Celebrin: the fix to Clearcasting was a simulation fix, not a game fix - the sim picked up the proc rate change automatically from the spell database, but the ICD had to be added manually.

I think part of Celebrin's question remains unanswered. One week ago the thought was arcane needed to manage mana to take advantage of the mana adept mastery, fire was in horrible shape with respect to mana (and either had to drop molten armor or weave in a lot of scorches), and frost didn't really have many mana concerns at all.

Now, suddenly, fire has pulled ahead and can push high dps for 8+ min fights and frost is mana starved after 5 minutes. What changed? Was it just choice of simulator? Were the mana problems with frost evident during beta raiding? If so, I don't remember anyone posting about them. While one can argue fire has a mechanism to deal with mana starvation (free scorches), frost has no such built-in mechanic.

I can add that frost have the lower mana cost and a talent that can reduce spell cost up to 10%. Maybe SimC isn't good with it. In beta, no one have to put a point in Clearcasting if I remenber correctly.

Well, one of the reasons I work with SimC is that I think it has a really strong overall model (the other is that it builds on a Mac). Even if buggy, the foundation is much better than the wee little sim I was working with during beta to model Frost stuff. So in general, I'm going to assume that SimC is producing more accurate results.

Of course that doesn't mean there won't be problems (obviously, as several major bugs have been corrected just in the past few days). So there are two possibilities: we're getting different answers because SimC is producing more correct results, or we're getting different answers because SimC has a bug. And that means examining the detailed output and trying to spot anything that looks wrong, so we can investigate and find out if it is a bug.

Fire, if you look at the detail output, is weaving in lots of Scorches. A whole lot of Scorches, in fact.

Now, of course, there's also the question of whether these fights represent what people were actually doing on beta (I, sadly, could not do any raid testing myself due to a big project at work). I've modeled, for example, 5 seconds of movement every 30 seconds. But what if it's every 20 seconds? That's more time that Frost doesn't spend casting while it soaks up regen ticks; if that's a more common scenario, then perhaps Frost's mana situation is usually better than the sim is showing.

Note also that Frost does indeed have a built-in mechanic to handle mana: Mage Armor. It's not an answer anyone likes, but what I found when playing with the profile last night is that you can either run Molten Armor + glyph and go OOM before 6 minutes are up, or you can run Mage Armor + Frostbolt glyphs, regem/reforge for more crit, spend more points on utility, and drop from about 22700 to 22500 DPS -- but never go OOM.

I encourage people to download the next beta, whenever it's released, and experiment with the Mage profiles. See if you can find ways to improve them.

The proc rate should be correct. The data is not scraped from the tooltips, it's actually retrieved from the database (this is how it can retrieve data not found on the tooltips, such as the coefficients). Looking at the Clearcasting uptime, I'm seeing an interval of just under 19 seconds, which is not unreasonable given a 15 sec ICD and a 40% proc rate.

ETA: New results uploaded, includes improvements to gear for all four specs, and bug fixes pushed for other classes.

I just browsed through the result of your latest Sim, but I can't find combustion anymore in the ability list or in the damage source circle diagram.

Is there any reason that it's listed no longer, because I'm quite sure it was there in the past.

Good catch, Takolin. My pyroblast fix broke Combustion. It's fixed now, and some of the new results will be uploaded shortly (the rest this evening).

Is it still worth to spend 2 points in Early Frost?

On live server the cast time is cutted at 1s.
My normal cast time (without early frost) is 1.43s, but the cast time with early frost is always 1s.

Every other haste effect seems to work properly.

As far as I can see, the glyphs in the OP seem to be off. In post #276 (and following) it's clearly mentioned the primary glyphs should be Molten Armor, Deep Freeze and Frostfire Bolt. Although Molten Armor is under discussion again due to mana issues in simulations, the other two should be changed or at least noted.

Glyph are indeed wrong. Lhivera has glyphs as:

• Frostbolt
• Deep Freeze
• Frostfire

but he is running Mage Armor in his latest sims. Vontre has Glyphs as:

• Molten Armor
• Deep Freeze
• Frostfire

but he is running Molten Armor not Mage Armor.

Is this agreed upon? Does Glyph of Frostfire beat out Glyph of Frostbolt? I know FFB does more damage but those are only occasional casts (and a lot of FFBs damage contribution comes from Ignite, which doesn't effect the glyph contribution). If this all accepted, I'll updated the OP.

Another issue that concerns me is that both arcane and fire seem to scale better with new gear. I'm assuming this looking at the values of both primary and secondary stats in short overviews posted before. Frost seems to have the lowest gain per stat on average.