Warlockery: By the Numbers

I just want to start off by saying: I’m a bad bad blogger. I’m thrilled to death when I get email from a reader.. but unfortunately it’s taken me -way- to long to respond to this one.

A while back I got an email from Megan.

Megan wrote:
Hi there, I was reading your blog today and looking at the list of lock gear available from 10’s 25’s etc and I couldn’t figure out how you got your point values on each item as well as for the slot, is there another post I’m missing that explains the point system used?

I wanted to answer Megan immediately.. but the quick answer wasn’t going to be the full answer and really that question needed a full answer. Add to that, the numbers I used originally to generate those lists are no longer completely correct since 3.1 fiddled our values around. I’m currently working on updating those lists.

Sokay. Here goes. The numbers in my gear lists are based on the pseudopoints I calculated for the gear based on the output from Simulationcraft and implemented in the addon PAWN and lootranks/wowhead.

The grunt work

If you load up Simulationcraft, and turn the crank, Simulationcraft spits out the following scale factors.**

Warlock_T7_53_00_18 Str=0.00 Agi=0.00 Sta=0.00 Int=0.36 Spi=0.91 SP=1.58 AP=0.00 Exp=0.00 ArPen=0.00 Hit=1.80 Crit=0.91 Haste=1.06 Wdps=0.00
Notes: assumes you use your succubus, have 4-piece T7 and spec 53/00/18

Warlock_T8_53_00_18 Str=0.00 Agi=0.00 Sta=0.00 Int=0.23 Spi=0.71 SP=1.51 AP=0.00 Exp=0.00 ArPen=0.00 Hit=2.19 Crit=0.77 Haste=0.96 Wdps=0.00
Note: assume you use your succubus, have 4-piece T8 and spec 53/00/18

Warlock_Lilac Str=0.00 Agi=0.00 Sta=0.00 Int=0.17 Spi=0.67 SP=1.37 AP=0.00 Exp=0.00 ArPen=0.00 Hit=1.19 Crit=0.69 Haste=0.77 Wdps=0.00
Note: assumes you’re me.. wearing my clothes.. using my spec. MMV.

Note the plethora of caveats. Regardless, here’s the stats side-by-side and graph.

Warlock_T8_53_00_18 0.23 0.71 1.51 0.77 0.96
Warlock_T7_53_00_18 0.36 0.91 1.58 0.91 1.06
Warlock_Lilac 0.17 0.67 1.37 0.69 0.77


Things to note: Regardless of which gear set/tier proc is chosen the relative weight of the stats remain constant. Haste doesn’t suddenly become more important than spellpower at a certain gear level (at least according to this limited sampling).
When using these to establish a Pawn or lootrank scale I do not include “hit” in the dps calculation scale.

In my opinion, getting to hit cap is -the- most important thing you can do. When you’re swapping around gear pieces to get to that cap it’s important to know both how much “pure” DPS a piece will add AND the actual +hit number. I track these values separately. If I’m AT the hit cap then any additional +hit is worth 0 BUT in the future that piece with hit MAY be worth something if I need to swap gear around to make the hit cap.

The Nitty Gritty

So based on the output of Simulationcraft for me based on my gear, here’s my bare-bones Pawn scale:

( Pawn: v1: “Warlock (dps)”: Intellect=0.17, IsShield=-999, CritRating=0.69, Spirit=0.67, IsMail=-999, IsLeather=-999, IsPlate=-999, HasteRating=0.77, SpellPower=1.37 )

To this scale I add my personal rating for gems. I choose “Use the correct gem color for each slot” and then enter a calculated scale value for that color of gem.

For example: If I have a red gem slot I’m going to put a Runed Scarlet Ruby into it. So for a red gem I take the stat value of the gem (19) and multiply it by the scale for the stat I’m getting from that gem (spellpower = 1.37). So in my red gem slot I’d put a scale factor of 26.03 (19 * 1.37).

