*These are the results from using the latest trunk build of SimCraft (build 3251) which is updated for the live version of 3.2.0 and PTR build 10433 of 3.2.2.
Do not use for earlier versions.*
These scaling coefficients are for Shadow Priests with BiS T7, BiS T8, no-TotGC 25man BiS T9 and full BiS T9



TL;DR version:
------------------
With 2*Reign of the Unliving:
3.2.0 T9 BiS PP scale factors: Int=0.2525 Spi=0.2085 SP=1.0000 Hit=1.8757 Crit=0.8428 Haste=0.7225
3.2.2 T9 BiS PP scale factors: Int=0.2585 Spi=0.3367 SP=1.0000 Hit=1.9065 Crit=0.8541 Haste=0.7368

With 258 Reign of the Unliving + Flare of the Heavens:
3.2.0 T9 BiS PP scale factors: Int=0.2387 Spi=0.2151 SP=1.0000 Hit=1.7919 Crit=0.7890 Haste=0.7035
3.2.2 T9 BiS PP scale factors: Int=0.2441 Spi=0.3395 SP=1.0000 Hit=1.8217 Crit=0.8000 Haste=0.7185



The Non-Attention Span of a Gnat version:
---------------------------------------------------


The tests were run with 40000 iterations.

optimal_raid=1 was set in SimCraft (so all the raid buffs, except for any +spellpower that a deep Demo Lock brings above what ToW brings are covered).

The SimulationCraft gear sets tested were:
============================

1) BiS T7 gear set: Priest_T7_13_00_58.simcraft
No gear introduced in 3.1 is allowed in this set.
2) BiS T8 gear set: Priest_T8_13_00_58.simcraft
No gear introduced in 3.2 is allowed in this set. All 3.1 gear and earlier is allowed including Ulduar 25man Hard Mode loot.
3) BiS T9 245 gear set: Priest_T9_13_00_58_245.simcraft
This gear set doesn't allow any loot from the 25man Heroic version of Coliseum, season 7 Arena gear or any loot from 25man Hard Mode Yogg-Saron, Algalon, Mimiron or Freya.
4) BiS T9 258 gear set: Priest_T9_13_00_58_258.simcraft
All 3.2 gear allowed in this set.
5) BiS T9 258 gear set but with Flare of the Heaves instead of the ilvl 245 Reign of the Unliving: Priest_T9_13_00_58_258_foth.simcraft

The Results (3.2.0)
===============

In terms of DPS gain per stat point:
-------------------------------------------


In terms of PP (i.e. normalized against Spell Power) (3.2.0):
----------------------------------------------------------------------

Code:

  Priest_T7_13_00_58          Int=0.2151  Spi=0.2115  SP=1.0000  Hit=1.4637  Crit=0.7102  Haste=0.6479
  Priest_T8_13_00_58          Int=0.2186  Spi=0.2197  SP=1.0000  Hit=1.5066  Crit=0.7054  Haste=0.6237
  Priest_T9_13_00_58_245      Int=0.2302  Spi=0.2142  SP=1.0000  Hit=1.5464  Crit=0.7493  Haste=0.6311
  Priest_T9_13_00_58_258      Int=0.2525  Spi=0.2085  SP=1.0000  Hit=1.8757  Crit=0.8428  Haste=0.7225
  Priest_T9_13_00_58_258_foth Int=0.2387  Spi=0.2151  SP=1.0000  Hit=1.7919  Crit=0.7890  Haste=0.7035

Lootrank urls (3.2.0):
==============

Priest_T7_13_00_58
Priest_T8_13_00_58
Priest_T9_13_00_58_245
Priest_T9_13_00_58_258
Priest_T9_13_00_58_258_foth

Wowhead urls (3.2.0):
==============

Priest_T7_13_00_58
Priest_T8_13_00_58
Priest_T9_13_00_58_245
Priest_T9_13_00_58_258
Priest_T9_13_00_58_258_foth

Pawn strings (3.2.0):
==============

Priest_T7_13_00_58:


