Shadow Priest 10.1 - Talents and analysis

Classes Priest
1

I have spent some time playing with Shadow Priest in 10.1 PTR and I would like to give some thoughts on some of the talent builds going into the patch.

Psychic Link vs Secondary Stats

After some testing, it doesn’t appear secondary stats affect the damage inflicted by Psychic Link. To clarify this, Psychic Link will always do 30% of whatever damage you dealt on the primary target, this includes the Crit/Mastery (Shadow Weaving) damage on the primary target, but the Link-ed target will not receive further damage increase from Mastery nor getting critically hit. The question here is the Crit stat. Previously, the Crit damage is applied individually when you cast Mind Sear on a group of mobs, however, in 10.1, due to this “all or nothing” nature (if you Crit on primary target, all Psychic Link-ed mobs also receive Crit damage), I can imagine Crit becoming a very important stat, possibly surpassing Mastery in AoE situation.

Top rows talents

Unfortunately, They are pretty much set and stone. You have the options to sacrifice survivability for shorter Silence cooldown/stun, or completely forgo Silence for better survivability.

Middle rows talents

Edited: Mental Decay has some uses due to its flat damage buff to Mind Flay or Mind Spike (the DoTs extension part is negligible). Dark Evangelism has some niche use in pure ST, otherwise it is close to useless.

Edited 2: Due to the changes in Idol of Yogg-Saron, as well as many talents that buff the damage Shadowy Apparition significantly, Void Eruption (Void Bolt) is much stronger than Dark Ascension. Void Bolt itself spawns Shadowy Apparition, in addition to that, more Void Bolt = more insanity = more Devouring Plague = more Shadowy Apparition = faster Idol of Yogg-Saron. Void Bolt is also currently the only way to extend your DoTs, which is extremely useful in m+.

Mind’s Eye vs Distortion Reality
The purpose of Distortion Reality is to prolong the Devouring Plague debuff thus increases the damage via Mastery: Shadow Weaving, Insidious Ire and Phantasmal Pathogen. Distortion Reality has some use in ST or Council-like fights where it allows you to maintain a much higher DP uptime. In AoE (such as in m+), you should be able to generate insanity much quicker thus maintaining the debuff via Mind’s Eye, and you also want the DP DoTs to expire sooner before the mobs are dead, prolonging DP is counter-productive. Distortion Reality also has counter synergy with our 4pc set bonus.

Mind Spike vs Mind Flay

At quick glance, besides being less appealing aesthetically, Mind Spike is a slightly more empowered filler spell. Mind Spike will likely sim with a higher number, however, if you consider the following scenarios, Mind Flay is actually a more practical spell and will most likely outperform Mind Spike in many scenarios:

  1. When channeling Mind Flay, it does 6 ticks of damage over 4.5 seconds (6 ticks over 3s with MF: Insanity). Within the same amount of time, you only get to cast 3 Mind Spikes. What this essentially means is, any spells/talents that can be proc-ed by Mind Flay or Mind Spike will have 2x/3x the chance if you cast Mind Flay instead (mainly Idol of C’Thun and Dark Evangelism). In addition to that, with our further reduced mobility, you will be forced to cancel a spell more frequently. You’d lose 1.5s of cast time on Mind Spike, whereas with Mind Flay, you probably would have done at least 1 or 2 ticks of damage (and also thus 1 or 2 more chances of proc-ing Idol of C’Thun)
    Clarification: “2x/3x the chance” is probably not the best description. Mind Flay has 2x or 3x the chance to proc the maximum number of Idol of C’Thun within the shortest amount of time. Assuming this hypothetical scenario: you have the Godliest RNG and every cast will proc an Idol of C’Thun, and the maximum proc is 4. With Mind Spike, the earliest possible for you to proc all 4 tendrils is 6s into the fight (4 x 1.5s), whereas with Mind Flay, it’d be 3s (4 ticks of 0.75s); and with Mind Flay: Insanity, it’d be 2s (4 ticks of 0.5s).
  2. Surge of Insanity - Mind Flay: Insanity does 259% spell power over 3 seconds; Mind Spike: Insanity does 187% spell power with 1.5s cast time. But let’s not forget - ONE Devouring Plague only gives you ONE Mind Flay: Insanity or ONE Mind Spike: Insanity. That means over the 3 seconds, TWO Mind Spikes really only do 187% (Insanity proc) + 66% (normal Mind Spike) = 253% spell power damage. (Yes, it’s “only” 6% less, but why would you want to waste 1 or 2 talent points to only do less damage?)
  3. If you opt out of Mind Spike and Mind Melt, you can free out 2 talent points and pick something that actually buff your overall damage (such as Ancient Madness and Mind Devourer).

