The Value of a Soul Shard

The Warlock community has always had an excellent group of theorycrafters, partly because the community could always count on SimulationCraft to do a lot of heavy lifting, and Zakalwe, a Warlock, is a very active SimCraft developer. Since MoP, however, the changes to Warlock mechanics made it so that SimCraft has struggled to provide comprehensive answers about optimal Warlock play. A large part of this is the introduction of Pandemic, making DoT refreshing more of an art than a science; the other part of this is because Warlocks are an ‘actual’ resource based class for the first time – Affliction has Soul Shards, Demonology has Demonic Fury, and Destruction has Burning Embers.

To compound the problem, these resources aren’t like the other resources in MoP – the build up for them is fairly long and slow, and the proper use of these resources dictates that they should be pooled. So while Hunters can calculate the marginal value of a point of focus just by looking at the DPS of a rotational ability like Arcane Shot, figuring the marginal value of a Soul Shard is not quite as simple.

Why is it important?

Up until this point, knowing the value of a Soul Shard wasn’t particularly important – it’s nice knowledge to have, since it allows us to craft more efficient sequences for situations such as the opener or in making decisions like Soulburn: Soul Swap versus Haunt, but now that 5.2 details are slowly being released, not being able to peg the value of our resources is a huge hindrance to the community.

The early details and iterations of the Tier 15 Warlock set bonuses focuses on these resources. Whereas in this past tier, Tier 14, we saw a very simple 2-piece (rotational spell does more damage) and a cooldown-based 4-piece bonus (Dark Soul cooldown reduced by 40s), the Tier 15 bonuses up to this point are almost all resource based:

Revised Affliction T15 2P Bonus:​ The Tier-15 2-piece set bonus now causes Dark Soul to increase the duration of Haunt by 4 seconds (instead of the previous Nightfall modifier).
Revised Demonology T15 4P Bonus: Increases Demonic Fury from all spells by 10% (up from 5% on certain spells).

Initial Warlock T15 2P Bonus: While Dark Soul is active, your chance to trigger Nightfall is increased by 30%, Chaos Bolt and Shadowburn consume 20% less Embers, and abilities cost 30% less Demonic Fury.
Initial Warlock T15 4P Bonus: Your Malefic Grasp and Drain Soul deal 5% more periodic damage, your Incinerate has a 8% chance to generate twice the normal Embers, and your Shadow Bolt and Soul Fire grant 5% more Demonic Fury.

A Piece of the Puzzle: The Value of Haunt

Warning: a lot of Math follows

There are 2 things you can do with Soul Shards in any normal, Patchwerk-esque fight as Affliction – Haunt, and Soulburn: Soul Swap. We’ll look at the latter later – for now, we’ll consider a shard as equivalent to an extra Haunt. Figuring the DPS value of any given Haunt goes something like this:

  • Haunt – base damage: 1869, spellpower coefficient: 175%, sacrifice modifier: 50%, buffs all other damage by 25%, cast time: 1.5s
  • Malefic Grasp – base damage: 214, base spellpower coefficient: 20%, base DoT damage: (280 + 214 + 256) / 2 = 375 * Mastery Modifier, DoT spellpower coefficient: 35% * Mastery Modifier, sacrifice modifier: 50%, tick time: 1s
  • DoTs (assume fully stacked Agony) – base damage: 280 + 214 + 256 = 750, spellpower coefficient: 70%, tick time: 2s

In addition, on the other side, you give up Haunt’s cast time worth of Malefic Grasp ticks. The ratio of Haunt’s cast time to MG’s tick time is 1.5 ticks.

So if we cast Haunt on a target with a full set of DoTs, then MG for the duration of the Haunt, the equation looks like this:

Haunt Damage + .25 * ((MaleficGraspTick + MaleficGraspDotDamage) * (Ticks - 1.5) + DoTTickDamage * DoTTicks)

Add in secondary scaling:

(1+1.03 * CritChance)*((1869+1.75*SP)*1.5) + .25 * (1.5*(214+.2*SP+(375+.35*SP)*MasteryMod) * (8 * HasteMod - 1.5) + (750+.7*SP)*MasteryMod * 4*HasteMod)

and reduce:

(1+1.03*CritChance)*(1.4* HasteMod* MasteryMod* SP+1500* HasteMod* MasteryMod+0.4* HasteMod* SP+428* HasteMod-0.13125* MasteryMod* SP-140.625* MasteryMod+2.55* SP+2723.25)

The numbers are slightly different with the T14 2-piece and the T15 4-piece, but I’m going to ignore that for now.

