Ah, now, I'd envisaged quite a different set-up for arcane from the one you mentioned. I was also extrapolating from those wonderfully vague sentences, but I'm operating here under two fundamental assumptions:

1) Mana spell costs are going to remain high enough to mean that you can't waste any mana regen from evo.

2) The damage bonus will not be an analogue value that changes gradually over time, but rather a set of brackets (e.g. 0-20% mana = X amount of damage, 20-40% mana = X+1 amount of damage, etc). I think this is a reasonable assumption to make, as otherwise it would force a player to do a vast amount of theorycrafting to maximise their dps, and Blizzard have stated that this is something they would like to see less of (hence the stat reforms to make the gameplay more intuitive).

With this in mind, I see Arcane's rotation unfolding as follows:

You start at 100% mana, and you are in the top damage bracket. You want to stay there for as long as possible. Therefore you use a less mana consuming rotation (e.g. drop the stack after 2 with a barrage, use Missiles whenever it's up etc) and gem as soon as you won't waste any. Your rotation means that you're doing less than maximum dps, but the mastery boosts that a sufficient amount to cause you to be actually doing more damage overall because you're staying in that top damage zone for longer.

As soon as you leave that top damage zone, or leave the zone where the mastery bonus will no longer patch up your rotation, your aim is to burn as fast as possible down to 40% and evocate back up. This will most likely involve straight Blast spamming, but who knows. You'll be losing mastery spell power, but the higher dps rotation makes up for this. The result is a fairly steady dps.

The thing is that it's more or less impossible to do any real theorycrafting without numbers. We can guess, but at the end of the day anything could happen, including the entire scheme being scrapped.

More thoughts on the Cat changes:
http://doyoumindifiplay.blogspot.com/2010/04/lets-do-time-warp-again.html

Reply