Bottom rows talents

Interestingly, this is the first time we’re allowed to pick 3 capstones. Because of this, it is almost a no brainer to not do so. There are 2 possible exceptions:

  1. Deathspeaker vs Idol of N’Zoth. Idol of N’Zoth seems to be providing a more consistent damage and allows you to do (some) damage beyond 8 targets. Deathspeaker on the other hand provides more instant cast thus alleviates some mobility issues.

  2. Void Torrent + Malediction build. As opposed to the spell description, Idol of C’Thun is guaranteed to spawn when you channel Void Torrent. You can opt to drop Idol of N’Zoth and Screams of the void to get to Void Torrent + Malediction. On the other hand, if you still prefer to have 3 capstones, you can drop Insidious Ire.

Shadow definitely felt “dumbed down” (much easier to play). I do not know its dps performance in comparison to other dps specs. In strictly 8 targets AoE, it seems to be doing very decent damage.

11 Likes
2

Nice breakdown.

Gave me things to consider.

Good job.

3

Really really fantastic summary.

This has given me quite a bit of claity now, particularly in the scope of flay vs spike.

4

Thank you for the info/break down.

5

I havent really been keeping up but you didnt take mindbender/ysharaj stuff at all on st or aoe? That’s shocking based on how strong it was considered before.

6

As of right now, (Mindbender + Inescapable Torment + Idol of Y’Shaarj) build is very weak in both AoE and ST due to the following reasons:

  1. Mind Blast, SW: Death now only extend the Mindbender’s duration by 0.7s (as opposed to 1s in 10.0.7).
  2. Shadowy Insight proc no longer add one additional charge of Mind Blast, meaning any proc from Shadow Insight doesn’t increase the number of times Inescapable Torment is triggered (and thus further extending the duration)
  3. In order for Inescapable Torment to be useful, you also need to pick up Deathspeaker, but this will require you to sacrifice one of the capstones (a big NO)

In 10.0.7, without any Shadowy Insight proc, I managed to get 9 Inescapable Torment out for the entire duration of Mindbender (7 Mind Blast + 2 SW:Death). However, it is relatively easy to get as many as 3-4 procs, thus a total of 12-13 Inescapable Torment is pretty common.

In 10.1, without Deathspeaker, it is a fixed 7 casts of Inescapable Torment.

2 Likes
7

Oh wow good to know thanks for the detailed response.

8

Where did you get your build from? The Discord theorycrafters have all been saying VF is at least 2% stronger than DA in all content, so I’m 2 paragraphs in and your breakdown is already throwing up gigantic red flags for me.

Edit: your next paragraph, DR is actually 8% weaker on ST. Where are you getting this stuff from?

Edit again: There are no current builds where Mind Flay outperforms Mind Spike builds. This whole post is misinformation city.

2 Likes
9

You made a bunch of claims without providing any explanation, besides “the Discord theorycrafters” said so.

My analysis is based on repeated dummy testing. For each talent builds, I performed the rotations on the dummy over 8 minutes to get a more accurate average damage.

But of course, I have not done every single combinations possible out there thus it is entirely possible that I may have missed something.

It is also worth mentioning that I did NOT do any of the test using the new 4pc tier set bonus, thus regarding VR, I did mention it is possible the 4pc will offset the benefits you gained through VR.

I’m curious about the Mind Spike build though, perhaps there is a build that doesn’t involve Idol of C’Thun that I’m not aware of?

10

You’re right, I should have linked the PTR APL with all the builds tested hundreds of thousands of iterations. But sure, your training dummy testing is just as valuable. I mean wtf lmao.

I mean, what did you think those “Discord theorycrafters” are working with, cotton candy and butterflies? They’ve ran so many potential builds in so many situations that Shadow has already been solved for 10.1 barring new changes. Weeks ago.

If you want the “proof”, join the Priest discord. It’s chock full of proof, overflowing with proof.

1 Like
11

Yes, I’m in the discord and I’m well aware of their numbers. But even their numbers don’t provide much explanations because they are based purely on raw numbers, using computer to simulate (sim) the numbers.