My current stats are 16501 INT, 25081 SP, 6638 Haste, 2578 Crit, and 6499 Mastery. Well Fed, Flasked and Raid Buffed, my Haunt with no procs is worth 252.5k damage. 1 extra Haunt for the entire fight (of 450s length) is worth roughly 560 extra DPS. 1 extra Haunt every 2 minutes is worth 2100 extra DPS. SimulationCraft puts my DPS at 119k, so 2100 DPS is about 1.77% of DPS.

There are a few caveats when it comes to the equation above – for and against more DPS. It counts fractional ticks as real DPS for one, which is a big issue, but it also does not account for procs at all (a fully procced Potion + DS + Heroism Haunt is worth 602k for me, as a comparison). As the tick benefit is rather low in the “base case” Haunt (252k total damage, but the Haunt direct damage is worth 132k, and you’d only be losing a fraction out of 1 tick out of 8 or more total), I think the 2100, “no proc” number is very close to the lower bound for my current stats.

The Second Piece: The Value of Soulburn: Soul Swap

What about the other situation – if we use our Soul Shard for Soulburn: Soul Swap? What you essentially gain when casting a Soulburn: Soul Swap is that you get 2 GCDs, the GCDs that would go into casting whichever 2 DoTs you don’t cast with the first GCD. The value of those 2 GCDs, at the low end, is 2 GCDs worth of MG and then the damage from Soul Swap. The equation for that:

1.5 * 3 * (MaleficGraspDamage + MaleficGraspDoTDamage)

and with secondaries:

(1 + 1.03 * CritChance) * 1.5 * 3 * ((214+.2*SP)+((375+.35*SP)*MasteryMod))

and reduced:

(1+1.03*CritChance)*(1.575* MasteryMod* SP+1687.5* MasteryMod+0.9* SP+963)

Note that there’s no Haste in here – Haste in this situation is ignored because 2 GCDs is always equivalent to 3 ticks of Malefic Grasp.

With these two equations, we can compare the 3 ticks we’re gaining of Malefic Grasp with the value of the Haunt. With the same stats as above, those 3 ticks of Malefic Grasp do 139k damage. This is clearly worse than just using the Soul Shard to Haunt. In the fully buffed example, those 3 ticks while fully procced, with a Potion, Dark Soul and Heroism are worth 227k for me, a gain of 88k over the baseline situation, whereas in the Haunt example above it was worth 387k.

A big reason for this is that Soulburn: Soul Swap doesn’t scale to Haste at all. While you cast Malefic Grasp quicker with more Haste, you would gain the same benefit from hardcasting your DoTs. No matter how you cut it, you save 3 ticks of Malefic Grasp. In addition, the gain is far more limited even at first blush – 3 ticks of Malefic Grasp as opposed to our hardest hitting direct damage spell and 8 seconds of 25% more damage.

So Why Do We Soulburn so much in our opener?

The short answer is: to take advantage of buffs. To figure out the benefit of doing something like this, consider the gap between SB:SS and Haunt. The optimal way to play Affliction is to pool Soul Shards for Dark Soul and/or buffs, so that we can Haunt and Malefic Grasp as long as possible. If we enter our buffed, Dark Soul phase with 4 Shards (and we do, for our opener), we should end up using 4 or 5 shards during it – 3 Haunts, a Soulburn: Soul Swap to start, and then an optional 2nd Soulburn: Soul Swap. The trade off is 1 Haunt during a period where no buffs are up. We can modify our equations from above slightly and set them equal to each other to figure out if this is worth it:

(1 + 1.03 * CritChance) * 1.5 * 3 * ((214+.2*(SP+X))+((375+.35*(SP+X))*MasteryMod)) * 1.25

=

(1+1.03 * CritChance)*((1869+1.75*SP)*1.5) + .25 * (1.5*(214+.2*SP+(375+.35*SP)*MasteryMod) * (8 * HasteMod - 1.5) + (750+.7*SP)*MasteryMod * 4*HasteMod)

The calculation is slightly different because there is 1 assumption we can make about those 3 MG ticks we’re bypassing during the opener – they will benefit from Haunt’s debuff effect.

In order to do this calculation, all the stat values need to be filled in. Filling in my values of 23.01% crit, 73.89% Mastery, 21.39% Haste and 30001 Spellpower, I can solve for X (or WolframAlpha can). The result is that if my procs are worth 13961 Spellpower, then it’s worth it to Soulburn that 2nd time in my opener. My opener procs a 2/2 upgraded Relic of Yu’lon, a 0/2 upgraded Normal Light of the Cosmos, Jade Spirit, Lightweave, along with a pre-pot, to give me 14148 extra INT, passing the threshold for it to be okay to SB:SS that second time.

