A demo warlock Warcraft Logs parse-reading

Get it? It’s a demonology warlock raid parse comparison but also an example warlock comparison using Warcraft Logs? It’s a demo warlock raid parse comparison!

Puns aside, I see Warcraft Logs introductions, and they usually show how to upload a log and then “hey, it makes pretty graphs and is a pretty powerful tool,” and then never goes any further. Perhaps it shows off the tools — look, you can make pins! and here are all the things you can do with pins! — but the guides I’ve seen that aren’t actual walkthroughs of a raid parse specifically meant to improve a specific person often just stop with the abilities of Warcraft Logs rather than what you should do with them.

Handing someone a full toolbox doesn’t automatically make them a skilled carpenter. The information sitting in the powerful & useful tool that is Warcraft Logs is still overwhelming to the casual raid parse reader, because they don’t know what to do with all of it.

However, I haven’t quite figured out how to generally write or make a guide that explains why to use a pin or a graph or a table or the replay to judge J Random General DPS Concept. Why? — It’s very specific per spec and per fight, even per attempt depending on what you’re doing, and even per question depending on what you want from it all. Trying to anticipate all the questions is just overwhelming by itself and hardly solves any of the overwhelming feeling.

The only part I’ve been able to figure out is that reading a parse is not so much about bringing a checklist of frequent problems to see if any apply to you. Reading a raid parse efficiently is more about bringing your knowledge about your spec, your class, and the specific encounter to the raid parse and critically applying that knowledge to the various results you have in the graph/table/chart/replay.

You have to already have questions or possible reasons why you didn’t do so great when you come to the parse, and the parse will help you answer those. The parse doesn’t magically conjure answers for you — you have to already be looking for them.

To start figuring out how to write a guide for reading a raid parse in order to better yourself, I recorded my own, where I compare myself as a demonology warlock to another demonology warlock who did better DPS on this one fight than I did.

First off — I’m sorry, I didn’t realize the resolution was that terrible until it was already uploaded & processed on Youtube. Which took multiple hours (because, well, it’s an hour-long video itself). And I didn’t feel like redoing it. Sorry. It’s probably actually because I recorded it at some low setting. It’s clearest to read in the smallest player, but I think it’s fine, since it’s mostly the graph shapes/animations that I’m focusing on, not the text.

My video is actually uncut as well as unscripted. I figure this is better so you can see me meander through all the different tabs, discovering new information and arriving at new conclusions organically rather than me telling you what to look for. I use myself as the worse warlock, so I’m only embarrassing myself in particular, not anyone else. And it’s also part of my after-raid homework to log-dive my own logs like this.

I was very, very tired when I recorded it — insomnia is never fun, and you’ll hear me yawn a couple times — so I actually make quite a few obvious knowledge mistakes. I usually catch myself, but there are others that are there that I now facepalm myself at. Sorry! When I’m writing articles, I am actually awake and actually remember things like Grimoire of Service is 2-minutes not 1-minute, and realize that the guy has higher DPS overall instead of just having a higher opener. The intent of the video is less to show off how bad I am at demonology and more that, hey, look, I’m imperfect at this spec and even I can get some ways I can improve out of a raid parse. (It’s also easier to show the differences between a not-great warlock and a good warlock than it is between a good warlock .and a slightly-worse warlock!)

This is not meant as a be-all end-all to Warcraft Logs guides. It doesn’t even cover everything there is to use — I don’t look at how tanks or healers could use the parse information, because, well, I’m not playing one right now. I’ve done a tank or healer parse comparison as part of being a veteran raider rank in my guild, in order to pinpoint why or how a guildmate needs to improve, but I don’t normally play a tank or healer, so I won’t pretend to know 100% how to read those parses. I can certainly open up a 101 guide to a guildmate’s class & spec & figure out the really obvious ways they are screwing up, but let me be clear, the only class I can possibly be an expert in is warlock, and even that I’m, well I’m good, but I’m not perfect.

This is meant as something to help get you started at reading a Warcraft Logs parse, especially if you play a warlock (or maybe even a DPS in general). So hopefully it helps you logdive on to figure out why you may or may not be doing things incorrectly and how you could do better.

A demo warlock Warcraft Logs parse-reading

Draenor warlock guides from me


In March 2014, when the class columns ended at WoW Insider, I joked that it was my fault. I’d beaten the Warlock Columnist Curse twice over, and it was so mad it lost that it had to take all the class columns with it.

Dear Warlock Columnist Curse:

Look at all those WI alums. Long live class columns!

— Poneria

P.S. Magdalena & Chaithi are also awesome people anyway.

I had to double-check the email when it came in the summer from BradyGames — it couldn’t possibly be true that I’d be invited to contribute to a real, official, printed-on-paper book, right? But it was true — Anne had put my name into the pot and I got to work with her again! (Ken Schmidt is another other editor/writer on the project.)

It’s surreal for me to see my writing in the same context as the smell of a newly printed book, but there it is. That’s my writing there, all over the Warlock section. That’s mine. What? Wow, that still repeatedly blows my mind.

I’m keeping my copies for personal reasons, but a couple of the other section authors are holding giveaways for one of their copies:

  • US entries: Chase Hasbrouk aka Alaron @ The Fluid Druid
  • AUS/NZ entries: Matt Sampson aka Binkenstein @ TotemSpot
  • Sorry, EU! None that I know of

I think maybe Bink’s giveaway is over (it was limited to the end of BlizzCon), but Alaron’s is still going on for another day.

If you’re reading me here, you probably already know all of the information in the Warlock section. It’s a newbie’s guide — veteran players will know where to find more advanced DPS tricks online — but I’m still super proud of the work I did from July to the beginning of October.

Right after I sent in my final drafts, Wowhead approached me to update the warlock class DPS guide, so I did that, too. It’s got some typos in it as I still haven’t figured out how they copy my text into the tooltip generator, but I’m still working against burnout to clean some of it up.

Burnout? What burnout?

I’ve been guide-writing in some form or another since July, or 4 months straight. Moreover, other IRL aspects of my life have started demanding more. While I’d love to put out quality like WoW Insider or Wowhead can do, I simply don’t have as much time to spend on it as I once did. I got a promotion in my “day job,” so I basically did IRL work, came home, did guide work, slept, repeat. I didn’t have much recharge time for just me aside from mindless D3 slaughtering.