But as we all know, sim don’t reflect reality of gameplay for many reasons - from bad RNG to casting the incorrect spells due to unexpected procs, or to simply dodging mechanics such as one of my examples above - if you have to cancel Mind Spike to dodge mechanics, you’d lose the entire 1.5s of cast time, whereas with Mind Flay, you’d have at least done 2 ticks of damage (that is 2 more chances of proc-ing Idol of C’Thun vs none)

They are also using the new 4pc tier set bonus which I did not (one of the reasons is because I’m trying to compare the baseline performance of shadow priest in 10.0.7 vs 10.1)

Edited: Okay, so there are also some changes that I was never aware of. For example, I just read the “Why You Should & Shouldn’t Play Shadow Priest in Dragonflight Season 2 - Patch 10.1” and learned that Void Torrent will guarantee a Idol of C’Thun spawn (the spell description only says “a chance to spawn”). Thus explain why Distortion Reality is weaker (because you will likely drop Insidious Ire for Void Torrent and Melediction, and Void Torrent generates more insanity)

Edited 2: I’m looking at one of their guides on GitHub. This statement here:

Note that if you want more single target just drop Distorted Reality for Mind’s Eye.

don’t make any sense at all. If anything, it should be the opposite, you want Distorted Reality for more ST and Mind’s Eye for more AoE. The reasons are very simple, in AoE (I’m assuming m+) even in a very high key, mobs usually don’t live more than a minute or so, so you actually want the dots to expire ASAP (to deal all of the accumulated VP damage), with Distortion Reality, chances are mobs are dead with still plenty of DoTs remaining, whereas in ST (likely in Raid), the boss will most likely survive the entire DoT duration so nothing goes wasted. This is what I meant by sim vs gameplay, in this case, sim is probably assuming all mobs will die at exactly 12s (interval).

1 Like
12

I think when accounting the rate of insanity generation and talent synergy should heavily factor in if you want a lower cost Devouring Plague or a longer duration of it.

With Insidious Ire, you would want DP up as often as your Mind Blast and Void Torrent are usable to maximize the damage enhancement synergy. The same can be said to other talents that factor in DP as a mechanism to increase your damage of other aspects of your toolkit.

With multiple targets, if they live long enough and you have enough insanity generation to use 2 longer duration DP before the 1st falls off, then that would be the justification for multiple targets. Where as if you had so much insanity generation that you risk overriding 2 longer duration DP on a single target situation where you loose damage, then that would justify a lower cost but lower duration DP. But again, that I think is only a secondary factor followed by talent synergy and rate of insanity generation.

13

Shadow being dumbed down is actually what has made me quit playing it. It had a complex playstyle like nothing other, and I really enjoyed that.
Idk what I’m gonna play now because nothing comes close to 10.0.5 shadow lol

2 Likes
14

That may put it over the top for me, but the issue is you cannot reliably count on this proc to move on demand - you’ll more than likely not have it when you need to.

Honestly, this raid tier was really rough for casters or ranged DPS that can’t do most of their DPS on the move and the fact Blizzard have basically doubled down on so many casters like Shadow being near 100% immobile has me really worried given how “movement happy” they were this raid tier and how you basically had to run football routes John Madden style for half the bosses.

BTW: As if it wasn’t enough, really appreciate the massive buff and design-revamp to Ret… making it one of the few uncomplicated, actual fun classes to play, Blizz. Now 20-25% of my raid rerolled Ret, I now have 4 additional characters rolling against me for the tier token next raid. Thanks for that… but hey at least the Catalyst is coming out way after the raid starts.

Looking forward to having a functional character again sometime this summer.

2 Likes
15

What are the opinions regarding Misery? :thinking:

No doubts that it is a strong talent, but doesn’t it basically makes SW:P meaningless for any kind of spriest build out there? Since it’s a mandatory talent, I’d question why it exists as one and not baked in VT to begin with.

Now, if it was all the other way around… SW:P appies VT, it would feel much more interesting because of the instant cast aspect of the former, and could potentially help a little bit with AoE outside of Shadow Crash.

16

Misery has always been a QOL talent since its inception.

It basically outshined every talent it competed with because of how well it integrated our 2 primary DOT spells.

It’s fine the way it is, no need to add or change anything.

Think of it this way, if Misery was in the 2nd or 3rd section of the talent tree, it would always be taken but it would feel bad because you have to make more sacrifices. But in 1st section of talents, it’s a no brainer and just a stepping stone to actual real choices later on that won’t conflict or detract picking Misery.

17

While your explanation of psychic link is accurate, your conclusion is not. Whether or not each individual damage event has an X% chance of critting, or if one parent event’s X% chance of critting affects all of its children, this does not change the chance of a given damage event critting. If crit is better, which it doesn’t appear to be, it would be because of changes to the mechanics of the spec and the rework itself, not psychic link.

