[Priest] Shadow DPS Guide

[Vampiroth]

I've moved the spell priorities around in simcraft to where VT, and SW: Pain were above DP3, and it showed a DPS loss. test it for yourselves.

1 DPx3 Did 178.1k damage (FULL SPELL, DIRECT DMG AND DOT)
1 Mind bender did 163.8k damage (FULL MINDBENDER, NOT 1 SWING)
1 SW: Pain did 70.1k damage (FULL DOT NOT TICKS)
1 VT did 65.9K damage (FULL DOT NOT TICKS)
1 Mind flay did 42.1k damage (FULL CHANNEL, NO CLIP, NOT TICKS)
1 shadowy app did 25k damage
1 halo did 72k damage. (1 HIT, NOT MULTIPLE)

So when you use that 1 GCD of DP over refreshing a VT or SW: Pain, you hardly lose any damage, but you gain the massive direct damage from DP, plus the dot damage of it, then right after that GCD you get to put sw: pain back up, then VT. Also, Getting DP up first does more damage than mind bender. AND it does more damage than halo.

In theory, the logic behind getting DP up faster is with procs over the entire duration of the fight, you could get an addition, or multiple additional DP's up over than if you waited. More DP's, More damage overall. It's just like mind flay clipping where you are trying to get more ticks of MF.

[Kaesebrezen]

I've moved the spell priorities around in simcraft to where VT, and SW: Pain were above DP3, and it showed a DPS loss. test it for yourselves.

Moving VT/SW:P above DP is a DPS loss in simcraft. You have to change the top priority DP line to include MB/SW:Smiley cooldowns.

That actually doesn't result in any change on DPS, simply because in a 8 min duration fight you won't get an additional Halo nor Mindbender, and Simcraft's DoT Uptime is already "perfect", so there isn't much to improve there.

[Vampiroth]

Doesn't really matter what simcraft says, it's fairly clear using a gcd on dp > refreshing a dot or using anything else.

[mindp]

Wow, way to not acknowledge me, Derevka!

He was expecting you to be watching iCarly.

[Juicebox]

Once and for all, it IS ok to delay your DP too, for example, refresh your dots as long as you don't lose out on an orb doing so.
Saying anything else is wrong. You're not losing any damage at all by delaying that DP. You do, however, lose damage not refreshing your DoTs.

Also, if you're simming, make sure VT and SW:P only has priority over DP3 until 25%.
After that, the orb generation is too high.

[Erto]

Doesn't really matter what simcraft says, it's fairly clear using a gcd on dp > refreshing a dot or using anything else.

It's clear you're not reading what we're saying, so this will be the last thing I say.

Delaying Halo or Mindbender results in a DPS loss if it means you get less of them in across the fight.
Letting DoTs fall off results in a DPS loss.
Delaying DP (as long as you're not losing out on Orbs and therefore further DPs) doesn't make a difference to how many you get in across the fight.

Therefore, delaying DP so you can refresh DoTs, cast Halo or cast Mindbender instead will give you a DPS increase. It seems like you're thinking way too much into just the damage of the one GCD/Ability, instead of damage across an entire fight.

Also, I mentioned something about the Divine Insight talent earlier, which can of course affect if you'll want to hold off on DP in case you do get a Mind Blast procc, but I don't have the Maths to suggest if you'll want to cast DP straight away just in case you get a procc. I haven't looked at the Mists fights too much yet, but I'm still not too sure on the Divine Insight Tier being a 'One Talent Fits All' Tier. There's likely to be fights where you're going to want the burst of Power Infusion, your Guild is working on the execute phase or there's a regular flow of adds that need killing.

Even if not, it's also likely they'll reduce the damage on Devouring Plague and increase the damage of Vampiric Touch and Shadow Word: Pain. I know this is not what we're talking about and shouldn't be basing theorycraft around, but it is very likely that it'll happen, because at the moment a spell we only use around every 28 seconds (less with proccs) in >20% phases is doing way too much of our damage (thinking more of PvP) and making one talent a very obvious winner.

[Vampiroth]

Not taking divine insight would be completely asinine. Why are you discussing theories you haven't actually tested?

