I'm fairly fresh to theorycrafting mage talents and DPS numbers, my mage has generally been about my 4th in line, so up until the patch, I didn't put much time into reading the 3.0.2 talent adjustments.
I'm playing and interested in Fire right now, his gear level is a muddle of Kara and Lesser PvP piece's as well as a few Badge items.
The World of Warcraft Armory
There have been a few posts within this thread already about situations where scorch spam, living bomb and hot streak pyro's may have a chance of being at least as effective as using fireball as the primary nuke.
Right now I'm using the standard rotation with fireball as the mainstay, but I only raid 10's and I don't usually get the same complete set of mana regen options as I might in 25's. Our DPS on these runs isn't top notch either, so they are not all 2 minute Kara boss's. I am forced to use Improved Mage armor over Molten armor as well as using Improved Mana gem.
What I'm thinking is that the mana efficiency gain of going pure scorch should let me swap to Molten armor. This is effectively adding 11% crit to my Hot Streak trigger (5% Molten, 6% scorch talent) With the way Hot streak proc chance scales up faster as crit rises I'm wondering if there is any value into me giving the spec a serious consideration.
I was thinking I could push it further towards its strength by re gemming to crit. (Do spells have a single table based miss/hit/crit roll like melee white hits, or is it like the rogue specials where your hits are rolled before your crit). Due to chaining scorch, the uncertainty and risk of letting the scorch stack roll off with large miss chances isn't a factor either.
If the fight is faster, or I realise partway in that I'll have ample mana, I figure that I can change to a scorch, scorch, fire blast (6 sec talented, also improved crit from talent) rotation to bump the DPS some more.
--
On the same topic, would someone mind clarifying the formula for working out the chance of a Hot Streak? I'm working on the assumption of simply crit*crit (so 50% -> .5 x .5 = 0.25 -> 25%) which would be 12.5% chance after each given spell cast as an average, but then something in the back of my mind has me thinking it's more involved that than, with the way crits can pair up with their neighbour anywhere in the log (ie. 1 and 2 might not might not pair up, then 3 and 4 might not pair, but 2 and 3 might have both been crits)
I'm sure there is a good mathematical expression for the answer, would love to know it :-). My gut tells me that I've oversimplified something.