I’ve been quite burnt out on writing guides for the past couple of weeks, especially when it comes to a really broad audience — optimizing gear lists, for example. While I have organized a few gear list notes for myself, I haven’t gotten around to doing the typical column Ultimate Bookmarkable Gear List guide stuff because getting it all “correct” is super stressful for me. There is also a wide variety to gear in Draenor, at least to start with, and while that’s fun for the player playing through the game, it makes guide-writing that much more stressful, because it’s more to organize.

But that doesn’t mean I’m stopping guide-writing. I’m just taking a break, is all. I’ve earned one.

Knowing myself, I will probably have lists worked out as I level in Draenor — I’ve started my raid loot lists, already. I also figure that the people who would use a gear list right out of the starting gate have probably already found their roadmaps and therefore don’t need my guides anyway. So I’m not worried about it.

But if you’re going to take your time leveling and you like my writing, feel free to read what I have to list, whenever it is that I share it.

Draenor warlock guides from me

Welcome to 6.0 – Warlock Guides

You might know that I provided the information in Wowhead’s warlock class guide when the boosted 90 feature came out in March. Wowhead came back to me to update the guide for level 100 with some tips for the pre-expac patch.

It’s mostly the same tips, except we obviously don’t have the level 100 talents or some of the balance from Draenor Perks.  The core mechanics are still the same whether you’re level 90 or level 100.

While I’d been writing my own guide (that we’ll see…llllllaaaaterrrrr…) concurrently, the warlock community’s opinions on various things (Which L100 talent? What stat priority? Etc.) recently started to solidify, so I did my best to double-check my advice for the Wowhead guide. It’s possible I still have some slightly incorrect top-end things, as I had to write & commit-to-submit* some things before the other warlock guides were published.

*By “commit-to-submit” I mean I may not have submitted the whole piece yet, but there’s a point where you have to call a section “finished” even though it’s not quite done yet because you don’t have time to both finish it AND completely write another section.

So here it is!

But Pon, you got things WRONG! OK, so tell me nicely & I’ll correct them. If you tell me nastily, I’ll probably still verify & correct them, but you’re still an asshole.

It’s always cool to know where other resources are if you don’t like my windy, wordy ways. The warlock theorycrafting community appears to be split more or less across either the Icy Veins or MMO-Champion forums. The EJ warlock forums have been pretty dead since 5.2. I’m still looking around to see what places pop back into activity with the new expansion.

  • Gahhdo is the main SimC warlock dev (that I know of) and Icy Veins’ Zagam did the Icy Veins guides that have updated for 6.0.2 (aff/demo/destro)
  • Lockybalboa/Zagam/Liquidsteel did stickied affliction/demonology/destruction guides (respectively) on the IV Warlock forums
  • On MMO-C Warlock forums, we have Cabana with the affliction guide and Woz aka @NotsWarlock with the demonology guide. Brusalk is usually the destro go-to on MMO-C forums, but there isn’t a destro stickied guide as of writing.
  • My guide on Wowhead is all three specs in one, but is a more general & condensed version than the above detailed spec guides.

Do you want a transition guide?

What’s Next?

I plan on doing guides for each spec on the level 100 Silver Proving Grounds gating mechanic for random heroic dungeons. I want to do the PGs as faithfully to a new player who just hit level 100 as I can, so I haven’t been doing them as much on beta as I’d like to partially because of the premade PvP gear that the beta experience was assuming you had.

I also had problems finding blocks of time when the servers were also up to do it. But I digress.

I’m of the opinion that just because I am a pure DPS class doesn’t mean that the Proving Grounds can stop at one spec for completion. It’s one thing to spec to complete the hardest of the hard content (e.g. Mythic raid content), but to spec just to complete a simple gating mechanic seems stupid. So while I’ll probably blow through with destro (it’s easiest), I want to do PG guides for all the warlock specs, possibly up to Endless.

“But, Pon” — you say — “Silver is pretty easy! You don’t need a guide!”

True, if you’re the type of player who knows how to find a guide online outside of the game, you probably already have enough smarts in your head to figure out Silver waves without needing a guide.

I’ve never been great at doing bleeding-edge kind of guides where you learn the best of the best way to do things, so I’m going back to what I’m good at and what I enjoy — doing guides for newer warlocks who are stepping up in skill. I plan on doing PG guides per spec that show off what this wave level shows you about your toolkit as a [spec] warlock. For example, Endless as destruction really teaches you about Havoc use.

Another thing I want to do is going back through all the retro raids — again, as each spec with an eye for non-warlocks learning on an alt. This seems like a massive project, and I don’t really have a timeline for it. It’s a thing on the bucket list more than it’s a thing on the project list. One difference is that I plan to take not only my main (fer mounts, duh!) but my quest-greens/boost-90 warlock through to see what you can do with a fresh alt versus the completely decked-out main.

Finally, a lot of vocal followers liked my prose-style gear lists, so I figure I will do those again for fun. I’ll start with heroic dungeon loot & maybe crafted pieces, and then move on to raids/world bosses.

Welcome to 6.0 – Warlock Guides

Warlords of Draenor Beta: Healthstones

Self-healing is a weird and intricate balance for warlocks because we tend to use our health as a semi-resource bar often enough for healing to matter. There’s the obvious Life Tap that we’ve had for many years, but more recently, Mists of Pandaria brought abilities that cost health, like Unbound Will or Burning Rush.

I feel like this is a subject that the devs still don’t know how to balance correctly yet. The changes tend to swing from not enough to far too much/overpowered, as we know well enough with Drain Life’s evolution, or even with how Harvest Life evolved in MoP beta. We had the Life Tap fiasco in MoP beta where our self-healing was so nerfed so far such that we couldn’t recover from Life Tap, so we eventually either went OOM or dead. So we warlocks have this weird balance where we need enough self-healing to make up for the health-costing abilities, but not so much self-healing that we can basically ignore boss mechanics when we feel like it or an otherwise lack of outside healing.

The major change for Healthstones is that it’s going to be a 1-use per fight (down from 3-5 use) and that it’s tied to the same CD as a healing potion (in Warlords, Healing Tonic). Additionally, the amount a Healthstone heals for also got nerfed.

Better-Good Design

When the newest build came out, Zinnin () asked Celestalon if Healthstones needed to be nerfed further, because they were already below the potion’s heal. Celestalon replied that the Tonic-Healthstone relationship was supposed to be more like the 300-275 stat food relationship, where one is technically better but the other isn’t absolutely terrible.