For reference those playing searless on live this season haven’t touched crit at all.

This is also incorrect, because you’re ignoring the massive increases in power that idol of yogg saron has seen in recent 10.1 changes. The value of voidform is mostly in void bolt and its additional apparition triggers and the base damage of the spell itself. While you do need more insanity generation / gear to get uptime, the increased yogg triggers make up for it. Not to mention 3 devouring plagues in a voidform is really, really low.

The opposite is actually true, because distorted reality allows you to have comparable devouring plague uptime on 2-3 targets at once, instead of just one, increasing mastery and pathogen value in AoE. Where the talent is completely awful is Single Target because less devouring plague casts is less ghost and yogg triggers, and you aren’t typically struggling for mastery uptime on Single Target.

However you are correct, a 12 second dot as a spender just doesn’t work in keys at all. The talent is much more for council fights.

This isn’t the case because the procs which refer to our filler spells work on RPPM, not fixed percentage. Having more instances of damage can be beneficial to force RPPM events to trigger closer to their intended maximum number of times, but it doesn’t factor in to this situation because you still cast spike frequently enough to maximise your hedging.

But flay also takes up more globals, you’re ignoring the loss of damage / insanity / value from the gcd you get back from using mind spike which deals more compressed damage. Especially when during voidform your top priority is casting void bolt on cooldown.

Nzoth sucks in ST, a lot. Even if it’s a capstone that doesn’t make it better than the talents above it in the situations it wasn’t designed for.

1 Like
18

There is a lot I dont understand about what you’ve said, but wanted to chime in here specifically since you called out my post.

Distorted Reality means you cast less Devouring Plagues to get more uptime. On paper that by itself would be better, but in practice (and in sims which account for RNG and movement) this means you get less uptime on Screams of the Void, less procs of Surge of Insanity, less Shadowy Apparitions, and less uptime on the tierset.

All of those factors weigh heavily against Distorted Reality in single target.

I wanted to cover this bit real quick as this is also incorrect. The sims that we run do extensive tests with swings of RNG. We take real RPPM data from spelldata or from empirically testing in-game and matching proc rates to get an accurate assumption of how procs smooth out over the course of many thousands of iterations. This gives us a swing in dps and proc variance we can use to pinpoint where the average is for most people.

Furthermore ALL of our sims have a degree of movement. None of our simulations rely purely standing still the entire time, as that is highly unrealistic. We take into account movement or forced mechanics that would cause you to cancel casts of any spell. All of this is done to get close to a real world environment.

1 Like
19

Having more instances of damage can be beneficial to force RPPM events to trigger closer to their intended maximum number of times, but it doesn’t factor in to this situation because you still cast spike frequently enough to maximise your hedging.

You’re correct about the RPPM. But that’s not what I was trying to say. In a shorter fight (such as trash packs in m+), you won’t cast enough Mind Spike to “trigger the intended maximum number of times”. Think of this hypothetical scenario: you have the Godliest RNG and every cast will proc an Idol of C’Thun, and let’s assume the maximum proc is 4. With Mind Spike, the earliest possible for you to proc all 4 tendrils is 6s into the fight (4 x 1.5s), whereas with Mind Flay, it’d be 3s (4 ticks of 0.75s); and with Mind Flay: Insanity, it’d be 2s (4 ticks of 0.5s). In raid, this is less impactful because the boss fight usually last for many minutes. But in m+, 6 seconds is a lot! You’d lose out a bunch of dps if the mobs die before your tendril can finish channeling its full duration.
I guess the better way to say this is Mind Flay has a higher chance to “trigger the intended maximum number of times” in the shortest amount of time.

But flay also takes up more globals, you’re ignoring the loss of damage / insanity / value from the gcd you get back from using mind spike which deals more compressed damage. Especially when during voidform your top priority is casting void bolt on cooldown.

During VE + PI (and possibly lust), Mike Spike performs slightly better due to the reasons you mentioned above. Outside of CD, Mind Flay build still beat Mind Spike by a considerable amount due to all 3 reasons I mentioned in my original post.

I have tested the rotation repeatedly on the 6 dummies, and this is what I found (with my current gears):
With Mind Spike build, the dps spiked to about 275k during the first VE + PI, then averaged to 205k dps over the course of 4 minutes (I popped VE + PI again on minute 2 but the dps struggled to get any higher).
With Mind Flay build, the dps spiked to about 265k during the first VE + PI, but then averaged to 220k dps over the course of 4 minutes.

