You shouldn't be making the trinket decisions in a vacuum. Gale has an insane amount of haste on it, 157 more haste than a reforged Volcano, so there is a spot based on your other gear where it can get you the 7th immolate tick that volcano can't. Which really is in the full 359 to partial 372 phase of raiding ie. the most critical time to be min/maxing.

This is also true of heart in the lower 359 tier where you have a lower amount of haste, where the bulk of the burst haste is useful and not wasted in meaningless haste over 50%.

As for trinkets, here's a bit of math I did on the official forums for the main trinkets straight out of heroics:

Witching Hourglass(Proc not personally checked)

285 intellect weighted at 2.9303 per point gives 835.1355 DPS.

Proc gives 1710 haste for 15 seconds, appears to have an ICD of 60 seconds and a 10% proc rate, meaning it's up every 75 seconds (about), giving 342 haste on average for the entire fight. This gives 357.6636 DPS.

1192.7991 DPS overall.

Anhuur's Hymnal

1710 spellpower proc for 10 seconds. 45 second ICD with a 10% chance on spellcast meaning it procs every 60 seconds, on average, thus giving 285 spellpower on average. This gives a 639.426 DPS increase.

285 Hit. Assuming all hit is used, this is a 533.093 DPS increase from the passive stat.

1172.519 DPS increase total.

Sorrowsong:

1710 Spellpower, 10 second buff with 100% proc rate and 10 second cooldown, 50% uptime under 35%.

1710 x .5 x .35 = average of 299.25 spellpower increase for the entire fight, giving you 673 DPS.

285 mastery valued at 0.8497 DPS per point gives you a total of 242.1645 DPS.

915.1645 DPS increase overall

Tendrils of the Burrowing Dark:(Proc not personally checked)

1710 spellpower, 15 second buff with 10% proc chance and (supposedly) 60 second cooldown. Assuming a destro rotation of about 1 spellcast every 1.5 seconds, that's a proc about every 75 seconds, or a 20% uptime.

1710 x .2 = 342 spellpower increase for the whole fight, or a 767 DPS from the proc.

285 Mastery weighted at the same value gives 242.1645 DPS from the passive bonus.

1009.1645 DPS increase overall

Gale of Shadows

17 Spellpower, stacking 20 times. When full stacked, it gives 340 spellpower, or a net 762.824 spellpower increase.

285 haste valued at 1.0458 per point gives 298.053 DPS.

1060.877 DPS increase overall

Soul Casket

321 Mastery at 0.8497 per point gives 272.7537 DPS.

20 seconds of 1926 spellpower every 2 minutes gives 321 spellpower average for a 720.1956 DPS increase.

992.9493 DPS increase overall.

Stump of Time

321 hit valued at 1.8705 per point gives 600.4305 DPS.

Assuming the same proc rate as Anhuur's Hymnal, this works out to an average of 321 spellpower throughtout the fight, giving a 720.1956 DPS increase.

1320.6261 DPS increase overall.

NOTES:

On any fight where the boss starts with less than full health, Sorrowsong becomes more and more valuable. If the boss starts below 87.5 percent, Sorrowsong's spellpower bonus will trump that of other 346 ilevel trinkets. Similarly, any fight with adds which the warlock will be killing (Magmaw, Wyrmbreaker, Etc), or any fight with two or more bosses which don't share a health pool (Conclave of Wind), Sorrowsong's proc will again beat the other two 346 trinkets.

Factoring in build up time for Gale of Shadows isn't required, because it procs off of hellfire (even if hellfire hits no targets). Therefore, a warlock can (and probably should) have zero ramp up time on almost every boss. If there is ever a boss fight where for 20 seconds or more the warlock will not be DPSing any target, the value of GoS drops dramatically.

Passive Haste/Mastery can and should be reforged to hit or crit, depending on what you need. If you do this, Sorrowsong, Gale of Shadows, Tendrils of the Burrowing Dark, and Soul Casket all increase in value.

GoS refreshes every time your dot does. So the boss would have to wipe your dot and disappear for 20 seconds. But if that happened you have plenty of time to LT and hellfire to maintain stack.