Priest_T8_13_00_58:


Priest_T9_13_00_58_245:


Priest_T9_13_00_58_258:


Priest_T9_13_00_58_258_foth:


The Results (3.2.2):
===============

In terms of DPS gain per stat point:
-------------------------------------------


In terms of PP (i.e. normalized against Spell Power) (3.2.2):
-----------------------------------------------------------------------

Code:

  Priest_T7_13_00_58          Int=0.2185  Spi=0.3376  SP=1.0000  Hit=1.5019  Crit=0.7238  Haste=0.6645
  Priest_T8_13_00_58          Int=0.2177  Spi=0.3370  SP=1.0000  Hit=1.5555  Crit=0.7200  Haste=0.6332
  Priest_T9_13_00_58_245      Int=0.2273  Spi=0.3437  SP=1.0000  Hit=1.5774  Crit=0.7578  Haste=0.6458
  Priest_T9_13_00_58_258      Int=0.2585  Spi=0.3367  SP=1.0000  Hit=1.9065  Crit=0.8541  Haste=0.7368
  Priest_T9_13_00_58_258_foth Int=0.2441  Spi=0.3395  SP=1.0000  Hit=1.8217  Crit=0.8000  Haste=0.7185

Lootrank urls (3.2.2):
==============

Priest_T7_13_00_58
Priest_T8_13_00_58
Priest_T9_13_00_58_245
Priest_T9_13_00_58_258
Priest_T9_13_00_58_258_foth

Wowhead urls (3.2.2):
==============

Priest_T7_13_00_58
Priest_T8_13_00_58
Priest_T9_13_00_58_245
Priest_T9_13_00_58_258
Priest_T9_13_00_58_258_foth

Pawn strings (3.2.2):
==============

Priest_T7_13_00_58:


Priest_T8_13_00_58:


Priest_T9_13_00_58_245:


Priest_T9_13_00_58_258:


Priest_T9_13_00_58_258_foth:


Analysis of the results:
----------------------------

1) The spirit PP value remains pretty much steady as you'd expect because there's a direct correlation between spirit and spell power via talents. The only other factor is how often Improved Spirit Tap procs and how often Glyph of Shadow procs.

2) The value for Haste is at it's lowest point with the T8 set. This is probably due to this set having the 4pc T8 bonus which means that on average it's haste value is higher than the other sets which in turn tends to diminish the value of the stat a bit.

3) In the T9 sets the relative value of Intellect and Crit see a bit jump. This is in large part due to the 245 set having 1 Reign of the Unliving and the 258 set having 2 Reign of the Unlivings. These items gain nothing from spell power or most of our talents, but do gain from having more crit% and both Intellect and Crit Rating boost our crit%.

4) Haste on the T9 258 set also sees a boost over the T9 245 set. This would probably be a combination of having 2 Reign of the Unliving's (the more haste the closer you can get to the minimum time between procs) as well as having the +5% crit on Mind Flay which means Mind Flay has a larger share of our DPS and losing the 2pc T8 bonus so Devouring Plague has a lower share. Mind Flay gains more from Haste than Devouring Plague and so the increase is expected.

5) Hit of course increases relative to the overall DPS. The more DPS a set does the more Hit is worth compared to Spell Power.

The gear sets in detail (using the SimulationCraft format).
----------------------------
Notes:
1) The alliance version of gear was used. The horde gear has the same stats, just a different name.
2) The Night Elf race was used because it's one that has no bonuses that directly impact Shadow Priests and thus is somewhat "neutral".


Priest_T7_13_00_58.simcraft:


Priest_T8_13_00_58.simcraft:


Priest_T9_13_00_58_245.simcraft:


Priest_T9_13_00_58_258.simcraft:


Priest_T9_13_00_58_258_foth.simcraft:

Thanks for the analysis. Interestingly it has made almost no impact on the top 10 items in each slot (one or two minor shifts) although (if my calcs are correct) it has lowered the 'requirement' on the CSD to be better than the ESD from around 4350 to 4175dps. And makes JC even more tempting a switch

Cheers
K

How much more work would be required to do this process to find stat values for Holy/Disc? No other site seems to compare to the timely and active theorycraft seen on these forums so it's kinda amazingly hard to find proper info like this else where.