I'll say it again, 1 halo or 1 mind bender does not equal or do more damage than 1 DP. So if you get 1 DP in place of 1 halo or 1 mind bender you do MORE damage. It's really not that complicated.

And just to clarify, I'm not saying don't use halo or mind bender, I'm saying that if its a choice between the two, you should go with DP.

[Zarqa]

Vampiroth, what they are trying to say is that if you look at 1 single GCD, yes DP will do the most damage. But if you look over a longer period of time, the TOTAL damage done will increase because you will still have cast the exact same amount of DPs in that period, but have had more DoT ticks because they never fell off. Whether or not you would have had more halo or mindbender casts will depend on how long you set the period.

They are really doing a good job trying to explain it to you, with excellent examples. You then replying that the opposite is 'fairly clear' without any examples or proof sounds condescending to me in all honesty. You clearly don't understand what they are saying, but apparently also are not trying to understand it.

Once again: all they are trying to explain is that with their priority list you will cast exactly the same amount of DP, but you will have more uptime of the other spells. The only tricky part in their equation is the orb generation from Divine Insight. They are absolutely NOT trying to argue that DP is not your biggest damage spell. It clearly is.

[Vampiroth]

The most important part of the equation IS Divine Insight. If you have tested it at all on beta, you would know it can proc back to back to back. Also, sub 20% orbs are generated by SWeath as well.

Clearly refreshing dots would be better without Divine Insight, as you have nothing to worry about at all because MB is for sure not going to come back up.

[Brusalk]

Please read what I posted.

Even with divine insight, as long as you don't get a second DI proc before you can MB again, you haven't lost anything.

Here's a simcraft spell-priority that has it at a wash either way, but because of that there's situations where you can potentially gain dps in a real fight by delaying.

actions=shadowform
actions+=/volcanic_potion,if=buff.bloodlust.react|target.time_to_die<=40
actions+=/devouring_plague,if=shadow_orb=3&(!ticking|remains<tick_time)&(buff.divine_insight_shadow. up|cooldown.mind_blast.remains<1.8)&target.health.pct>=20
actions+=/shadow_word_pain,if=(!ticking|remains<tick_time)&miss_react&!(buff.divine_insight_shadow.u p|cooldown.mind_blast.remains<1.8)&target.health.pct>=20&shadow_orb=3
actions+=/vampiric_touch,if=(!ticking|remains<cast_time+tick_time)&miss_react&!(buff.divine_insight_ shadow.up|cooldown.mind_blast.remains<1.8)&target.health.pct>=20&shadow_orb=3
actions+=/mindbender,if=cooldown_react&!(buff.divine_insight_shadow.up|cooldown.mind_blast.remains<1 .8)&shadow_orb=3
actions+=/devouring_plague,if=shadow_orb=3&(!ticking|remains<tick_time)
actions+=/shadow_word_death,if=target.health.pct<20
actions+=/berserking
actions+=/mind_blast,if=cooldown_react
actions+=/shadow_word_pain,cycle_targets=1,max_cycle_targets=5,if=(!ticking|remains<tick_time)&miss_ react
actions+=/vampiric_touch,cycle_targets=1,max_cycle_targets=5,if=(!ticking|remains<cast_time+tick_tim e)&miss_react
actions+=/mindbender,if=cooldown_react
actions+=/mind_sear,chain=1,interrupt=1,if=num_targets>=5
actions+=/mind_flay,chain=1,interrupt=1
actions+=/shadow_word_death,moving=1
actions+=/mind_blast,moving=1,if=buff.divine_insight_shadow.react&cooldown_react
actions+=/shadow_word_pain,moving=1
actions+=/dispersion



If you still don't get that it's not a dps loss to hold off on DP unless you waste an orb gen, then I don't think anything we can say can change your mind.

[lolalola]

I'll try and explain what they're saying Vampiroth.

Devouring Plague should only be cast while we have 3 orbs. As a result, after getting 3 orbs we have at 8 seconds before we can get an orb again, at which point we want to have cast DP already so we don't potentially waste either MB casts/fight or orb generation.