In actuality, the threshold is a bit lower than this – INT conveys extra crit, and if your guild puts down a Skull Banner or Stormlash Totem to start the number should be much smaller. Just adding the Skull Banner, for example, the value of X drops to 7340 Spellpower for me.

A Small Tangent – The Fel Flame in the Opener

While I’m on the topic of the opener, I think it apt to mention Fel Flame in the role of this. Some people (me, Evrelia) choose to Fel Flame during the opener. Some other Warlocks (Sparkuggz, notably), do not. The math for this is not especially difficult, so I think it’s proper to walk through it quickly. Fel Flame itself does damage, and the other gain is 6 seconds of Corruption and Unstable Affliction – with buffs. If you choose not to Fel Flame in your opener, then you get 1.5 ticks (1 GCDs worth) of Malefic Grasp. Fel Flame’s stats are as follows:

  • Fel Flame – base damage: 762 to 841 (801.5 average), spellpower coefficient: 75%, sacrifice modifier: 50%, cast time: instant (1.5s GCD)

Doing this quickly with my own stats, Fel Flame’s direct damage is worth 87.8k damage during my opener. The 6 seconds of Corruption and Unstable Affliction are worth 224.3k damage. On the other side, 1.5 ticks of Malefic Grasp is worth 113.7k damage. There’s a small penalty in terms of mana for using Fel Flame as well (about 1/7th a GCD), but it’s clear that the gap is large enough to warrant using the Fel Flame.

Affliction’s Tier 15 2 Piece

In the course of this post, I’ve covered a possible number for the value of a Soul Shard, and translated it directly to the value of Affliction’s 2 Piece. But while this number (1.75% of DPS or so), might be good lower-bound value for a single-target fight or a Patchwerk fight, in a multi-target scenario the value of the set bonus approaches zero. We’ve seen a few fights this tier where Haunt is used very sparingly (Stone Guard, Wind Lord Mel’jarak come to mind).

The reason Haunt isn’t used very much is because Haunt’s value is tied to the amount of Malefic Grasp uptime. The higher the MG uptime, the higher the value of Haunt. Conversely, the value of Soul Swapping DoTs onto a target is inversely proportional to MG uptime. MG uptime decreasing indicates that you don’t have enough free GCDs to do everything that needs to be done – and it’s something that’s apparent in multi-target fights.

This is why the first iteration of the 2-piece was actually a bit stronger (despite calculations pegging it at .7 shards/2 minutes, as opposed to the shard/2 minutes of the current bonus), as it doesn’t rely on Haunt, a spell that gets severely devalued in the number of targets go up. It’s possible that designers meant for this to be the case – after all, Affliction is leading the pack on many multi-target fights, so this might be a way to keep that in control.

Advertisements

13 thoughts on “The Value of a Soul Shard

  1. The +4s on haunt does not translate to an additional shard per se. It is actually only worth 8 free seconds of haunt debuff, which is actually a bit less good than a shard (because of a Haunt’s own damage).

  2. Surutcra@Hyjal.eu my Hero from Wol…
    I have been studying your (and Brokestorm’s ) opener to try and emulate your dps ( I fail miserably but now I have a good opener at least )and it seems you do three 3 SS during it – one before Misery, one just after you got all the procs up (+ misery and racial), and and one juste before the first buff wears off (around mid Misery).

    I can see the logic behind your opener which is to have all buffs up for all your dots during the full duration of the Misery but is it really worth wasting One Haunt + a SS lasting merely 3 before being applied again? (I admit it’s efficient in practice, juste wondering it’s the VERY best possible opener)?

    Have you considered a variation where you pop misery+SS immediately, fel flame as soon as all procs are up? effectively saving one shard for an additional Haunt?

    • I only found out about that possibility not so long ago. I’ve actually discussed it with Brokestorm and we do not see any very noticeable discrepency between both of them. My guess is the fel flame opener is probably better unless you are too close to loosing procs because you are capped at 4 shards. Also from the first dot application to the refresh before procs fall off, the time is going to be 15-20s (roughly, 15s of lightweave + time to proc lightweave). You want to have all your dots above 50% base duration after that time, which is guaranteed if you used SB:SS but not with FF. Actually taking 18s and with FF, you “loose” 14/2 – (14+6 – 18) = 5s of fully buffed UA, as well as 18/2 – (18+6 – 18) = 3s of fully buffed Corruption. I’m eager to see what kind of insight rusjive here can give us on that matter !