The difference between 275 & 300 stat foods is 25 stat,. If you represent this in a ratio, 300:275, then we get 1.09:1. It was further of a minor difference because in Mists, 25 Intellect (e.g.) was almost nothing of a boost compared to what you’d get from a gem or an enchant.

Let’s see how healthstones & tonics compare.

Healthstones: Build 18738

I got to thinking about the influence of Versatility and Blood Pact’s healing increase component, so I asked around on Twitter for a body to eat a handful of cookies for me and Dayani (@healiocentric /Healiocentric blog ) volunteered. After doing some gear switches to play with Versatility a bit, with both of us eating cookies and sharing numbers in party chat, we figured out that the tooltip takes into account the Versatility the user has, but not the Blood Pact passive. The total heal takes into account both.

I’ve unfortunately lost all the exact numbers because I didn’t record our party chat until half of it was cut off by the chat log’s limit. But the important thing is that I can replicate them!

The flat Healthstone heal is 15%.  My maximum health will always be 289,740. (Unless I exchange some gear or drink a stamina flask, but warlocks always have the 10% raid buff due to the Blood Pact passive.)

Then you take into account your Versatility. On my premade orc in 660 PvP gear, I have 1.63% Versatility. 15*1.0163 = 15.2445%. The tooltip still says 15%.

Then we need to factor in how I’m a warlock who is at least level 80, which means I have the Blood Pact passive. Blood Pact is a complex passive in Warlords that has 3 parts, but the important part here is that line about “increases all healing you receive by 10%.”

So we take our 15.2445 and multiply it by 1.10, which gives us 16.76895% of max health, which is 48,586.35573 health. When I take a health dip (by using the equipment sets feature to go full nude then fully geared again in one click), then use a healthstone, I receive a heal for 48587. It’s off by 1, so I know from Theck’s lessons that there’s some rounding going on in there. But I don’t care enough about 1 health point to go find out where.

When I put raid buffs on with a cauldron in Shattrath, my theory holds.

Base heal (15%) * Versatility (1.0463) * Blood Pact healing passive (1.10) = 17.2% heal. I should get something around a 50020 heal and I got 50021.

Dayani’s cookie-eating corresponded — on a shaman, her Versatility affected her cookie heal. (Also A’dal’s pesky Shatt buff affected it!)

Healing Tonics 

At level 95, I’m still in my raid gear when I got copied over, which gives me a max health of 148,980 and some Versatility of 1.22% (because I’m human, so I get free Versatility). I also unglyphed Healthstone and didn’t have Grimoire of Sacrifice buffed when I logged in. I was buffed with Dark Intent, but that’s spellpower and multistrike, so who cares.

My crafted Healing Tonic says it restores 56,681 health. When I actually drank the Healing Tonic, it healed me for 62350.

That’s accurate, because 56681*1.10 (to account for Blood Pact) = 62349.1.

The tooltip for the Healing Point takes into account Versatility, as it turns out. The Wowhead tooltip scaled down to level 95 for Healing Tonic is 56000. 56000*1.0122 = 56683, which is close enough for government work. Same thing when you apply it to the level 100 version — 68000 — though the Versatility % tooltip must do some rounding. Eventually you do it the long way through rating and Dayani is amazing again with math and figuring out the rounding spot, and you get basically the correct answer where the 68000 Tonic heals a warlock for 68000*1.0163 = 69109 on the tooltip, but effectively 76019 when we include Blood Pact.

Comparison: Tonics & Stones

So here’s the deal. Give or take some rounding errors, Healthstones work like this:

% max health heal =15% max health * (1+Vers) * Blood Pact

And Tonics work like this:

# heal = 68000 * (1+Vers) * Blood Pact

Assuming a warlock standing around, you get Blood Pact; if you’re some other class or situation, you can get some other healing done/taken effect going on (like A’dal’s Swiggity Swattrath Buff).

It comes down to whether a 68k heal is better than 15% max health. If 68k was 15% of your maximum health, you’d have 453,333.3333 repeating health. So maybe when we get to the end of the expansion and you’re sitting at 453,334 health, your Healthstone will finally match Healing Tonic.

Right now, in 660 PvP gear, we have 289,740 max health (15% = 43461). Even if I use Shadow Bulwark off a voidwalker sacrifice, I can only get up to 376,662 health (15% = 56499.3).

68k right now is about 23.5% max health. That’s almost 1.5 times a Healthstone’s heal. That’s not a better-good competition, that’s a good-crap competition.

At 1.09, Healthstone heals 62385 health, or 21.5% max health.

At 1.10, Healthstone heals 61818 health, or 21% max.

At 1.15, Healthstone heals 59130 health, or 20% max.

At 1.20, Healthstone heals 56667 health, or 19.5% max.

At 1.25, Healthstone heals 54400 health, or 18.8% max.

If I take a look at the Shadow Bulwark situation, where I increased my maximum health without touching my Versatility, where I increased my healthstone’s heal without increasing the tonic’s heal — the tonic is still better, 68000 to 56499 being 1.20 ratio. You’re only going to get to a stamina-only increase like that through a temporary raid CD. Stamina flasks don’t even begin to compare to that (30% max health is in the realm of 85k; a stamina flask is 18k health).

I don’t feel like or think that’s balanced. 15% is way too low.

If 68k is the accepted base heal for a potion, the Healthstone should be brought back up to a 20% max health heal, to at least give a little choice. To go to the 1.09 ratio of the stat foods, Healthstone would have to be buffed from live realms to 21.5% max health.

Otherwise, if we’re tuning to Healthstone’s 15%, the Tonic needs to be nerfed to something like 48k to get to a ratio of 1.09. The compromise 1.20 ratio would be a 52k heal.


Buff Healthstones and/or nerf Healing Tonics.

Edit: Already Changed

I @’d my post to Celestalon, and he replied that Healthstone has already been changed in the next build to be a flat base heal of 50k instead of based on max health percent. So now both warlock formulas for Tonics and ‘Stones are: # heal = [68000 (Tonic) / 50000 (Healthstone)] * (1+Versatility) * Blood Pact.

I’d reasoned out a 1.20:1 ratio as fair, which left me at thinking buffing Healthstone to heal 52k. Not too shabby for speculation!

