I think the most important statement is, that for most of the encounter, a early heroism is a win (exclude fights like Maexxna, Malygos etc). Beyond most of the named advantages, there are some more for the 'casual' ones:

• No explicit cooldown math for every fight - just use cooldowns after the first stacking with Lust at everyones choice. It simplified the cooldown management for most of us.
• Lower mortality - a dead DPS Player for a late Lust will screw up all other (possible) advantages

With this in mind, there is one more question: when to use combustion? I guess its best to use in 35%-0%, indifferent when Bloodlust is casted (beside, of course, the encounter allow you to cast it twice or more often).

So, use early BL and enjoy. As an Arcane Mage it was Hell to time Cooldown with late Heroism / different encounter, that was the main reason i swiched to FB.

I don't think I can totally agree here, since the CD starts ticking as soon as you make your 3th crit.
So with that in mind, the faster you get those, hopefully consequetive, crits off, the sooner you're able to use it again.
Which obviously can only have the potential of critting more, "can" and "potential" being the keywords here.
I wouldn't make a big deal out of it, but when discussing min-maxing...

On the other hand, I'm sure I can agree on that though...

3.1 new numbers are out on simcraft, looking dire for mage dps at the moment. weve fallen into hybrid range.

SampleOutputPTR - simulationcraft - Google Code

Historically, how have these numbers compared to actual outcomes?

Hardly a point in despairing before they release the mage changes.

Why is the 57/3/11 mage using Mage Armor instead of Molten? Also, you could try adding the Living Bomb glyph for deep fire and see what that does, although you'll have to make some guesses as to how it works.

Apologies if I missed a post regarding this, but how is LIVE_57_3_11 5892dps, and PTR_57_3_11 5649dps? Are there Arcane nerfs in 3.1 that would cause a 4% dps drop?

Kitty DPS actually went up over 7k after I fixed a gcd-related bug in Cat Form......

Arcane appears much more mana-starved in 3.1...... Unlike other classes who also had their "while-casting" talents increased, Mages cap out at 100% so they cannot fully recover from the 40% drop. Since the SimulationCraft Arcane AI really isn't all that clever in mana-starved scenarios it may be misrepresenting the spec.

The mana consumption charts actually look similar between Live and PTR simulations. But I still would not trust these numbers:

Mage_T8_57_03_11
10955 arcane_missiles
6799 arcane_barrage
4913 arcane_blast

Mage_T7_57_03_11
11211 arcane_missiles
7272 arcane_barrage
5060 arcane_blast

Well, it is more complicated than this. I'll use some arbitrary numbers for simplicity sake.

Pre 3.1: Arcane Meditation was 30% regen for 100 mp5. Mage Armor was 30% regen for 100 mp5. Mage Armor Glyph was 20% regen for 66 mp5 for a total of 266 mp5.
Post 3.1: Arcane Meditation is 50% regen for 100 mp5. Mage Armor is 50% regen for 200 mp5. We are now regen capped at 200 mp5.

Total loss of 66 mp5.

This change only effects those who were using the Mage Armor glyph (which I understand to be the minority since it was understood to be a dps loss overall).

Improved Scorch / Winters Chill drops spell crit by 5% across the board.
Slightly offset by Blood Frenzy / Savage Combat changes.

SimulationCraft uses a full-burn 5min scenario (which should probably be reconsidered) that really stressed the Arcane Mages.

Unfortunately, the Mage module does not have a dedicated owner..... so it just gets a small piece of me. The AI could almost certainly be improved.

No mage specific nerfs announced yet, but the expected loss of 5% crit from winter's scorch (to match 3.1 ISB) is in and of itself a DPS loss for all casters, barring other changes.