No, dots from LB will not reset the critical strike counter of Hot Streak, just the same as the dot from FFB will not reset it.

Yes, the tooltip reflects the damage in this way, but the tooltips with talents are and continue to be completely inaccurate with respect to +damage gear. Taking Fire Power (+10% damage to fire spells) will result in an increase of 10% to the tooltip value of any Fire spell, but testing easily shows that it is in fact an increase of 10% to the final damage after spellpower gear modification. AB is the same way.

I added the 3.0.8 mode to Rawr.Mage. Barring any mistakes in modeling here are the results I found regarding arcane. I've tested a range of possible cycles, but it's possible that I still missed some interesting variations so this should not be taken as conclusive.

The highest cycles of the ones tested is Arcane Blast spamming with Arcane Missiles on proc, not casting any Arcane Barrage. However its mana consumption is quite high and the dpm conversion is not favorable. Out of all other cycles under normal haste conditions the usual AB-ABar with MBAM replacing AB is still the highest dps cycle.

For extreme haste conditions, specially with stacking of other cooldowns it's worth spamming AB with AM on proc. Otherwise from what I tested AB-AB-ABar with replacing AB with MBAM on proc is the best cycle when normal cycle starts clipping into ABar cooldown.

Mana tradeoffs in general have not changed, it is still better to use molten armor over mage armor unless mana starved. In general mana regen remains of very low value for arcane spec in terms of increasing dps.

But how does it compare to FFB and deep fire spec?

By "frosty goodness" Lhivera is not referring to the essential single target DPS talents, but the flavour talents like Permafrost, Ice Barrier and Improved Blizzard that previously were available due to only requiring 11 points in Arcane/10 points in Fire (depending on your preferred subtree).

Apart from the obvious idiocy of having a talent that boosts damage by at least double the accepted standard of 2% per point, the placement of the talent in tier 4 Arcane gives frost mages who wish to be competitive no choice in where to put their points.

Now a lot of min-maxers are probably just going to say "So what?" but for me personally - and I suspect for many other frost mages - the charm of Frost is in those non-essential flavour talents. I'd rather spec into a FFB build (and actually have more than one button to press!) rather than dump all the flavour for a 12% DPS gain that _still_ isn't enough to bring Frost up to par with the alternatives.

Theres a new glyph of arcane blast. 5% more damage on the arcane blast debuff (?).

Oooo, I need to model this.

By "Frosty goodness" I did not mean DPS talents. I meant all that non-DPS stuff that people use to justify the spec having lower DPS. I figured that would get through in the context, but perhaps not.

Making it 20/40/60%? That would be very nice indeed. Though I suppose it could also be 20/35/50 and still be nice.

Shouldn't 18/0/53 produce significantly better damage than 0/18/53 with the TTW change? +12% overall, plus some extra damage from the Fireballs.

And, I'm sorry, I was kinda talking "should" rather than "is" on the FFB spec vs. Fire spec comparison. Frost, Arcane and Frostfire specs all seem at first glance to have more advantages in more encounter types than Fire, so I'd kind of expect it to place a bit higher with the other three being equal to each other, in a perfect world.

Sorry it's 18/0/53

I have stat comparisons! Ahoy! Comparison is listed as dps gained per point of stat (NOT item point). Taken at Manly's gear level by adding 100 of stat.

Frost spec: spower = 1.44, int = .15, spirit = 0, crit = .49, haste = 1.31

Arcane spec: spower = 1.36, int = .71, spirit = .55, crit = .69, haste = 1.05

Fireball spec: spower = 1.56, int = .2, spirit = 0, crit = 1.23, haste = 1.05

Frostfire spec: spower = 1.58, int = .23, spirit = 0, crit = 1.41, haste = 1.23

There you go.

The changes to Arcane still mean that it can't bring the 10% spell crit buff to the raid, without a significant point spend in an alternate tree. Seems odd from a design perspective that a raid buff that can only be brought by the non-hybrid mage class, is dependant on the spec of that class.

I've scribbled down some maths, and reached the same conclusions as Kel S'jet. In particular, it's possible to derive a couple of inequalities that tell you which rotation is better.

The damage of a cast of AB I call simply AB. The damage of a cast of ABr I call simply ABr.

First of all, the dps of {AB-ABr}. Call this D1.
Second, the dps of {AB-AB-ABr}. Call this D2.