Now, spells like Halo which have a GCD are limited in the max number of times they can be cast per fight based off of how close you actually are to casting them on GCD. What they're saying is that if you hold off on casting DP for a DoT refresh, or casting a CD spell like halo or mind-bender, you're not losing any potential damage from Devouring Plague but you are gaining potential damage from being able to cast a CD spell more in a fight or having higher dot uptime. I actually agree with them that because of the orb generation DP+3 shouldn't be highest on priority, and instead CD spells or other dots should be higher. However, that only holds for the first 6 seconds(4 gcds) after getting 3 orbs.

Now, where it gets a bit complicated is that if you have divine insight, rather than it being a set "DP+3 is lower on priority than x" for 4 gcds after getting 3 orbs, you have to be concerned with getting 2 DI procs in those 6 seconds. As haste increases it gets more and more likely that we will get a proc in those 8 seconds from a SWP tick.


Probability of getting a proc off of one tick: 15%

Worst case we have 3 ticks of SWP in those 6 seconds. (first just after the MB cast that gets us 3 orbs)

Probability of getting getting 1 or fewer procs in three ticks: 93.9%
Probability of getting 2 or 3 procs in 3 ticks: 6.1%

As you can see, the probability of getting 2 or 3 procs in those 6 seconds is quite low. (The thinking here is that if you get a DI proc you have until you get a second one to use it up, otherwise you lose the first one. Also, I'm guessing that the number of times you'll get a DI proc mostly negates the effect of the case where you don't get a DI proc before the CD on MB comes up from the first proc'd cast)

As a result, it's most likely still a DPS increase to cast a dot refresh or a cd spell before DP until you get a DI proc, at which point you should cast DP and then MB to reduce the chance of munching the first DI proc with the second.

Can someone explain why the orb generation from DI is completely RNG whereas the one from Nightfall apparently isn't?

What I don't understand is why you say probability of getting 1 or fewer procs in three ticks. If you get no proc, then its alright, but as soon as you get one proc while having 3 orbs that is a general DPS loss over perfect RNG (meaning it'd proc right after you popped DP). Every time it does proc and you use Mind Blast your CD on MB is max again so the sooner you use MB again the higher the chance is DI procs (if DI procs with MB on CD you're getting nothing out of it) so that you're getting quicker towards 3 orbs (that it s the goal so to say). What someone who is good at maths has to calculate is what is the DPS loss of missing out on one orb (you got 3 orbs and DI procs) versus the lack of reapplying one or two of your DoTs (so SW:P, VT, or both). This is completely besides the fact that DP is a short burst dmg + DoT and that the chance of say popping DP and then having 3 orbs again within 6 seconds and then not being able to pop it again because it is still up. Instead, where I get confused in my rotation is what I should do when my DoTs are about to fall off with while I get just gotten my 3rd orb. Should I have reapplied my orbs earlier? [EDIT]And what about having 2 orbs, should I then reapply my DoTs earlier to avoid having RNG biting me?[/EDIT]

Also if you have a fight with trash, Twist of Faith could proc there with also SW:D which allows more DPs. So the Twist of Faith mechanism would in these cases have more uptime than merely on execute phase. Think: Sorrowsong on Ragnaros.

PS: To make SW:D show as-is instead of as SW you can use [noparse]SW:D[/noparse] on UBB-compatible forum.

[Yvaelle]

Sorry for not participating up until this point, didn't check thread and it exploded overnight

Delaying Halo or Mindbender results in a DPS loss if it means you get less of them in across the fight.
Letting DoTs fall off results in a DPS loss.
Delaying DP (as long as you're not losing out on Orbs and therefore further DPs) doesn't make a difference to how many you get in across the fight.

All these things are true.

If you still don't get that it's not a dps loss to hold off on DP unless you waste an orb gen, then I don't think anything we can say can change your mind.