Spell clipping is a very real thing, especially now we have even reduced mobility. I have tried this in a few 23 and 24 keys where I had to dodge some mechanics and I’d frequently lose out 1.5s of cast time on Mind Spike, whereas with Mind Flay, I was able to do at least 1 or 2 ticks of damage, and kid you not, every so often, that 1 or 2 ticks actually helped spawning a tendril.

This is also incorrect, because you’re ignoring the massive increases in power that idol of yogg saron has seen in recent 10.1 changes. The value of voidform is mostly in void bolt and its additional apparition triggers and the base damage of the spell itself. While you do need more insanity generation / gear to get uptime, the increased yogg triggers make up for it. Not to mention 3 devouring plagues in a voidform is really, really low.

This I agree with you. After running a few keys, I do notice VE is way stronger than DA. I wasn’t aware of the Yogg-Saron change, but also mostly because Void Bolt can extend the DoTs which works very well on bigger mobs pack.

@Publik

There is a lot I dont understand about what you’ve said, but wanted to chime in here specifically since you called out my post.

Distorted Reality means you cast less Devouring Plagues to get more uptime. On paper that by itself would be better, but in practice (and in sims which account for RNG and movement) this means you get less uptime on Screams of the Void, less procs of Surge of Insanity, less Shadowy Apparitions, and less uptime on the tierset.

All of those factors weigh heavily against Distorted Reality in single target.

Basically, what I was trying to say is, you want the DoT to last its full duration on the mobs (or as long as possible). Because of this, the longer the DoT stretches out, the more DPS you will potentially be wasting.

For others who don’t quite understand what I’m saying, think of the math below (not based on real value):

The mob only has 1500 remaining health. Regular DP deals 1000 initial damage + 3 ticks of 400 damage. Distorted Reality DP deals 1200 initial damage and 6 ticks of 240 damage.

So, by casting one last Regular DP, the mob would die after the 2nd tick, with 700 DoT goes to waste.
With Distorted Reality: The mob would also die after 2nd tick, but with 1140 DoT goes to waste.

Now, this is just a single cast on a single mob. Imagine overlapping DP on a group of Psychic Link-ed mobs (which is usually the case in m+), the DPS loss can easily grow exponentially.

Any DoT that goes to waste is the DoT that you cannot apply on another mob, thus DPS loss.

I wanted to cover this bit real quick as this is also incorrect. The sims that we run do extensive tests with swings of RNG. We take real RPPM data from spelldata or from empirically testing in-game and matching proc rates to get an accurate assumption of how procs smooth out over the course of many thousands of iterations. This gives us a swing in dps and proc variance we can use to pinpoint where the average is for most people.

Furthermore ALL of our sims have a degree of movement. None of our simulations rely purely standing still the entire time, as that is highly unrealistic. We take into account movement or forced mechanics that would cause you to cancel casts of any spell. All of this is done to get close to a real world environment.

I honestly do not know how sims work behind the scene or what its algorithm is.

But since you’re responding to my post, I’m really curious to know if sim is assuming the mob would survive the entire duration of the DoT, or take into account mobs dying half way into the DoT’s duration.

When you mentioned your sim take into account movement or forced cancellation of spell casting. How often are those movement or forced cancellation? Given our reduced mobility, this should be quite frequent. In addition, cancelling a spell casting versus cancelling a spell channeling are 2 different things, does the sim takes into consideration things like “0 cast of Mind Spike versus 2 ticks of Mind Flay (thus 2 additional chances to proc C’Thun”?

I’m also curious to hear your thought on one of my posts above:

Think of this hypothetical scenario: you have the Godliest RNG and every cast will proc an Idol of C’Thun, and let’s assume the maximum proc is 4. With Mind Spike, the earliest possible for you to proc all 4 tendrils is 6s into the fight (4 x 1.5s) , whereas with Mind Flay, it’d be 3s (4 ticks of 0.75s); and with Mind Flay: Insanity, it’d be 2s (4 ticks of 0.5s). In raid, this is less impactful because the boss fight usually last for many minutes. But in m+, 6 seconds is a lot! You’d lose out a bunch of dps if the mobs die before your tendril can finish channeling its full duration.

Does sim take into account “late” tendril spawn and thus mobs dying off before the tendril finished channeling its full duration?

1 Like
20

A perfect breakdown— I’m a “monkey see monkey do” kind of player, so seeing an in depth explanation of our options this season is fantastic.