Warlords of Draenor Beta: Healthstones

TC101: Multistrike & Versatility

I’ll revisit how to calculate Intellect when I get around to analyzing Spellpower, but for now, I can do more of the simple things like rating conversions.

Multistrike and Versatility are the easiest of the five secondaries to discuss because they have a rather uncomplicated and passive effect on our damage done. Critical strike chance, haste, and mastery all have particular spec implications of what they do for damage, even if you don’t consider the specialization attunements.


Multistrike is pretty simple to figure out.

The full 660 PvP premade gear on the PvP level 100 character has 590 Multistrike rating in total. By removing pieces of gear and comparing the percents, I figured the appropriate rating conversion.

  • 590 = 8.94% –> 590/8.94 = 65.9955 rating
  • 513 = 7.77% –> 513/7.77 = 66.02 rating
  • 411 = 6.23% –> 411/6.23 = 65.97 rating

I guessed that the actual rating conversion is 66 rating = 1.00% Multistrike, that the percentage was determined by gear_total/conversion, and that the final percentage was rounded.

  • 590/66 = 8.939393…rounds to 8.94%
  • 513/66 = 7.77272727…rounds to 7.77%
  • 411/66 = 6.2272727…rounds to 6.23%

And I think I’m right.

When it comes to enhancements, there are static enchants, a weapon enchant proc, a food buff, a flask, and a raid buff. These all appear to be additive by rating or flat percentage.

Remember, my base gear had 590 = 8.94% Multistrike.


Enchant Cloak – Breath of Multistrike

  • +200 MS
  • 590 + 200 = 790
  • 790/66 = 11.969696…

I had actually forgotten to test the static enchants, so that screenshot is on my undead warlock. The rest I was testing on the blood elf warlock. Racial pick doesn’t matter for Multistrike, anyway.


Draenor Multi-Strike Flask

  • Blizzard hasn’t copyedited how it wants to spell Multistrike, I guess.
  • +400 MS
  • 590 + 400 = 990
  • 990/66 = 15


Food – Rylak Crepes

  • +125 MS
  • 590 + 125 = 715
  • 715/66 = 10.8333333…


Flask + Food

  • + 400 + 125 MS
  • 590 + 400 + 125 = 1115
  • 1115/66 = 16.89393939…


Flask + Food + Buff

The buff is just a flat percent addition on top of whatever rating total you have.

  • 1115 rating + 5%
  • 1115/66 = 16.89393939…
  • 16.89% + 5% = 21.89%


Weapon enchant proc — Mark of the Frostwolf

Temporary enchants also just add rating. The weapon enchant proc gives 500 Multistrike for 6 seconds.

  • Started with flask, food, buff = total 21.89%, 1115 rating
  • 1115 + 500 = 1615
  • 1615/66 = 24.469
  • 24.47% + 5% = 29.47%

So, in summary, Multistrike chance = round(total_rating/66 + raid_buff_percent), where total_rating includes gear, enchants, flask, and food.


Versatility has two effects to it: damage or healing done and damage taken. I started referring to these as DD/DT, so I’ll often write “V” numbers as x% / y%. The DT number is always half the DD number. Thus, depending on what you want out of Versatility, you can find it by:

  • %DD = round(total_rating/130)
  • %DT = round(total_rating/260)

If you add anything to the DD percentage, you can only add half to the DT percentage to get the correct number. It’s easier to make that mistake of blindly applying a whole percentage (like the raid buff) to the DT number, so I think it’s better to think of them as two different rating conversions, depending on which one you’re aiming for. This way, the raid buff adds 390 rating, instead of trying to remember to halve the percent.

Versatility is also a good (current) example of when tooltips are lying out their textual asses.

If you equip or unequip gear, the rating number changes as it should. If you enchant gear, the number changes.

But if you eat, flask, buff up, or switch to a human (with the racial affecting Versatility) the rating doesn’t move, despite showing an increase in percentages. So when I write “212” (#), I mean that the tooltip currently says it’s 212 rating, but I know it’s actually the (#) rating applying.

Base Versatility

If you take off all your gear on a non-human, you get 0 rating with 0% DD or DT. On a human, you get +100 Versatility rating as a passive racial, but the tooltip will say 0 rating. To find the conversion, we consulted some numbers.

  • Dwarf 660 premade: 212 = 1.63%/0.82%
    • 212/1.63 = 130.0613
  • Human 660 premade: “212” (312) = 2.40% / 1.20%
    • 312/2.40 = 130
  • Human premade, naked: “0” (100) = 0.77% / 0.38%
    • 100/0.77 = 129.870

I estimated that the rating was 130, so the DT rating would be 260. The percentages are obviously rounded at the end.

  • 212/130 = 1.6307
  • 212/260 = 0.8153
  • 312/130 = 2.4
  • 312/260 = 1.2
  • 100/130 = 0.7692
  • 100/260 = 0.3846

Gear changes

This is my undead warlock with the premade gear, but with some changes. She has Crit/Versatility boots on (+102 V) instead of Crit/Mastery boots, and I’ve enchanted her cloak with 200 Versatility.


  • 212 starting V + 102 (boots) + 200 (static enchant) = 514 rating
  • 514/130 = 3.95384% DD
  • 514/260 = 1.97692% DT

Flask + Food + Buffs — Additive

Same thing with flask and food that happens with gear, except the rating tooltip is bugged. The flask adds 400 rating while the food adds 125 rating.

The raid buff adds 3% Versatility, which adds 3% to the DD and 1.5% to the DT. Rating-wise, it’s as if you had added 390 rating.

  • +food = “212” (337) = 2.59% / 1.30%
  • +flask = “212” (612) = 4.71% / 2.35%
  • +buff = “212” (602) = 4.63% / 2.32%

Let’s do math:

  • 337/130 = 2.5923
  • 337/260 = 1.2961
  • 612/130 = 4.70769
  • 612/260 = 2.35384
  • 602/130 = 4.6307
  • 602/260 = 2.3153

Fully Buffed Human

If we get a fully buffed human, I expect:

  • 212 (premade gear) + 100 (human racial) + 125 (food) + 400 (flask) + 390 (raid buff) = 1227 rating
  • 1227/130 = 9.4384
  • 1227/260 = 4.7192

And what do we get? Remember, it’s “212” rating.