Sorrowsong doesn't improve Halfus DPS unless you are using the adds low health to multidot halfus. If you aren't multidotting halfus then it is the same as if it was all 1 boss and you are just getting your proc at a different time in the fight.

What makes Sorrowsong better than a lot of other trinkets though is when you get your damage increase. Chimearon, Cho'gall, Ascendant Council all have some steep burn phases. So even though the trinket wouldn't necessarily be better than the other heroic trinkets over the fight you are getting an effective 855 spell power during the only part of the fight where you are seriously screwed the longer it goes on. In this sense for those particular fights its best to use the 855 spell power number for determining which trinket to use.

I don't know about your scalers, but when I replace Sorrowsongs 300 SP with 855 SP it makes it by far and away the best trinket in the game for low health burn phase mobs by over 500 DPS (including heroic raid trinkets) during the zerg phase.

Another example when fight mechanics dictate a different approach than simcraft modeling.

I did a quick test with the Darkmoon Card: Volcano trying to get some numbers. Equipped with my pre-raid gear (without the cloak, because of the embroidery; affliction specced) I procced the trinket on the raid training dummy; with and without Fel Armor.
The tests showed an average fire damage of 1462 without Fel Armor, and an average of 1544 with Fel Armor. While I know that my tests are only a very small sample size, this could be a sign for spell power scaling.

Other interesting observations: the 1600 intellect from the proc do increase your maximum mana, but not your current mana. E.g. if the trinket procs at 50k/66k mana, you stand at 50k/90k mana while the proc lasts. Could be interesting for spells or effects that base on your maximum mana.

The trinket seems to only proc from direct damage speals. Based on the tooltip, I thought it would proc also from DoT damage, but this seems not to be the case. I wasn't able to proc it once while only holding corruption, unstable affliction and bane of agony up on the target. With searing pain casts, the proc usually occured within 3 casts.

Thanks for the testing - the only missing pieces of information right now is whether or not it scales with metamorphosis, demon soul: felguard, and the personal damage bonus part of demonic pact. But I guess this is the wrong thread for that.

@Apollexis:

This is neither a bug nor is it news. Haste rolls for the entire duration of the dot. Unlike damage or crit modifiers, haste modifiers won't be recalculated with every tick.

EDIT: I stand corrected, all player buffs/debuffs roll for the entire duration of the dot.

I would have no problem believe this except when i hover over immolate it says X damage every 2.3 seconds, and when the soulfire debuff wears off it says ever 2.7 seconds, Which inciniuates that they did not intend this.

EDIT:

The point im trying to get across, It would be a good idea to maybe add to the first post? About refreshing your immolate with cast or fel flame after you refresh your soulfire buff or gain any haste/crit modifier. I don't think alot of people know it still works like this.

You're right, the only modifiers that refreshes on every tick are debuffs on the target.

Obviously the tooltip would change, it always shows you the properties of a spell if you'd cast it right now. This does not at all suggest any of Blizzards intention for recalculating dots on every tick.

You're already not supposed to cast any dots unless ISF is up and ISF has top priority in any rotation, so there's really not much point in adding anything to the first post. Also, I doubt that it would be a DPS increase to recast Immolate whenever you gain a buff.

Hello all,

Picked up Theralion's mirror , which i spent about 15 min wondering if i should spend my gp on. The 321 static int was incredibly tempting but i am trying to find math on how to proc (1926 mastery) works out for us. Any help would be appreciated. Thinking about replacing the JC trink.

This is a chart I generated for myself a while ago, simulating profiles with a single trinket compared to a baseline profile with no trinkets:


The baseline trinketless profile was far from hit capped, so every trinket was reforged to hit where possible.

Yes, just like when using simulation to generate scale factors, this is only accurate for the particular gear setup used in the baseline. No matter what I do I can't provide a simple chart that's accurate for everyone.

I do think I can provide a more generally useful list by hit-capping the baseline profile and reforging the trinkets to haste wherever possible, but then there's other things to worry about, like haste thresholds, etc. Even if I removed all stats from all the trinkets and just simulated the value of the proc, the results would still be dependent on the exact gear set of the baseline profile, particularly for trinkets with haste procs.

So sure, flawed results, but this stuff will always be flawed unless you do it for yourself, importing your own character from the armory.