buffs to stack... so boomkin aura haste wouldn't stack with WoA.

Also, in all the WoLK gear I've seen I have yet to see haste rating, so I wonder if they aren't going to not make haste available for the first few raids... similar to what they did in TBC.

Wowhead lists plenty of items with haste on them, both for melee and casters: Activist's Signet of Blasting, Cloak of Azure Lights , Choker of Betrayal, Horn of the Herald are just a few that pop up if you search, and they're available at low levels. That's not even getting into all the Haste gems available. It looks to be a very readily available stat in WotLK.

Arazan, I'm in full agreement with them on not allowing all the different haste buffs to stack. I think the game would be very broken if they all stacked. As it is, the ret aura one gives the ret pally 3% haste on top of anything else he gets, and is an okay substitute for WoA if you only have 1 shaman, such as in a 10 man. Allowing WoA and WF to stack for the shaman who casts Heroism/Bloodlust is fine as well, since that gives shaman extra DPS cooldowns, while also providing a wonder group cooldown. And Moonkin can probably use some haste love of their own. But the idea of a mage getting every haste buff and getting down to a 1.11 second Frostbolt before he adds in his own haste rating from gear is downright awful. There'd be no real way to balance it.

Still, I was informed that tests in the beta were performed and they were stacking, so I had to take it into account, even if I personally feel that their stacking is complete bullshit. Plus, if they don't stack, the WF and WoA changes are actually shaman nerfs, and I have no problem with shaman getting a nerf. Currently, a shaman provides 30% haste from Heroism/Bloodlust, and chaining that is a plausible reason to bring more shaman. But if it isn't stacking, then it becomes a 10% melee haste buff or a 20% caster haste buff, assuming moonkin aren't in the group, which is a good step down from the 30% it currently is.

What i mean in terms of ArB not scaling with haste is that with enough base haste gear or Icy Veins or even the spell haste talent, you begin to not be able to pair instants ((0.0 secs)Arcane Barrage + (1.5)Fire Blast + (3.0)Barrage becomes (0.0 secs) Arcane Barrage + 1.2 fire blast + 2.4-3.0 sitting around + 3.0 Arcane Barrage) or substitute that fire blast with a polymorph, or an ice lance, and suddenly you realize that that arcane barrage cooldown really needs to scale with haste or it begins to really hurt