TC101: Multistrike & Versatility

TC101 Homework: Character Sheet Stat Calculations

We all see the finished guides and models from theorycrafters, but despite theorycrafters and guidewriters alike swearing they’re not actually gods who get everything right on the first try, we don’t really see the trial-and-error experiments or calculations behind theorycrafting.

I’ve done plenty of following along with theorycrafting over the years, but I’ve not actually tried to derive the equations myself until now.  Theck’s done some teaching; I figure now it’s time to show off some learning. This post will probably be full of mistakes, but that’s OK, because learning. I hope someone out there enjoys this sample of my theorycrafting “homework.”

  1. I do write out on my notes to myself like “let’s do X” “OK that didn’t work, why” “here’s why” when doing math/science things because I got used to doing that when explaining math to myself in school. Y’know, the whole “let X equal 5, let Y equal 6, *do funky equations*” stuff. So it’s not just my guidewriting slipping in — I really do write to myself like that!
  2. If I slip between singular and plural, that’s my bad — Theck’s using the first person plural because he’s walking us through it, and I was reading along with Theck while doing these calculations, but I typically write in first person singular, so um, yeah, confusion.

Determining Base Intellect

Pulling the base stats for warlocks, we get this:

Strength 551
Agility 985
Intellect 1042
Stamina 890
Spirit 1155

And pulling the racial modifiers for warlock-able races, we get this:

Human Dwarf Gnome Worgen
Strength 0 5 -5 3
Agility 0 -4 2 2
Intellect 0 -1 3 -4
Stamina 0 1 0 0
Spirit 0 -1 0 -1

My blog theme can cut off the table, so I split it into Alliance (above) and Horde (below).

Orc Undead Troll Blood Elf Goblin
Strength 3 -1 1 -3 -3
Agility -3 -2 2 2 2
Intellect -3 -2 -4 3 3
Stamina 1 0 0 0 0
Spirit 2 5 1 -2 -2

While I have created a level 100 premade of every warlock race on the PvP beta server, I decided I’d go with troll. I normally play a human, but troll is the race that SimulationCraft uses (particularly the affliction warlock model).

Variables: (Sorry, I’m not a LaTeX wizard like Theck is)

  • B = class_base + race_base = 1042 + (-4) = 1038
  • G = 2378


  • CS_Base = B = 1038
  • CS_Bonus = G = 2378
  • CS_Total = B + G = 3416

Basic math, woot! We can do this!


…1089 + 2496 = 3585.



I double-checked to make sure I was in no-selected spec. I switched to my human premade, and same thing — the numbers didn’t match.

The First Difference Between Warlocks and Paladins

Paladins can wear any of the 4 armor class items, even though they eventually prefer wearing plate. Warlocks, on the other hand, can only ever wear cloth. Paladins, like other higher armor class wearers, get a passive at level 50 that gives them 5% extra preferred primary stat if they stick to wearing all plate. Warlocks do not need this incentive to pick a particular type of gear, but to balance everything, we get a “congrats, you can only wear one thing!” 5% primary stat bonus at 50.

Another difference I noticed is that the theorycrafters I look to for learning theorycrafting — Theck, Arielle, Hamlet, etc. — are all players of hybrid classes. Theck’s first screenshot in the 101: Character Stats post where he’s determining the base stats clearly has his plate gear on — he just hasn’t chosen a specialization yet. The preferred primary stat for paladins can depend on spec — protection prefers Stamina, retribution prefers Strength, and holy prefers Intellect. So the passive doesn’t activate for paladins until you choose a spec.

Our problem above with the character sheet and Theck’s very first calculations example is that all warlocks like Intellect. That’s just how it is when you’re a pure DPS caster class. And the game knows this — that we only wear cloth and we only like Intellect. So it doesn’t need to take those considerations out when we’re not in a particular spec.

So let’s skip ahead to when Theck calculates that passive for paladins and see if it matches my numbers for warlock.

The Armor Skill Formerly Known as Nethermancy

Theck’s picking retribution; I’m picking affliction.

If I was right — that the base intellect really was supposed to be 1038 + 2378 from gear = 3416 total, what happens if we repeat Theck’s first error and just multiply by 1.05?

  • 3416 x 1.05 = 3586.8

The character sheet had displayed 3585 — which is close, much like Theck’s simple error is. So I’m probably right — the cloth armor bonus is already factored in for warlocks where it’s not for paladins. Let’s continue with the calculation.

  • CS_Base = floor( B x 1.05 ) = floor( 1038 x 1.05 ) = floor(1089.9) = 1089
  • CS_Bonus = floor( G x 1.05) = floor( 2378 x 1.05 ) = floor(2496.9) = 2496
  • CS_Total = CS_Base + CS_Bonus = 1089 + 2496 = 3585

Yay! The numbers match!

So, before we go to the next step, let’s reiterate all our variables:

  • AC_multiplier = 1.05
  • B = class_base + race_base = 1042 + (-4) = 1038
  • G = gear_stat = 2378

And all our formulas:

  • CS_Base = floor( B x match )
  • CS_Bonus = floor( G x match )
  • CS_Total = CS_Base + CS_Bonus

Buffed: 5% Intellect

I’m not a paladin, so I can’t buff myself with Kings, but I can stand in Shattrath with the beta cauldrons. I clicked a beta cauldron to get the raid buffs and clicked off everything except the 5% Strength, Agility, Intellect stats buff.


Like Theck’s ret paladin, the base for my affliction warlock hasn’t changed: it’s 1089.

If I apply the Kings buff like Theck has, where I multiply my total unbuffed Intellect by 1.05, I actually get a different comparative number than Theck does.

  • Theck multiplied his unbuffed 4023 Strength by 1.05 and got 4224.2, which would be correct according to his character sheet if he floored that.
  • I multiply 3585 Intellect by 1.05 to get 3764.25 — which still wouldn’t be the 3765 Intellect on my character sheet if I floored it!

OK. Well. Skipping ahead worked last time.

Theck’s all like “so the formula for the bonus is”:

  • CS_Bonus = G x match x multiplier + B x match x (multiplier – 1)

And I’m like, whoa whoa whoa, hold up, where’d you get that, Theck. So let’s try it out.

We got closest to the answer when multiplying the total unbuffed stat by 1.05. (We’re going to write this a little backwards so we can match Theck’s formula visually.)

