Archived

This topic is now archived and is closed to further replies.

Mearis

Shadowpriest Theorycraft 3.3 Edition - I get by with a little help from my friends

254 posts in this topic

Taking points out of Focused Mind or Vampiric Embrace is silly. You have plenty of spare points if you don't go with the Spirit Tap line (which has been proven to be awful at best).

It's better than losing a few spellpower.

Share this post


Link to post
Share on other sites
It's better than losing a few spellpower.

Imp Vampiric Embrace is better than spending 5 talent points for a tiny dps gain per point.

Share this post


Link to post
Share on other sites
Imp Vampiric Embrace is better than spending 5 talent points for ~3-4 dps per point.

And what about Focused Mind? It's a fairly small MP5 boost, and is even less useful now that we have an open glyph slot for Glyph of Dispersion.

Share this post


Link to post
Share on other sites

Did I say they're taking points out of Imp VE? No, I said they're turning 3/3 focused mind into 0/3 focused mind, 2/2 Imp VE, and 1/1 Inner Focus. Read my post before you make a big deal about something, please. 2/2 VE is amazing now, but it was sub-par on Anub when 1 VE + JoL was enough to keep a ranged group up and any overheal went directly into Anub's health pool.

I don't know how you can give up ST/Imp ST anymore with the spirit changes. Considering how many fights in ICC have adds and spirit's new buffs, proccing a spirit tap is a jump in several hundred spellpower and lots of delicious manareg. I find it to be far more useful than focused mind (which is what, 50 MP5? maybe if it affected DoTs it'd be useful) in any practical application.

Share this post


Link to post
Share on other sites

Hello not sure if this is the right place for this but...

Since the patch i believe Haste and SP are now a level field in the way they benefit shadow priests. I have been stacking haste and currently have about 28% unbuffed with around 18% crit unbuffed and 2780 SP (with inner fire) my question is. Is that to much haste should i be dropping some getting more crit / spell power, by changing some gear around in my bags i can get more haste but at the cost of crit. what is the ideal haste rating any comments would be great.

Link to my armory

The World of Warcraft Armory

feel free to comment on enchants \ gems if you think they should be different

Share this post


Link to post
Share on other sites

The issue I've always had with Imp VE is that most of the time when your healers are doing a good/really good job Imp VE is just a bunch of over-heals.

Here's my spec, The World of Warcraft Armory

I haven't gotten to play in a while and I plan on changing it to -3 shadow affinity, -1 focused mind, +4 Imp MB.

My spec is a good mix of mana regen, cost reduction, and gets those extra DPS points where it can.

Share this post


Link to post
Share on other sites
Hello not sure if this is the right place for this but...

Since the patch i believe Haste and SP are now a level field in the way they benefit shadow priests. I have been stacking haste and currently have about 28% unbuffed with around 18% crit unbuffed and 2780 SP (with inner fire) my question is. Is that to much haste should i be dropping some getting more crit / spell power, by changing some gear around in my bags i can get more haste but at the cost of crit. what is the ideal haste rating any comments would be great.

Link to my armory

The World of Warcraft Armory

feel free to comment on enchants \ gems if you think they should be different

Simcraft and Rawr are your friends. Use them; it's not our job to optimize your character for you.

The issue I've always had with Imp VE is that most of the time when your healers are doing a good/really good job Imp VE is just a bunch of over-heals.

Here's my spec, The World of Warcraft Armory

I haven't gotten to play in a while and I plan on changing it to -3 shadow affinity, -1 focused mind, +4 Imp MB.

My spec is a good mix of mana regen, cost reduction, and gets those extra DPS points where it can.

Imp mind blast is mandatory 5/5 to max DPS, and focused mind and shadow affinity are trash talents. Threat should never be an issue, and focused mind gives negligible cost reduction. Imp VE is good because when it counts, it saves lives. It should not be neglected in any serious raiding shadow priest's arsenal. Don't worry about mana regen as a shadow priest because we have Shadowfiend and Dispersion. Further, you should replace your glyph of mind sear as it's nigh-useless in most encounters. The common choice is Glyph of Dispersion as it gives superior regen to Glyph of SWP and reduces the cooldown on your Oh Shit! button.