      Opener :
      – 1.5s: pre-haunt
      – 0s: Dark Soul + long on-use effects + dots through SS from pre-pooled soulshard (ideally, you don’t want to throw short [<=10-12s] on-use effects here like orc racial because they may force you to refresh sooner than you would like at the end of your procs. Mine are still macroed, I'll try and separate them).
      – MG till all procs up
      – short on-use effects + refresh dots (with FF or SB:SS. If you already had a nightfall proc at this point you may prefer to use SB:SS, same if you are under BL, ultimately it does not really matter how you refresh your 2/3 dots)
      – MG & keep haunt up until procs are about to fade
      – SB:SS for full duration on all spells

  3. Unless i completely misread it you added haunts damage but forgot to add the MG dmg you would do when not casting that haunt and instead using SB:SS. Should make them pretty much equal with your stats.

    • I’m not quite sure what you’re saying. The article looks at 2 different values for a Soul Shard: Haunt, which is worth Haunt’s damage, the benefit from the debuff and -1.5 ticks of MG, and Soul Burn: Soul Swap, which is worth 3 ticks of MG (as compared to just hardcasting 3 DoTs), and concludes that the Haunt is worth more than the damage saved from SB:SS-ing. Can you clarify?

      • It´s a good thing i added the “unless i completely misread”. You´re right, I missed that you withdrew that damage from haunt where i thought it was to be added to MG. So disregard from that comment. Just me being too tired for my own good.

      • Although now that I know that I was mistaken in that part I think you´re defending that 2nd Soul Burn: Soul Swap in the opener too much. The alternative is “probably” not a completely unbuffed dot. For that too be true you would have to be able to keep haunt up for the first 40 seconds or so of the fight. That would require 7 Soul Shards, or 3 soul shards regenerated in the first ~30 seconds of the fight. I´m not certain that this is correct but I can say for a fact that you should atleast add 4000 int or so for the Haunt calculation to get a fair result (Atleast i´m not having full uptime on haunt even with fairly powerful dots (trinket + jade spirit for example))

      • You’re right in that I’m very charitable towards the second SB:SS. There are a couple extra things that I’d like to note however:

        1) In all calculations I’m leaving out the value of Haunt chaining, which is slightly better than letting Haunt fall off before re-Haunting. With an SB:SS there’s just no way to Haunt chain as you’re wasting too much time off of the Haunt debuff by hardcasting.
        2) As gear levels increase, the probability that you will have all the Soul Shards you need during the opener and 20 seconds afterwards increases due to more Haste. In this scenario, Malefic Grasp doesn’t get stronger (because of the 2 GCDs saved = 3 ticks equivalence). On the other hand, the probability of an unbuffed Haunt also decreases slightly (because you will get too many Shards in between every Dark Soul), although once we lose the T14 4-piece bonus that will help a bit as well.

        I’ve long suspected (after doing this calculation) that Warlocks that are in Normal, Raid Finder and Heroic blues might need to play differently on the opener. I think I’ll look into this subject a bit more. Thanks for engaging me in this discussion!

    • Thanks for the reply! One more thing though in favor of not using Soulburn: Soul Swap. That factor is heroism, without it I doubt you get enough shards to keep haunt up no matter how good your gear is and it also increases the value of Haunt. My pleasure to get into it. I know that I personally have been using Soulburn: Soul Swap way too much and this post was a pleasant wake up call.

  4. I believe it can be simplified a bit.
    Haunt is worth Haunt damage + 2 seconds of our normal dps (25% of 8 seconds currently). Most often last MG tick will not be buffed (can’t be perfectly lined up due to Haunt travel time), so a little less.
    SS:SB is worth 3 MG ticks.

  5. You would need to define “our normal DPS” to define precisely what you mean. As I see it, I would define it to be the DPS with haunt’s damage included, but without the added damage from the debuff. Then a soul shard used towards haunt is haunt_damage + 2s * ‘normal’_dps – 1.5 * MG_tick_damage.

    Rustjive’s method is fine to compute the optimal value of a shard (i.e. if you end up only casting MGs for the full debuff duration). Your method would rather give a lower bound on the value of a haunt, because it does not account for the fact that you actually try to haunt when you know you will be able not to refresh / LT too much for the debuff duration.

    I agree on the “value of a shard used towards SB:SS” ~= “value of 3 MG ticks” (non counting “second order” effects like the advantage you gain from refreshing all dots immediately as opposed to one after the other, say when wanting to immediately take advantage of a new proc).

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s