I'm pretty sure Vampiroth isn't denying that, his point is that the *risk* of a lost orb (via Divine Insight) itself outweighs the benefit casting Mindbender or Halo precisely on their GCD. There is a decent argument to be made in his favour as well, but I haven't been able to get simcraft to do it the way I want it yet, but I will try to explain it in words:

The long cooldowns on Halo and Mindbender mean the risk of having them on cooldown at the end of the fight is very high. In a 3 minute fight, you get 3 mindbenders. In a 3m59s fight, you get 3 mindbenders. The same effect occurs with Halo. The same effect is Not the case with Devouring Plague however. This means the quantity of Mindbender and Halo in a fight have more to do with the precise length of the fight than on their position in your rotation.

Because Devouring Plague has no cooldown, and such a short duration - the potential room for improvement on uptime is very, very significant. With it being such a colossally high DPET - the mere *risk* of a lost orb, or one third of a Devouring Plague may be enough (can't say until I can prove it) to make Mind Blast / DI procs more valuable than even our burst cooldowns (Mindbender / Halo).

Just to make this transparent (but not to suggest that Divine Insight isn't an obvious choice for pretty much all single target fights), if you didn't have Divine Insight - you would cast one DP every 25 seconds (8 second MB cooldown + subsequent GCD) for 6 seconds, with an uptime of only 24% - but still accounting for probably 25% of your DPS. This means that in a single target fight its probably close to impossible to be at risk of overlapping Devouring Plagues: every devouring plague will have full uptime.

Assumptions:
- Divine Insight
- (Hasted) Shadow Word: Pain on a single target
- Mastery will work as intended before live (procs ticks will proc DI)
- Full free pvp set on beta as with all other numbers (18.11% haste, 9.3% chance of mastery proc ticks)

8 ticks (including initial) per 18 seconds, 4 ticks (including initial) for the subsequent 7 seconds. That's 12 ticks, with a 9.3% chance of one of those getting a mastery proc, so the average is going to be around 13 potential DI procs per 25 seconds single target. So theoretically that means we could actually see near 100% uptime on DP - albeit it will be super rare. (But that means get ready for the other classes crying about all the WoL top dps parses being shadow because someone lucked out and had 100% uptime on a spell designed around 25% uptime - and dealt 75% higher DPS than what we were balanced around).

So we need a distribution for the likelihood of 15% DI proc rate on 13 average ticks - but I can already tell that's going to be a pretty likely occurrence. That, to me, is what I think Vampiroth is concerned with.

So having thought all that through, I can actually simplify that argument much better I think.

If you lose just 3 DI procs per fight (actually pretty likely if its not your top priority) then you lose one Devouring Plague (we can make caveats here, but I think that's pretty fair to say). If one Devouring Plague has a DPET of 180k, and Mindbender has a DPET of 157k, then even if you lost the full duration of a Mindbender, you would still be better off having got that one extra Devouring Plague in (better off by 23k). What's more, since I calculated DPET *without* mastery factored in, mastery actually increases this discrepancy since it benefits DP but not Mindbender or Halo.

I want to be able to sim it before I feel comfortable with it, but I think I'm siding with Vampiroth on this so far?


Edit: the more I think about this, the more the model itself bothers me. It means that the few massive but rare DPS parses are going to draw a lot of attention - and the likely response will be that shadow gets nerfed to be usually slightly under everyone else (like fire was in tier 11 and 12 before it got buffs). It makes our single target dps really random and less skill-based like in past expansions. It also means that in multi-target fights with the potential for big spikes in orb generation through multi-Paining things and varied execute times that remove the risk for overlapping DPs (since you can just put it on the other target if its already up) - we could see big multi-target DPS again just like Live, but for a very different reason: rather than because our dots are so big, because two targets = doubled resource generation for our single target damage.

[Brusalk]

Can someone explain why the orb generation from DI is completely RNG whereas the one from Nightfall apparently isn't?

What I don't understand is why you say probability of getting 1 or fewer procs in three ticks. If you get no proc, then its alright, but as soon as you get one proc while having 3 orbs that is a general DPS loss over perfect RNG (meaning it'd proc right after you popped DP). Every time it does proc and you use Mind Blast your CD on MB is max again so the sooner you use MB again the higher the chance is DI procs (if DI procs with MB on CD you're getting nothing out of it) so that you're getting quicker towards 3 orbs (that it s the goal so to say). What someone who is good at maths has to calculate is what is the DPS loss of missing out on one orb (you got 3 orbs and DI procs) versus the lack of reapplying one or two of your DoTs (so SW:P, VT, or both).