Share this post


Link to post
Share on other sites
The common choice is Glyph of Dispersion as it gives superior regen to Glyph of SWP and reduces the cooldown on your Oh Shit! button.

question - does anyone really need to hit dispersion more than once every 2 min as an oh shit button? if so i think they must be bad. And if we actually use it ever 1 min 15sec as mana regen it might give us more mana than glyph of swp but isnt taking 6 sec out of dps every min a dps loss? do we really need that much mana? i use shadow fiend twice every fight (once with heroism/bloodlust) and if i ever have a few extra sec or get below 50% mana i use symbol of hope + dispersion to top me off and give some random raid members a min-innervate... is this all just preference or is there something significantly better about glyph of dispersion over glyph of swp?

Share this post


Link to post
Share on other sites
Did I say they're taking points out of Imp VE? No, I said they're turning 3/3 focused mind into 0/3 focused mind, 2/2 Imp VE, and 1/1 Inner Focus. Read my post before you make a big deal about something, please. 2/2 VE is amazing now, but it was sub-par on Anub when 1 VE + JoL was enough to keep a ranged group up and any overheal went directly into Anub's health pool.

My post wasn't intended to be confrontational, but you seem to have taken it that way. Please, go back and read what you said:

Many Spriests are removing points from Focused Mind (used in H Anub to make mind searing less mana-hungry) and putting 2 back in Imp VE, and another in Inner Focus.

Your post implies they didn't have Imp VE in the first place, so having Spirit Tap would be via not having points in Imp VE. My argument was picking spirit tap line over imp ve line. I could have been more clear, but despite H Anub there is enough benefit to Imp VE that leaving it out for a very marginal dps gain was silly. It's obviously a stronger talent now than it was, but it was never anything to scoff at to begin with.

question - does anyone really need to hit dispersion more than once every 2 min as an oh shit button?

There are plenty of times where large predictable damage is coming, or where you may have a period where you can dispersion for a few ticks. Icehowl's Stomp, Jumungar burrow, Faction Champions (both for snares and survival), Twins Vortex, end of Anub p2, spikes on Marrowgar, etc. We don't even have hard modes yet, so you can imagine damage is only going to ramp up.

Share this post


Link to post
Share on other sites

Dispersion

Again, its one of those things you need until you need it. In later ICC fights, and months from now when HMs come out, there will be a greater need for the 'oh shit' button. Especially as your guild learns the encounters.

Lets look at it from another way: do you really need any third glyph? No, MF + Shadow are the only two that you really should have. So if you had to choose between mana you don't need (SW:P), DPS loss (SW: D), range you don't need (MS), or a shorter CD 'oh shit' button (Disp) which would you pick? So what if I don't use the shorter CD every time, it's still has the opportunity for greatness.

Share this post


Link to post
Share on other sites
Your post implies they didn't have Imp VE in the first place, so having Spirit Tap would be via not having points in Imp VE. My argument was picking spirit tap line over imp ve line. I could have been more clear, but despite H Anub there is enough benefit to Imp VE that leaving it out for a very marginal dps gain was silly. It's obviously a stronger talent now than it was, but it was never anything to scoff at to begin with.
I dealt with this in my post as well; many shadow priests were eschewing Imp VE because it led to unnecessary overheals on Anub--this freed up talent points for Focused Mind which made mana management far easier. I specced into Spirit Tap and Imp Spirit Tap because a marginal DPS gain is better than none at all, and Spirit Tap was a semi-potent dps and manareg buff as well, when you could proc it.

As far as glyph of Dispersion comes, I just have to say that a good player probably won't need the extra "Oh shit" button, but a great player will account for all possibilities. You're only gimping yourself if you don't take it. There are also plenty of opportunities to Dispersion in most fights without taking a DPS loss.

Share this post


Link to post
Share on other sites
AoE

More than 3 mobs:

Vampiric Touch freely and then Mind Sear, rotating the target of Mind Sear.

Vampiric Touch is almost always worth casting over Mind Sear, unless you are AoE'ing a lot of very low hit point adds.

Would it be possible to elaborate this slightly, or point to the relevant math? In particular, how many (long-living) targets does it take to make pure Mind Sear spam superior. I would assume this greatly depends on whether the 2T9 bonus is available or not.

Share this post


Link to post
Share on other sites
I don't know how you can give up ST/Imp ST anymore with the spirit changes. Considering how many fights in ICC have adds and spirit's new buffs, proccing a spirit tap is a jump in several hundred spellpower and lots of delicious manareg. I find it to be far more useful than focused mind (which is what, 50 MP5? maybe if it affected DoTs it'd be useful) in any practical application.