D1 = (AB+1.15ABr)/4.0
D2 = (AB+1.15AB+ 1.30ABr)/6.5

Now consider. If D1-D2 > 0, it's better to use {AB-ABr} all the time. If D1-D2 < 0, it's better to use {AB-AB-ABr} all the time (neglecting mana issues).


After doing the algebra, you get the inequality, that for D1-D2 >0...

ABr > 0.914 AB

This is almost always satisfied, and at higher spell-power it gets easier and easier. It confirms Kel S'jet's findings that the {AB-AB-ABr} rotation is a dps loss of {AB-ABr}, as for the numbers he gives for the average damage of each spell, this inequality is very much satisfied (2910 > 2868).

You can do the same thing to compare D1 with D3, the dps of {AB-AB-AB-ABr}.

The result is:

ABr > 1.04 AB

This is slightly better, but it won't take much spellpower to cause this inequality to hold and for {AB-ABr} to be the optimal DPS rotation once more. If you look at Kel S'jet's numbers, we see 2910 > 3263 and so the inequality is violated...but only barely. This is why the dps gain is so small.

Conclusions

Never cast {AB-AB-ABr} unless you're clipping against the cooldown of ABr due to haste buffs.

Don't even think about {AB-AB-AB-ABr} unless you have mana to spare and your spellpower is low enough that the inquality is violated. The better it is satisfied, the more the gain. Expect gains of about 10dps for every 100 that inequality is violated.

Now that's quite inspiring. Almost brings back the earlier builds on beta with 25% increase/debuff, which rises the question why didn't they try a 5% decrease in the beginning. Either way, as Lhivera stated is a possibility, does it end being 20/40/60 or 20/35/50?

Inequalities that need to be violated for {AB-AB-ABr} to be better than usual {AB-Abr} rotation are for 20/40/60 stacking:

ABr > 1.03 AB

And for 20/35/50 stacking:

ABr > 1.61 AB


Should be very easy to violate the latter, and in fact the dps gain for {ABx3 ABr} would be significant. The first case is unclear. Either way, more hopeful (although it may lead back to a case of AB-spam). More work needs to be done to settle which rotation would be best in these two cases.

With these scaling numbers wouldn't this make Frostfire the absolute best spec if it does competitive (or top) dps right now?

Personally I am mostly astonished by how much int gives to ffb specs .. (?) I guess my calculations were out of date.

Hmm, Rawr says that crit is lower then haste for Fireball spec. (That's with 3.0.8 module on and torment)

I'm assuming these aren't with the patch changes since spirit gains benefit from Fire and FFB specs. If you can, might want to add the values based on that, and up the value of crit due to Living Bomb crits being relevant to Hot Streak, if that is even a worthy increase in value.

These are with the patch changes... spirit only gives you mana. Fire and Frostfire don't get anything out of more mana.

Vontre: out of curiosity (and because I'm working on setting up Pawn), what were the respective values of hit for each spec below the cap?

Vontre, I have some questions regarding how you model the arcane spell selection. I've been trying to make sense out of the cast sequence display, but I'm having some problems seeing the logic behind it. Do you take into account that you don't see the proc immediately after AB and you can react only one spell later? From what I've read before I assumed you did, but I can't explain what I see in spell sequence. For example taking an excerpt from the output:

Arcane Blast - 108.06
Arcane Barrage - 110.17
Arcane Blast - 111.44
Arcane Missiles - 113.55
Arcane Barrage - 115.66
Arcane Blast - 116.93
Arcane Barrage - 119.04
Arcane Blast - 120.3
Arcane Barrage - 122.41
Arcane Blast - 123.68
Arcane Missiles - 125.79
Arcane Barrage - 127.9
Arcane Blast - 129.17
Arcane Barrage - 131.28

If the selection logic is to use AM as soon as possible then I can't explain this. Also you said that on AM you restart the cycle, but here you follow with an ABar.

Another question is regarding the amount of procs and location of procs. As far as I understand your simulation is supposed to generate deterministic output, unlike simulationcraft. If this is the case how do you select where are the procs happening so that you ensure it's statistically equivalent to average?

Also for anyone else that is doing simple analysis by comparing ABxN-ABar, while that analysis can be useful, take into account that any such analysis completely ignores potentially lost MB procs and how the procs are integrated into the selection logic.

Just looking over the new wintergrasp rewards. For pre-raid gearing is it worth using Flow of Knowledge alongside the badge trinket of the same stats? Would they share an internal cooldown? I can't test without enough WG commendations but my gut tells me they'll both work fine (stacking, not sharing ICD).