  • CS_Total = multiplier x ([ G x match ] + [ B x match ])

Expand it out:

  • CS_Total = multiplier x G x match + multiplier x B x match

The problem is that it’s attributing all of that to the stat bonus, CS_Bonus. CS_Bonus = CS_Total – CS_Base. So subtract a CS_Base equivalent from each side.

  • CS_Total – CS_Base = multiplier x G x match + multiplier x B x match – B x match

We can clean this up a bit.

  • CS_Bonus = multiplier x G x match + (multiplier – 1) x B x match

Which looks like Theck’s. Cool beans. Math!

  • CS_Bonus = multiplier x G x match + (multiplier – 1) x B x match
  • CS_Bonus = 1.05 x 2378 x 1.05 + (1.05 – 1) x 1038  x 1.05
  • CS_Bonus = 2676.24
  • Character Sheet bonus = 2676

Theck got a one-off; we got what we’d get if we floored it.

We’ll still do it Theck’s way for the sake of learning.

  • G = 2378
  • B = 1038
  • match = 1.05
  • multiplier = 1.05

And, with the help of some spreadsheets to make the calculations easier, we get for the first batch of flooring:

  • F1 = 2675.295
  • F2 = 2675.495
  • F3 = 2676.195
  • F4 = 2675.745

We actually get F3 for our answer — which doesn’t match Theck’s ret spec strength problem, but it matches his prot spec stamina problem. We find out that none of these formulas work. So it’s not our math that’s the problem. Reading on, Theck has walked us through an entirely incorrect hypothesis for the purpose of learning. Yay learning!

Hypothesis, Take Two

So, we’re at “A Change of Approach” in Theck’s post, and I don’t feel like retyping out that set of six formulas in nonLaTeX text. So go look.

Let’s bring up the picture of our total intellect with a stats buff on to refresh the numbers.


Our variables: B = 1038, G = 2378, match = 1.05, multiplier = 1.05.

  • T1 = 3765.195
  • T2 = 3765.395
  • T3 = 3765.195
  • T4 = 3765.745
  • T5 = 3764.25
  • T6 = 3765.00

In contrast to Theck, the warlock numbers can only rule out 1 formula — T5, which hilariously enough is the SimC formula for MoP, according to Theck. Theck’s retribution paladin strength problem rules out T2, T3, T4, and T6, so he decides to test T1 & T5.

To entertain Theck’s thought process, I looked at how he demonstrates that T5 can be off by 1 in Warlords stats. Our B times the match — 1038 x 1.05 — gives us 1089.9, which is similar to Theck’s made-up example of 919 x 1.05 = 964.95. Both numbers are super close to the next integer, which gives the one-off difference between flooring and not flooring.

Our T1 = 3765.195 and our T5 is one-off at 3764.25, so we already know that the more correct formula is T1:

  • CS_Total = multiplier * ( floor( G *match) + B*match )

Theck cleans up the formulas:

  • C = floor(G*match) + B*match
  • CS_Base = floor(B*match)
  • CS_Total = floor(C*multiplier)
  • CS_Bonus = CS_Total – CS_Base

A Case of Stamina

Just out of curiosity, though, how would a warlock test the T formulas if she didn’t have Theck’s post? Remember, we only ruled out 1 of the 6 formulas with the warlock numbers. You can’t just sit there and guess with that.

I figured I’d go back to stamina, like Theck did.

Let’s start at the beginning. Remember that we’re trolls.

  • B = 890 + o = 890
  • G = 3250


Theck still uses match = 1.05, because Stamina is what protection paladins get for their armor class match (retribution gets Strength). So he can use the same number for Stamina because it’s the same passive, just different spec & stats.

Warlocks can only get Intellect and thus can’t get a Stamina bonus through armor class, so our match can’t be 1.05. But we clearly have a modifier of some kind going on in our base stamina, because 979 definitely isn’t 890.

I hypothesize that it’s our Blood Pact passive. Let’s look at the tooltip (here’s the Wowhead double-check):


The 10% that goes with nearby party and raid members is the raid buff equivalent that warlocks bring. In Mists of Pandaria, our Dark Intent buff provides this along with the 10% spellpower buff, but in Warlords, Dark Intent is providing 10% spellpower with 5% Multistrike. Our old DI’s stamina portion is being rolled into our Blood Pact passive, which acts like an aura (like when a shaman provides mastery just by being a shaman). The other portion is our old Fel Armor passive.

This could go one of two ways. The base stamina could be counting both the Fel Armor portion and the raid buff portion together, since we will always have both parts simply for being warlocks. Or, the base stamina could just be counting the Fel Armor portion, and the raid buff is treated like a total stat modifier, much like the Kings/Mark of the Wild stats buff was treated.

Since I’ve already spent time on beta looking at warlock buffs, I have a strong feeling that it’s the latter case. If you click a beta cauldron in Shattrath, your scrolling combat text says “<Blood Pact> fades, <Stamina Buff>.” Similarly, if you click off the Stamina buff you get from the cauldron, you get text that says ” <Stamina Buff> fades, <Blood Pact>.” My stamina also does not change in either total, base, or bonus when I do these clicking on/off shenanigans. The buffs clearly share with & overwrite each other instead of stacking.

So let’s try this.

  • B = 890
  • G = 3250
  • Fel_Armor = 1.10
  • multiplier = 1.10

CS_Base = floor(B * Fel_Armor) = floor(890 * 1.10) = floor(979.00).  And what was our base tooltip again?



OK, time to test the multiplier effect. I don’t know if we have a high enough multiplier to prove the T functions like Theck did with the protection paladin, but we can try.

  • T1 = 5009.4
  • T2 = 5008.9
  • T3 = 5009.4
  • T4 = 5008.5
  • T5 = 5009.4
  • T6 = 5008

Ooof, not quite. For Stamina, T1, T3, and T5 could work. (They’re also all the same answer!) We already know that T5 doesn’t, since our Intellect didn’t work well with T5. So we’re down to T1 and T3.

The difference is whether you floor the G*match or whether you floor the B*match.

  • G*match = 3250*1.10 = 3575
  • B*match = 890*1.10 = 979

I think it has to do with how it’s a bit harder for me to get a non-integer number with the 1.10 multipliers. Both my B & G numbers are divisible by 10, so when you multiply them by 1.10, you’re going to get an integer again. So there’s no difference between flooring it or not flooring it.