My gem scaling:
Red: 26.03 Runed Scarlet Ruby
Yellow/Orange: 18.49 (12.33 + 6.16)Reckless Monarch Topaz
Blue/Purple: 17.69 (12.33 + 5.36) Purified Twilight Opal
Meta: 34.25 Ember Skyflare Diamond (I don’t have a way with the scale to account for “2% more int” so I left it out.. so I just need to keep in mind that a head piece with a meta gem socket is actually worth a little bit more than the PP will show)

So I end up with this as my final Pawn scale:

( Pawn: v1: “Warlock (dps)”: Intellect=0.17, RedSocket=26.03, CritRating=0.69, IsMail=-999, MetaSocket=34.25, IsLeather=-999, IsPlate=-999, HasteRating=0.77, BlueSocket=17.69, YellowSocket=18.49, SpellPower=1.37, IsShield=-999, Spirit=0.67 )

I use this “dps” scale to generate my “hit” scale:

( Pawn: v1: “Warlock (hit)”: Intellect=0.17, RedSocket=26.03, CritRating=0.69, Spirit=0.67, IsMail=-999, IsPlate=-999, IsLeather=-999, MetaSocket=34.25, HitRating=1.5, BlueSocket=17.69, YellowSocket=18.49, SpellPower=1.37, IsShield=-999, HasteRating=0.77 )

The only difference between the two is that in the “hit” scale +hit has a value. This is so when I look at a piece of gear, if the “DPS” value and the “hit” value are the same then I know the piece doesn’t have any +hit. If the number are different then I know the the piece has hit and the “DPS” number is the “raw dps” value of that piece of gear (good for swapping gear around to get the highest DPS value and yet still to all of my hit rating).

Lastly, from these scales I can generate gear lists from Lootrank or Wowhead. At this time I prefer Lootrank.

Lootrank links Pre-Naxx Gear, Pre-Ulduar, Raid/Badge


Pseudopoints – the number of arbitrary points one piece of gear vs another piece of gear. In theory the gear with the most psuedo points is the best piece. In actuality the points are only as good as your scale.. and sometimes you may need to stack certain less-valued stats for certain gear sets (ie, getting frost resist set, the frost gear has lower dps than regular gear but is more valuable because of the added survivability)

Simulationcraft – a desktop application written for the express purpose of theorycrafting World of Warcraft specs/gear/dps rotations to determine the “best”. (Simulationcraft Homepage)**

Scale Factor – The scale factor for a particular stat represents the increase in DPS per unit increase of that stat.

** A full explaination of Simulationcraft is WAY beyond the scope of this blog post. I recommend you consult the Simulationcraft documentation and WIKI and the relevant class-specific Elitist Jerk’s threads.

12 thoughts on “Warlockery: By the Numbers

  1. Hey, just found your blog a few weeks ago, revamping my ‘lock spec. So, I am trying to import these Pawn values but it keeps complaining about invalid tag.

    Be cool if you could repost that pawn scale or email it to me,


    Eric “Elkagorasa”

    Elkagorasas last blog post..clank clank clank went the rope on the flag pole

  2. I found that I needed to unwrap them. I copied them into Notepad and pasted the single line to PAWN without any issue. I guess it didn’t like it being on 3 lines.

  3. Hi, in your “hit”-scale how do you get the value for hit? Is it just an educated guess or does it come from any calculation? You use another value than the one given by SimulationCraft which puzzles me a bit

  4. Hit is uber.. it’s the #1 stat. I value it on a totally separate method from everything else. When I’m building a new gearset I do the following:

    1. Go thorough all pieces of gear and find whatever has the highest score in Pawn. Put it all on.
    – Is my hit at 6%?** if yes, save this as my “heroics” gear set. If no, goto 3.
    – Is my hit at 17%?** if yes, save this as my “boss” gear set. If no, goto 3.
    3. Dig through all my gear and find the next piece that -has- more hit than the piece I have on that decreases my pawn score by the least (this is fudged a bit if I’m also considering gems/enchants).
    4. Put that piece of gear on. Goto 2.

    **When considering hit I will include 3% from spec if I’m spec’d into suppression. I don’t count the 3% from others in the raid or the 1% from Draenei in my group since I can’t depend on having it. So usually for raiding I gear for 14% hit. I would rather have too much hit and NEVER miss than have too little and chance 3%-4% miss rate. This does mean my final DPS is lower than it could be in some raids. I’m ok with that.

  5. Thanks for the explanation. I am used to the “Rawr” application from my Druid, but since the Warlock part of it is still in development I am looking for alternatives.

    I found one which looks nice:


    A bit like Pawn, but:

    1. Takes hit rating cap into account if you wish
    2. The author also tries to model trinket effects

    I just tried to install it and will test as soon as my realm is online again.

  6. Ok. since I wrote this I’ve revised my numbers to use the scales from 3.3.. and it appears that the Chaotic Skyflare is better then the Ember πŸ™‚

Comments are closed.