I know why I can give up Spirit Tap. I don't have any spirit on my gear. The full set of BiS gear from Icecrown involves a whopping one item with spirit (a mainhand weapon). Even if you have some piecemeal spirit gear due to what drops your raid actually got, you're not going to have more than 600 spirit. A permanent 10% increase in spirit is +60 spirit, which is worth around 33 spell power. You're gaining 6.6 spell power per point spent on Spirit Tap. I agree it's an improvement, but not a terribly compelling one. I'm pretty sure I'm sticking with Silence, as having a ranged interrupt seems more likely to prevent a boss fight wipe than an extra 33 spell power.

The issue I've always had with Imp VE is that most of the time when your healers are doing a good/really good job Imp VE is just a bunch of over-heals.

Similarly you could say that you only need to bring 4 healers to fight X, because you only need four healers if no one stands in the fire. Or that you shouldn't enchant runspeed because you should always magically be standing in the right place ahead of time.

The reality is that we make these concessions-- more healers, enchanting runspeed, and speccing improved VE-- so that the raid as a whole has a wider margin of error, making you less likely to wipe. I'd rather have a raid that survives if at least one of two people does their job. Besides, how else are you going to spend those 2 points? Getting 13.2 spell power from Spirit Tap?

Share this post


Link to post
Share on other sites
Would it be possible to elaborate this slightly, or point to the relevant math? In particular, how many (long-living) targets does it take to make pure Mind Sear spam superior. I would assume this greatly depends on whether the 2T9 bonus is available or not.

I'm using numbers from my log.

Mind Sear

# of hits = 3945

avg hit = 2252.7

total dmg from hits = 8886924

# of crits = 2249

avg crit 3482.7

total dmg from crits = 7832530

Avg dmg per mob hit = (8886924 + 7832530) / (3945 + 2249) = 2699.3

Vampiric Touch

# of hits = 2230

avg hit = 2572.1

total dmg from hits = 5735712

# of crits = 1320

avg crit = 5303.5

total dmg from crits = 7000585

Avg dmg per cast = ((7000585 + 5735712) / (2230 + 1320)) * 7 = 25113.8

Assuming 0 haste, since it doesnt matter unless you're gcd capped, VT has 16742.6 DPET and MS has 2699.3 DPET.

So you need 16742.6 / 2699.3 = 6.2 mobs being hit to spam MS.

In a scenario with 8+ mobs MS spam is better than VT(assuming the mobs wont die shortly after you cast VT, if the mobs are gonna die before VT deals full dmg it might be better to just spam MS).

The data is from Anub 25 H fights. Results may vary slightly depending on raid buffs, mobs debuffs and gear.

Share this post


Link to post
Share on other sites
Assuming 0 haste, since it doesnt matter unless you're gcd capped, VT has 16742.6 DPET and MS has 2699.3 DPET.

So you need 16742.6 / 2699.3 = 6.2 mobs being hit to spam MS.

Thanks for the data.

For completion, you assume here that the T9 bonus.

If you don't have it, multiply VT damage by 5/7, and similar computations yields to 4,42 mobs being hit to spam MS.

Fore more complete results, if you estimate that your VT will have only N ticks before the mobs die or VT expires, you need 6.2 * N / 7 mobs per hit to spam MS.

Share this post


Link to post
Share on other sites
* Recast Shadow Word: Pain at 5x Shadow Weaving.

Is this suggesting that we no longer wait for other damage modifier and/or crit modifer (de)buffs (such as imp scorch/shadowbolt)? I was aware that rolling hasted corruptions was fixed, but I didn't see or hear one way or the other if this affected locking crit/dmg modifiers with SWP.

Opening Sequence:

Depending on if you have to run in or not, things vary slightly, however, the typical opening sequence is something like this:

Vampiric Touch

Devouring Plague

3 Tick Mindflay

Then SW:P

You might offer a note about using Shadowfiend as a DPS cooldown in the opening rotation. The chances of going oom within the first 3mins of the fight is pretty unlikely and you always have dispersion as a backup.

My opening rotation is usually something like this: VT DP MB Shadowfiend MF SWP. I'll also toss in a SWD (or another MF) before pain if I'm not satisfied with the current modifers (provided by shadowgreenlight) in hopes that the extra GCD will provide enough time for them to be applied.