I say 1 or fewer procs because as long as you don't overwrite an existing DI proc with a second, you can cast the DI proc after keeping higher uptime on DoTs. The assumption I'm making is that if you have DI, the chance of getting a DI proc before the next cd on MB is up kind of negates each other. In general DI complicates things, but here's my thinking: If you get a DI proc within 1 or 2 GCDs of getting 3 orbs, you have a GCD to react to queue up a DP after your current cast and then cast MB without losing any orbs.

Basic Illustration: x = swp tick, d = di proc, m = mb cast, _ = empty gcd where you can cast something, --- = a unit of time
So say you don't get a DI proc:
m --- _ --- _ --- DP --- flay --- m
In this case, you have no issues with anything, same number of DP/orb gen and you have higher uptime/cd usage.

Now say you get a DI proc after you've started filling up that first empty gcd. That means the next tick of SWP (and thus next chance of getting another proc) is at least 1 gcd in the future. Thus you can react to getting that proc and DP for the very next gcd, allowing you to consume that proc before there's even a chance at getting a second one. That being said, you will lose some time in the case where you get no DI proc before the natural cd resets. Needs to be seen whether higher uptime outweighs that potential case.


@Skilloverkill:

If you lose just 3 DI procs per fight (actually pretty likely if its not your top priority) then you lose one Devouring Plague (we can make caveats here, but I think that's pretty fair to say). If one Devouring Plague has a DPET of 180k, and Mindbender has a DPET of 157k, then even if you lost the full duration of a Mindbender, you would still be better off having got that one extra Devouring Plague in (better off by 23k). What's more, since I calculated DPET *without* mastery factored in, mastery actually increases this discrepancy since it benefits DP but not Mindbender or Halo.

I guess we're arguing slightly different things. I completely agree with you that losing DI procs is a very, very bad thing and losing even one will most likely outweigh any potential benefit from holding off on casting DP.

That being said, I guess I'm arguing that even with DI procs at every opportunity there's no reason (single target) as to why we would have lost a DI proc. At most we've lost some slight amount of MB casts in the one case where you don't get any DI procs before the natural CD of MB comes up. (And my gut feeling is that the increased uptime will outweigh the loss in that one very specific case. Probability with 3 ticks of getting 0 is 61%. Getting mastery procs only decreases that chance)


I hope that made sense to folks

[Yvaelle]

@Skilloverkill:


I guess we're arguing slightly different things. I completely agree with you that losing DI procs is a very, very bad thing and losing even one will most likely outweigh any potential benefit from holding off on casting DP.

That being said, I guess I'm arguing that even with DI procs at every opportunity there's no reason (single target) as to why we would have lost a DI proc. At most we've lost some slight amount of MB casts in the one case where you don't get any DI procs before the natural CD of MB comes up. (And my gut feeling is that the increased uptime will outweigh the loss in that one very specific case. Probability with 3 ticks of getting 0 is 61%. Getting mastery procs only decreases that chance)


I hope that made sense to folks

Makes sense to me, I can still see it going either way though - and we probably won't know conclusively until we have a model for it. I guess the bigger risk then isn't so much even the proc risk a second DI proc at three orbs (though this risk does skyrocket with multi-dotting which we haven't covered yet) - but the long cooldowns on Mindbender and Halo. If your guild kills Ultraxion in 6 minutes on progression, its pretty much impossible to suddenly get a 5 minute kill the next attempt - or a 7 minute kill (past enrage). If you always use Mindbender precisely on cooldown, then you get 5 in, or if you wait 10 seconds to use it after every cooldown - you still get 5 Mindbenders into the fight (the same is mostly true for Halo).