We can’t change the B number, since it’s based on race and class, but we can change the G number, which is based on gear. I unequipped my helm (+331 stamina), which leaves me with 2919 stamina from gear.

  • B*match = 890*1.10 = 979
  • G*match = 2919*1.10 = 3210.9

This should be a good number to test, because it’s so close to the next integer that flooring G*match should make a difference versus not flooring.

  • T1 = 4607.9
  • T3 = 4608.69

Aaaand it does make a difference of, ding ding ding, one. Let’s check with the in-game tooltip when I have my helm removed:


Woot! T1 is confirmed to be the correct formula for warlocks, backing up Theck’s theory.

Until Next Time

While Theck goes on to deal with racial bonuses, food, and flasks, I think I’ll stop this “homework” post for now.

Besides, all of the racial cases he’s testing — tauren’s Endurance, draenei’s Heroic Presence, and pandaren’s Epicurean — are not applicable to warlocks because warlocks can’t be any of those races. I feel good enough reading through Theck’s post to broaden my knowledge of it, but I don’t think many (if any) of the warlock races have flat permanent bonuses like that until we get into secondary stats.

I suppose that’s my next “homework” — finding all the base values for all the warlock races as well as the warlock specs and branching out into theorycrafting the secondary stats. Then I can get into flasks and food, since most of the flasks and food in Warlords involve the secondaries, not Intellect.

I enjoyed my “homework,” so I hope you did, too, and maybe I’ll see you next time with more shemathigans.

TC101 Homework: Character Sheet Stat Calculations

Racials for warlocks in Warlords of Draenor

Out of the recent Developer Watercooler and the resulting Twitter hour of clarification by senior game designers Celestalon and Holinka, there was one of a few major topics: the reworking of all the racials. With hit and expertise going away in Warlords of Draenor (WoD), a lot of racials needed to be redesigned, which made a perfect opportunity to create new ones for some races to bring them in line with others.

While some of the proposed racials like the night elf’s day-and-night one are interesting, I’m going to focus on the races and racial properties that are warlock-applicable. While I could list out all the racials possible for each category, I feel I’ll list out the racials I feel either win the category or are most important to discuss.

New or changed racials are highlighted in fel green. The factions are also represented in colors: Alliance and Horde.

Removed racials

  • Gnomes had a 1% expertise racial to daggers or 1H swords in the main hand.
  • Humans had a 1% expertise racial to 1H swords in the main hand.
  • Orcs had a 1% expertise racial to fist weapons, which affected a wand in the main hand, oddly enough.
  • All the other expertise racials dealt with weapon types not usable by warlocks.

Expertise as a stat is gone in Warlords, so these racials are gone as well.

Perks: Spell & CC resistances

We’ll get to CC-breaks later. There’s a category of racials that provide some resistance to a certain spell school, and then there’s a couple races that provide a duration reduction for certain crowd control types.

  • Orc: 10% Stun reduced duration (Hardiness)
  • Troll: 15% movement-impaired reduced duration (Da Voodoo Shuffle)
  • No resistances: human, goblin
  • Spell resistances: Frost (dwarf), Arcane (gnome, blood elf), Nature (worgen), Shadow (worgen, undead)

With spell resistance as a stat gone from gear and flasks, etc., there’s nothing for a player to control spell resistance, so encounters aren’t designed around spell resistance being a deal-maker. But we can discuss it anyway.

Warlocks have Twilight Ward which provides a big absorb against shadow or holy damage. The current PvP 2-piece for warlocks also allows Twilight Ward to absorb from any magic school, so having a spell resistance perk is really just that, a perk. The “tanky warlock” stance changes Twilight Ward into Fury Ward, which absorbs all magic schools plus physical damage. When Twilight Ward doesn’t apply, we will use Unending Resolve or our health and self-healing in response.

The crowd control duration reductions don’t draw a lot of attention either, as everyone usually focused on the active ability CC-breaks which provide a complete escape from crowd control. But the usual PvP counter to a warlock is a burst, CC-crazy melee like a warrior. Reduced duration against tools a melee uses to lock a warlock down is a plus in PvP.

Perks: Profession bonuses

Races offer some kind of bonus to a profession, whether primary or secondary. Humans and trolls don’t have a profession bonus, but what they have falls under the category of “this is useful when you’re in the progress of doing something.”

  • None: Orc, Undead
  • Secondary profs: Archaeology (dwarf)
  • Primary gathering profs: Skinning (worgen)
  • Primary production profs: Alchemy (goblin), enchanting (blood elf), engineering (gnome)
  • Reputation gains (human)
  • 20% XP from killing beasts (troll)

The troll racial makes me wonder how they came up with that one. It replaces Beast Slaying’s extra damage done to beasts, so I see the theme of killing beasts, but, really? An XP bonus? I guess if you’re desperate, you can race-change to troll for leveling and then race-change back when you’re done.

The professional bonuses are obvious when leveling a profession, and perhaps are actual advantages for twinks, but at endgame once everything is leveled, there’s no real difference. If it matters to you, goblin, blood elf, or gnome are the winners, as warlocks in both PvP and PvE endgames tend to run double production professions. (If there were such a thing as a Tauren warlock, we could have an Herbalism bonus, whose haste through the active Lifeblood ability we might like.)

The only racial that really stands out for PvE endgame is the human reputation racial, because it makes grinding for rep a little bit easier, whether through killing things or doing quests.

Advantages: Passive stat boosts

Generally the passive stat increase is the main reason a race is picked as best in PvE. With an extra 1% of a favored stat, players can use different gear to further maximize another stat without losing effectiveness. To make up for missing passive expertise racials, many of the races got new passive stat boost racials.

Several tweets and blue posts — over especially the new night elf racial that switches its secondary stat bonus depending on the time of day — have said that stats will be a lot closer together in value for Warlords. All I can say is that we’ll see about that. I have to wonder how not having haste breakpoints for DoTs will go for haste’s priority. DK theorycrafter Magdalena brought up the point of possible readiness “breakpoints” where you can get some cooldowns to line up with other cooldowns or possibly cooldowns lining up with specific boss phases.

Haste, crit (aff/demo), and readiness fall into the category of stats you want to have “enough” of; mastery, crit (destro), and multistrike tend to be “stack ’em high” stats. Usually the best passive stat bonus racial goes to helping out the “enough of” stats, so that players can go for the gear that has more of the “stack ’em high” stats on.

