Making blanket statements that like this is always dangerous. Just like you shouldnt assume +dmg or +dmg/+haste gems are always best.

Actually, the value of Int VERY much depends on how much haste, spellpower and crit you have AND the nature of the fight. In Sunwell gear in TBC, I end up with some yellow INT gems being clearly best in slot based on Rawr sims. The thing that is not obvious is that more INT equals longer AB spam times due to both mana pool size and mana regen boost, not to mention int-> crit and int-> spellpower.

Your mileage may vary greatly.

Yes, but it requires you to be within 26 yards and increases odds of ignite bug. I have played with scorch/fireblast/pyroblast rotations for PVP/mobile fighting and it's interesting but fireball or frostfire spam beats it in raiding situations, and scales better with gear. The scorch/fireblast/pyroblast rotation is similar in DPM to a straight scorch/fireball rotation, so the mana is there. It's just more dangerous to the caster and nets to less damage.

A normal raider will have a significant percentage of all spells cast in his rotation as pyroblast (15%-25%) whereas scorch will at most be ~11% (unless you're overscorching). 6% crit rate on a spell that happens 1.5-2 times as often and does more than twice the damage is pretty self-evidently better than 6% crit rate on the lesser spell, even with the increased odds of adding a hot streak proc.

Using Fire Blast at a Hot Streak proc is pointless unless the Fire Blast crits. Might as well just had another fireball queued up rather then wasting GCD on it. Correct if I'm wrong on this one.

edit - never mind. I don't understand what he's doing with "potent hot streak".

To Clarify Arcane Potency:

The buff you gain isn't consumed until a spell HITS the target. If you gain the buff then launch a fireball then hit fireblast, both spells will gain the benefit of the +30% crit chance buff. This can easily be verified on target dummies.

Hence with a naturally high crit chance already in a raid environment, getting a proc of Arcane potency can give a VERY high chance of causing it to also proc hot streak.

I just realised I got 2 phrases crossed on my original post on the subject. Sorry for the confusion, fixing it now.

Ok that makes sense. It's similar to how combustion or shatter can be used to sometimes proc hot streak. I don't think you need to go to target dummies though, the math should be something like this.

You would work out how many pyroblasts you expect with a given crit rate over a given amount of time, ie a normal scorch/fireball/pyroblast rotation

You would work out how many more pyroblasts you'd expect with this combo+add in fireblast damage and work out how much time that adds to your model (1 GCD per extra pyroblast and for each fire blast cast)

You would normalize over extra gcd's added by fire blast and compare the two.

(you would also hope that fire blast didn't eat any ignites, or factor that in somehow).

I think there might be some edge conditions here, because a pyroblast could be your "clearcast" spell in the rotation..although the fireblast could still benefit from the crit and improve odds of a following non+30% fireball of critting. Also you have to factor in the idea that the fireball would follow a prior crit and you'd have hot streak anyway without the fire blast. The math is hard enough to do right that it probably is easier to get a statistically relevant sample at the target dummies for most of us.

I did something similar to model the effects of combustion. (combustion turned out to have a very small improvement on DPS vs not having it. It's not a very good DPS talent, but it's helpful for "burst on demand". If it wasn't for dragon breath I probably would not take it)

This looks great, thanks !

Has anyone done the math on 4pT7 set bonus? "5% increased critical strike damage" Does this mean, it is roughly a 2.5% damage increase assuming a 50% crit rate? If we make a simple conversion, how much +spell damage is it worth to sacrifice? Assuming my above assumption is true, 2.5% increase in damage is roughly ~ 60 hit rating. What is the latest say on this?

Is ClassTimer updated for 3.0, then? I'd been using Ellipsis on the beta, but it doesn't have an option to break out bars for your focus target separately (that I saw) which I really like for sheep tracking.

The writer of Ellipsis informed me that you can't separate out your focused targets, but you can sort them to the top of the primary list, or show only the focus.

The one I linked works on beta, so I would assume it would work in 3.0. Can't remember the last time I sheeped something in a raid/instance on beta though. With all the new aoe talents most every pack gets aoe'd down.

Not sure if anyone had reconfirmed it, but any ice lance past rank 1 = not working as intended still.

It took them 2 years just to confirm that it actually bugs. A fix before WotLK is very unlikely. And the blue post didn't actually say that they are working on a fix more like it is on the todo-list. I was thinking Blink when I read it. With Hot Streak ignite is hurting my intended firedps more than usual actually.