Do you mean a Holy/Disc DPS spec? Or for healing?

The numbers I posted are from SimCraft runs. It only does DPS currently. Nothing on the immediate horizon to handle healing.

As there really are no healing "rotations" as there are for dps it's kind of hard to come up with anything decent. When I was healing as disc I could sometimes end up spamming fheal while keeping pom/shield/penance on cooldown and at other times just jump around waiting for the tank to get hurt. DPS is easy to simulate, healing isn't

Could you please explain what exactly that means?
I assume by increasing/decreasing you mean positive/negative values for scale_*. I'm wondering why the sign would actually matter. Do you (or dedmonwakeen) have an explanation for this? Did you use different absolute values, too?
And would you mind posting the gear sets?

What he means is that for the first set of tests they increased the various stat values (spellpower, crit, haste, etc) in turn to see how much the overall dps increased for each change.

For the second set of tests they decreased the various stat values in turn.

In theory you would expect the results to be the same both increasing and decreasing, and in practice thats pretty much what you get (given expected margins of error). I guess the main reason for doing the test is just to confirm that it actually works out properly in practice.

The differences in spirit and int between increasing and decreasing these values are perhaps slightly larger than I would expect, but the differences are still pretty tiny. Certainly nothing that will make a meaningful difference when making gear choices.

What would be the average dps over the 3 sets?

Could you maybe elaborate a bit on what that bug is and what impact it has? Does it mean that any values you get from the normal way can't be trusted no matter how many iterations you run? At a quick glance these values doesn't look wildly different compared to what we've been seeing on the SampleOutputPTR page. Except maybe the hit rating value making a big jump when going to the t8 gearset.

I have a fewe questions for one of you simcraft experts

T7 gear rotation: http://simulationcraft.googlecode.com/svn/trunk/Priest_T7_14_00_57.simcraft

It states the following priority system (pain excluded): MB > DP > VT > Flay

T8 gear rotation http://simulationcraft.googlecode.com/svn/trunk/Priest_T8_13_00_58.simcraft

This one states: VT > DP > MB > Flay

So my question is, how come the T8 priority system prioritizes VT over DP but the T7 system prioritizes DP over VT? We get talents and a setbonus in 3.1 which makes DP better than it currently is so I don't get the swapping around.

My 2nd question is, how come MB has such a low priority with the T8 gear? I thought the haste-setbonus-effect would make this spell more desirable to keep on cooldown.

A remarkable conclusion is that when you're not hit-capped, gemming for +hit is desireable on all gear levels.
you'll probably be better off getting a new +hit item, but still.

This is because upon testing of the T8 profile I saw that VT>DP>MB>Flay did better than what it was in the T7 set and so I changed it for the T8 one. I just didn't bother changing the T7 one too. (The T7 profile will very soon be obsolete anyway. The current T8 profile might very well be renamed to T7 once 3.1 hits and the new T8 profile will have T8 BiS gear).

Aha I see. Did you try with MB prioritized higher so the haste-bonus has a higher uptime?

Is it possible to come up with some formula or spreadsheet where you enter in your total spellpower, crit, haste (and whatever other stats), and you can find out what the actual pseudospellpower values of various stats is?

I know they change as your gear lvl increases, and it seems the better gear you have, the more important crit becomes over haste.

Updated the first post to reflect the latest build of SimulationCraft as well as adding in BiS T9.

The number of decimal places in the scale factors increased to 4 to reflect the size of stats these days.

The T9 258 set is using both the 245 and 258 versions of the Reign of the Unliving/Reign of the Dead trinket.

If it turns out that we're unable to use these together I'll update it replacing the 245 version with Flare of the Heavens which is the next best choice.