So even if the risk is small that you would lose an MB proc by waiting on it, if you knew the rough length that your guild was going to take to kill a boss - the risk of losing Halos or Mindbenders by waiting is mostly removed, but the risk of losing Devouring Plagues can never be removed because its RNG. So all the above arguments so far are predicated on not knowing the rough length that a fight will take you to kill - but in practice we will know roughly how long the fight will take - and that should give us a pretty big window (48 seconds of downtime on Mindbender per minute, 39 seconds of downtime on Halo per minute) to always get the same number of cooldowns.

Changing my answer to the more politically correct if less assertive - elvish answer - "I think it's definitely one of the two answers "

[Slashgordon]

Has anyone had any thoughts regarding clipping Mind Flay? With MF not really doing anything beyond damage, is it worth clipping it if, say, you get a divine insight proc that would in turn give you a third shadow orb? Is it worth clipping it if SW:P / VT would fall off before you could refresh them? Is it worth clipping if you enter the execute phase mid channel?

Would it be a situation similar to how aff locks used to play, where they would clip drain soul to refresh haunt / Shadow Embrace / UA (pre cata changes) but only do so immediately after a tick? Or is clipping MF a straight no no?

[Danii]

Has anyone had any thoughts regarding clipping Mind Flay? With MF not really doing anything beyond damage, is it worth clipping it if, say, you get a divine insight proc that would in turn give you a third shadow orb? Is it worth clipping it if SW:P / VT would fall off before you could refresh them? Is it worth clipping if you enter the execute phase mid channel?

Would it be a situation similar to how aff locks used to play, where they would clip drain soul to refresh haunt / Shadow Embrace / UA (pre cata changes) but only do so immediately after a tick? Or is clipping MF a straight no no?

Yes, it is worth clipping, you should have a 99% uptime at least on all of your DoT's, which sometimes requires you to clip your MF, however, you should practice how and when your MF clips so that you are prepared for it.

[Vampiroth]

Skill explained what I'm saying, also you said you lose mind blast casts.... Losing 3 mind blast cast = losing 1 dp.

1 dp > 2 vts, or 2 sws, or 1 mind bender, or 1-2 halo.

Honestly, the rotation isn't going to stay the same. If it does, we will be vastly underpowered compared to other classes if they don't get nerfed. But realistically, they should bring us up vs bringing multiple classes down.

[Yvaelle]

Honestly, the rotation isn't going to stay the same. If it does, we will be vastly underpowered compared to other classes if they don't get nerfed. But realistically, they should bring us up vs bringing multiple classes down.

I always think this is the way to find balance - keep the classes that are working fine just as they are - buff the classes that are underperforming, tune bosses based on balanced classes.

Ghostcrawler always seems to respond to this by claiming that would make everyone OP though.. which is mildly infuriating (because making everyone perform roughly equally is the exact opposite of making people OP).

[Felade]

Ghostcrawler always seems to respond to this by claiming that would make everyone OP though.. which is mildly infuriating (because making everyone perform roughly equally is the exact opposite of making people OP).

he does have a point - we don't want an arms race either. He's also said that if there's one spec underperforming and the rest are at one level, they'll bring up the underperformer (I believe that was in reference to t11 resto shammy).

[Erto]

I always think this is the way to find balance - keep the classes that are working fine just as they are - buff the classes that are underperforming, tune bosses based on balanced classes.

Ghostcrawler always seems to respond to this by claiming that would make everyone OP though.. which is mildly infuriating (because making everyone perform roughly equally is the exact opposite of making people OP).

No it's not. Classes aren't just balanced around eachother, they're balanced around bosses too. If everyone's "overpowered" then yeah, there's balance between the classes, but it makes bosses easier than they should be. Same goes the other way by making fights harder than intended by nerfing classes to the same level as the others.

Also, with regards to Vampiroth about the Divine Insight talent, on the whole it's of course the best talent, by a long shot. But that doesn't mean it will always be used. If you're progressing on a fight with it's last phase starting at 30%, and that's the bit your Guild is working on and the rest of the fight is going fine, Divine Insight might help you top the metre by giving you more damage overall, but Twist of Fate will help your Guild actually get the kill, by providing more damage to the boss during the difficult part.