Share this post


Link to post
Share on other sites

Regarding the VT vs. Mind Sear spam discussion, an important fact that is being left out is that you can easily put VT on all the mobs before they are grouped up, so you aren't really losing any dps, since you wouldn't be Mind Searing anyway. A classic example of this is Anub, and just about any add fight where Mind Sear is viable would present a similar opportunity. If aggro is an issue (which it shouldn't be) you can always fade while you DoT everything up, which should give the tanks sufficient time to gain aggro over your one or two VT ticks.

Share this post


Link to post
Share on other sites

Some more thoughts on the "closed-form" approximation:

Now that haste has effect on dots, I tried the following approximation :

  1. Run the cycle (with priority VT > DP > MB > MF), and compute its dps. Restrict to 3-ticks MF. That's easy to do, because every spell has cast time and cooldown that are multiplicative of the GCD. You can get a closed form solution pretty easily (it's just long and boring to do).
  2. Add now the haste, by multiplying the previous "cycle dps". Still a close form solution.
  3. Add the dps from other spells, mainly Shadow Word:Pain and ShadowFiend. You're still in closed-form area.

I know they are several approximation / difficulties :

  1. Cycle is changed by haste, because the cd of MB is not decreased. I guess that the effect is not so huge in fact (but that's a guess).
  2. Several procs / temporary buffs are difficult to model. Pretty easy solution is to average the buffs / effects, leading to a "correct" approximation.
  3. I restricted the cycle, using no SW:Death neither 2-ticks MF, because they simplify the cycle. Actual dps can only be better.

Before doing the actual closed form, I've been lazy and wanted to check that the approximation is not too far from reality, and made a quick programm to predict some results in a simple case. Can anyone confirm me that the model dps prediction is somewhat close to simcraft/ Rawr prediction in the following case ? I don't have simcraft installed, and Rawr would be tricky to do it.

Case modelled:

  1. Spellpower is a flat 3500.
  2. Only glyph is the glyph of MF (for 10% more dps on MF). Glyph of shadow is "included" in the 3500 spellpower).
  3. Spec is a classical one, without Improved ST. No inner focus (I hate procs, did I already told you ? ).
  4. Crit is 35%, with all talents / buffs / mod debuff, for all spells (I know that there is some little difference for spells, I've been lazy).
  5. Haste is 35%, everything taken into account also.
  6. Shadowfiend is a flat 290 dps (value comes from Althor's simulation in 3.2 thread).
  7. No proc / on-use trinket. (Eye of the BroodMother, Illustration of the Dragon Soul are so nice, and guess what : they don't proc).
  8. Don't ask my why. but I included the Earth&Moon effect, and no other raid buff (but the only one is 3% more dps).

Here are the results :

Gear + Buff : 3500 spellpower, 0.350000 crit, 0.350000 haste

Earth&Moon(or EP or CoE) is the only raid buff that is included.

Priority : VT > DP > MB > MF. Full MF only.

VT - DP - MB - MF - MF - MB - MF - VT - MF - MB - MF - MF - DP - MB - VT - MF - MF - MB - MF - MF - VT - MB - MF - DP - MF - MB - MF - VT - MF - MB - MF - MF - MB - MF -

----------------------------

Damage per VT : 16593

Damage per DP : 15105

Damage per MB : 6382

Damage per MF : 9742

-----------------------

51 GCD cycle : 5 VT, 3 DP, 9 MB, 17 MF

Total damage : 351355, cycle dps : 4592

-----------------------

Hasted cycle dps : 6200

SWP dps : 587

Shadowfiend dps: 290

---------------------

Total DPS : 7078

Is it realistic ?

I can easily put other values for spellpower / haste / crit.

I can also release my code if anyone is interested.

Share this post


Link to post
Share on other sites

I put up a character with:

3459 Spell Power

31.37% Crit (Thats 35.37% on MB, 37.37% on DoTs)

36.57% Haste

291 Hit Rating

Damage pr VT: 15753 * 1.13 ~= 17800

Damage pr DP: 15615 (+3995 Imp. DP) ~= 22159

Damage pr MB: 5779 * 1.13 ~= 6530

Damage pr MF: 8841 * 1.13 ~= 9990

SWP DPS: 643.69 * 1.13 ~= 727.37 DPS

Rawr suggests a rotation of:

VT - DP - MB - MF - MF - MB - VT - MF - MF - MB - DP - MF - VT - MB - MF - MF - MB - MF - repeat

7085 DPS before Shadowfiend.

BiS, 4pt10:

(Trinkets are Illustration + Eye of the Broodmother due to newest trinkets not fully implemented)

3877 Spell Power

47.13% Crit

37.64% Haste

269 Hit Rating

The closed form solution is 30 spells before repeating.