Listing the data I have received for the 4/5 Frostfire Set Bonus

Tested totally naked except for the T7 4pc, no helm (no meta), had 464 spellpower.
Untalented: Hits for 494/495, crit for 754
With Spell Power (no DPS increasing talents): Hits for 494/495, crits for 877/878

Added a helm/cloak to activate CSD, went up to 658 spellpower:
Untalented: Hits for 549/550, crits for 864/865
With Spell Power: Hits for 549/550, crits for 1014/1015

Naked except 4pc, 464 spellpower:
Untalented: 492 tick, 751 crit
With Ice Shards: 492, 997 crit

With CSD, 658 spellpower:
Untalented: 520 tick, 818 crit
With Ice Shards: 520 tick, 1101 crit

Full gear with CSD, 1959 spellpower:
Untalented: 706 tick, 1110 crit
With Ice Shards: 706 tick, 1495 crit

1600 spell damage, Arcane Missiles (rank 13), no CSD, no talents.
Hit: 819 arcane damage Arcane Missiles
Crit: 1249 arcane damage Arcane Missiles


Conclusion
It looks like it's a 5% increase like Ice Shards, meaning a 2.5% on the total crit modifier only, 3.5% with Ignite.

The final formula for crit damage then seems to be:
(1+(0.5+0.045[CSD])*(1+0.5[Spell Power]+0.5[Burnout]+1[Ice Shards]+0.05[4T7]))*(1+0.4[Ignite])

**update** As per Laekoth, added 10/51/0 ... a new winner!
**update** Changed 0/50/11 to moten Armor ... pulls ahead

Here are my new numbers for level 70 post 3.0.2 dps. This time I am leaving out TtW because of assumptions that it will not work on raid bosses:

Top Scorers
Mage_105100
Mage_005011
Mage_005308
Mage_115000
Mage_134800

All mages are using Molten armor. (Mage_005308 still has 1% downtime with mage, but loses DPS)
All Mages have Glyph of Mana Gem
Mage with Arcane Power has Glyph of Arcane Power
Mages with Imp Scorch have Glyph of Imp Scorch (Even though only one is scorching)

The Mages are using these Glyphs because they are the only once availible at 70, and that the sim uses.

Focus Magic "circle jerk" is as follows:
Mage_530800 -> Mage_144700 -> Mage_115000 -> Mage_530800

Stats for all of the mages are taken from my T5/T6 mage as if I re-gemmed for hit cap:
gear_stamina=409
gear_intellect=419
gear_spirit=204
gear_spell_power=1116
gear_spell_crit_rating=282
gear_spell_hit_rating=176
gear_haste_rating=69

Thanks to the creator of Simulationcraft for making this possible.

I have some annoying nitpicking results to share.

The way this crit formula works results in CSD providing more than 3% boost to crits.
(1+(0.5+0.045[CSD])*(1+0.5[Spell Power]+0.5[Burnout]+1[Ice Shards]+0.05[4T7]))*(1+0.4[Ignite])

Consider full frostfire bolt spam.
No CSD:
(1+0,5*2.55)*1.4 = 2.275 * 1.4
CSD
(1+0,545*2.55)*1.4 = 2.38975 * 1.4

Thus CSD is larger than no CSD by a factor of nearly 5%.

Which begs this question, are you sure the formula is correct?

The "greater than 3%" affect has been observed in-game on any spell with greater than 50% crit modifier.

EDIT: SimulationCraft treats CSD (and the old Elemental Oath) a little bit differently, but virtually the same:

(1+(0.5)*(1+3*[CSD])*(1+3*[EO])*(1+0.5[Spell Power]+0.5[Burnout]+1[Ice Shards]+0.05[4T7]))*(1+0.4[Ignite])

Essentially.... Talents are additive. All others are multiplicative -and- x3.

I linked that post for level 70 comparisons in the original post.

Could you maybe add some more information (like gear levels, molten/mage armours, glyphs) to the numbers?
Do you have an older post with that information that you would like to maintain?

Also, how did you setup Focus Magic in there? I'm still curious whether 11/50/0 would do more total damage than 0/53/8.

Since the differences are pretty small there, it depends on how the Focus Magic setup is and whom it got attributed to now.

Chaotic Skyfire Diamond
It behaved like that for roughly a year now. If you suspect it has changed, just check it with Blizzard or Arcane Missiles who do large fixed damage.

Thanks for your numbers Raven. And for reference, can you add specs and castsequence to that post please?
So we get rid of all the extra posts, regarding that issue.

SimulationCraft will always "correctly" defer Ignites safely without munching them because it does not model travel time.

I've seen tons and tons and tons and tons of anecdotal information...... but nothing hard enough for me to model.

It seems -something- has changed in Beta..... but I just don't have enough information.

This isn't directly true. Ice lance ranks 2 and 3 do not apply triple damage to frozen targets if you have Molten Fury. Without Molten Fury, these ranks work fine. Molten Fury also breaks Torment the Weak's effect in conjunction with Frostfire Bolt's chill effect, for some reason. It's a very odd and specific bug and I'm not sure how in the world that slipped in.

removed

Its pretty easy to speculate that blizzard didn't want both talents stacking, much like the old combustion/pom not stacking.