9749 DPS before Shadowfiend.

Quick estimation of DPS with Muradin's Spyglass and Dislodged Foreign Object puts us at around 10k DPS before Shadowfiend.

Share this post


Link to post
Share on other sites

Elambras, I have worked on a closed form solution for modelling shadow DPS in the past, and I find myself continually frustrated by the number of exceptions and problems I find in the model. If you have the patience to keep developing something like that, more power to you, but I'm content to use simcraft. If you have excess coding energy, you should contribute on that project.

Your assumption 2 bothers me. The cooldown on Mind Blast does not change with haste, so that couldn't be in the haste multiplied equation you're developing. Certain haste levels would change Mind Blast's position in the cycle.

I think I like [iTEM]Black Magic[/iTEM] over 63 SP. It simcrafts for me as a 50 DPS upgrade over 63 spellpower, which is interesting because it is only usually listed as a ~10PP difference. It doesn't really seem to affect my Flay cycles.

Share this post


Link to post
Share on other sites

My main reason (and what I can only assume to be the primary reason of most) for not using Black Magic over 63sp is because its theorycrafted value assumes that you are allowed to be entirely stationary, which is obviously not always the case. At least as far as the fights that I've seen so far in ICC (both on live and the ones I experienced on ptr), I don't think that a stationary state is necessarily reliable enough to justify using Black Magic. Perhaps if it had a bigger advantage over 63sp than it does, but even at a 50 DPS upgrade I really don't think I could justify switching.

Share this post


Link to post
Share on other sites

Thanks a lot, The Not So Evil.

I'm not sure I'll be motivated enough to do it, but an approximate closed-form solution seems to be feasible.

In fact, I was more interested in the "theoretical" part than in the result. Now that I know it's feasible, and how to do it, the interest greatly decreases. And I do agree that Rawr and Simcraft are better tools. The only interest is that they don't have any closed-form formula, which sometimes gives intuition.

@jdgaynor : I know about Mind Blast. One way to "correct it" would be to consider that the number of Mind Blast stays the same, and replace the additionnal Mind Blast in the cycle by Mind Flay dps. But that wouldn't be correct neither, because in fact, the whole cycle change. That was the point of the approximation : to not recompute the cycle.

However, MF dps is about 75% of MB dps (if my numbers are right).

So if I'm taking the previous approach, I shall keep the cycle dps, and just replace MB with "1/2 MF" for haste concerns. MB is 9 out of 51 GCD in the cycle, and about 16% of the cycle damage. Replacing it with MF leads to only 12% damage instead, multiplied by haste. So the "error" is around 0.12*haste, in percentage. That's between 4% and 6%, still "fine" for me.

Share this post


Link to post
Share on other sites

Rawr tries to build a closed-form solution. Before 3.3 this was often not so successful giving some unfortunate haste results. With 3.3 and hasted DoTs however, its accuracy has increased a lot. In the event it cannot create a closed-form solution it will use a time limited solution instead. (Which is basically the amount of spells you can cast inside the selected fight length). A lot of work has been put into Rawr between 3.1 and 3.3 to reduce/eliminate the impact of rotation changes due to varying haste.

All in all I feel Rawr is pretty accurate when it comes to the estimation of a players expected DPS when left at default settings.

And for those of you who have been living under a rock lately. Rawr estimates 4pt10 to be about 462.17 DPS in BiS gear. (238.23 PP for those of you who prefer that.). Get it. Really.

Share this post


Link to post
Share on other sites

If these closed-form solution are not too varying one from another, it could be helpful to give a few meaningfull instances (for example, a different tier level gear). In fact, the exact expression may vary, but I expect that the difference will not be huge for similar gears.

Share this post


Link to post
Share on other sites

Biggest single problem with closed form solutions is that each haste proc requires a new solution. Taking the worst example I can think up, its a Troll Shadowpriest with 2 use/proc Haste Rating effect trinket, Engineering Glove enhancement, Black Magic weapon enchant, Haste Potion and Heroism. Depending on how the Troll stacks and/or avoids stacking this haste you get different solutions. (And quite many of them actually.)

Share this post


Link to post
Share on other sites
Guest
This topic is now closed to further replies.

  • Recently Browsing   0 members

    No registered users viewing this page.