  • 1% real haste: gnome, goblin
  • 1% crit: worgen
  • 2% crit damage: dwarf
  • 2 secondary stats: human
  • 2% pet damage: orc
  • Health regen: troll
  • Drain attack: undead

The goblin racial was clarified to be 1% real haste, as opposed to only cast speed; I assume the addition of the gnome haste racial would be the same. Real haste can affect other things like RPPM mechanics, so this is a subtle change with big DPS impacts.

There’s another subtle different between worgen and dwarves. Worgen give +1% crit chance, which lets you crit more often. Dwarves give +2% crit damage/healing, which lets you crit harder. The default meta gem for warlocks has been Burning [Whatever] Diamond, which had some intellect like numerous meta gems did, but the second effect was 3% increased crit damage, and that’s why we took that meta gem over others. Since haste can be a crappy stat for single-target destruction warlocks, they still even debate whether to use the old Burning meta gem or the legendary meta gem (LMG). Dwarves have a pretty good racial here.

Orc’s 2% pet damage is a boost to demonology, and troll and undead see limited benefits in PvP. Generally, if it was down to passive stat bonuses, goblin wins out for the Horde.

The new Human Spirit (no longer 3% Spirit) is going to be interesting. The biggest question I have is something Celestalon tweeted about; whether the racial will offer enough of each stat to make a competitive difference. If human doesn’t provide enough stat to make a difference, dwarf may win out as the default go-to race. The biggest benefit to human racial is a ridiculous amount of min/maxing such that you can effectively change your passive throughout a tier or maybe even fight to fight without needing to spend money on an entire race change.

If haste or crit is valued a lot, gnomes and worgen will beat out humans for desirability in the passive stat field. If mastery, multistrike, or readiness take control of stat priorities, human will be the go-to passive stat bonus racial. If priority is balanced between one of each bunch, then human may be competitive depending on how much it offers for each stat. “Pure-stat” races will probably be the “best” in the beginning of an expansion, but as gear gains more secondary stats, we see priorities shift more toward the stack-em-high stats, which include multistrike and mastery, neither of which is in a racial beyond human.

Advantages: Active abilities

The active abilities are often what makes or breaks a race beyond the passive stat boosts in PvP activities. Alliance active racials are usually defensive or reactive racials (though you can argue using Darkflight’s movement speed boost offensively like in flag carrying). Horde active racials are usually offensive or proactive racials.

Instead of going by faction, I put similar active abilities near each other.I realize there are other racials like undead’s Cannibalize, the worgen self-mount, or the goblin personal bank option, but those aren’t the racials that get brought up for endgame balance. Similarly, goblins have a jump-forward ability; if you really need a jump ability, you’re going to be a demonology warlock and Demonic Leap everywhere.

  • Dwarf: Self-dispel (poison, disease, bleed, magic, curse) + damage reduction + non-CC break (Stoneform, 2min CD)
  • Undead: Self-dispel (Charm, Fear, Sleep) + shares 30 sec CD with similar effects (Will of the Forsaken, 3min CD)
  • Gnome: CC-break for snares/roots (Escape Artist, 1min CD)
  • Human: CC-break for just about anything + shares CD with PvP trinkets (Every Man For Himself, 2min CD)
  • Orc: Spellpower increase for 15 sec (Blood Fury, 2min CD)
  • Troll: 15% haste for 10 sec (Berserking, 3min CD)
  • Worgen: 40% movement speed boost for 10 sec (Darkflight, 2min CD),
  • Blood Elf: Restore 3% mana + silences NPCs for 3sec (Arcane Torrent)
  • Goblin: Fire damage ability (Rocket Barrage, 2min CD)

The “miscellaneous” is pretty obvious: worgen (speed boost), blood elf (silence + resoure restore), and goblin (extra damage). Blood elf might be useful in PvP, but usually the worgen would be the best miscellaneous racial. In Cataclysm and Mists of Pandaria, worgen was typically the “best” Alliance race, and PvE/PvP sometimes showed that having an extra movement speed boost really helps.

Dwarf got an upgrade to Stoneform to remove just about any type of debuff, plus its usual damage reduction. However, the ability is not going to be a CC-break and will be unusable while CC-ed. The dwarf racial will be best for reacting to a damaging debuff in PvE raids or as a defensive proactive ability in general.

The CC-break abilities — undead, gnome, and human — all have a variety of three characteristics:

  • What CC it breaks: movement speed impairments, loss of control, or both
  • Whether it shares CD with any similar break abilities: No sharing, partial sharing, or complete CD sharing
  • Its own CD: short, medium, long
Undead Human Gnome
Breaks loss of control CC Breaks both movement & loss of control CC Breaks movement CC
Partial shared CD Completely shared CD No shared CD
Longest self CD (3min) Medium self CD (2min) Shortest self CD (1min)

Gnome will break only snares and roots, which leaves the warlock still able to cast both defensive and offensive spells. Roots are only dangerous to us when we try to peel from a melee, and we can use our CC to help us in that regard. Thus, the really short CD on a not amazing CC-break.

The human EMFH ability is still being looked at; primarily, its fate rests on how the PvP trinkets that share its CD get designed. I presume that the undead racial gets the larger CD because it only partially shares a CD with other effects, rather than completely locking out both options.

Finally, when it comes to straight-up throughput, the orc Blood Fury (spellpower) and troll Berserking (15% haste) are the winners, and not just for warlocks. Berserking might not be as big a deal with haste breakpoints gone, but it’s still a powerful haste boost. The biggest deal from these racials is that their cooldowns line up with many of our cooldowns so we can stack them for bigger spikes of damage. Typically, orc wins out in the beginning of the expansion where spellpower is a massive boost while we played the secondary stat breakpoint game, and troll wins out in the late expansion as we have enough secondary stat on our gear to support going balls-to-wall on stacking a particular stat.


Of course, the “winner” is completely subjective based on what a player cares about when picking a race. A player with pure performance in mind may choose a different race than someone who picks using the casting animations.

They took out the expertise racials and put in new passive stat racials, but the active abilities that are thrown about as OP didn’t get touched too much beyond Berserking getting nerfed a little. We’ll need to discuss passive stat racials when we get closer to simulating gearsets, but that’s quite a long ways off.

However, we shouldn’t get too excited about the passive stat bonuses, since supposedly the stat priorities will be a lot closer to each other than the ridiculous “hit X haste breakpoint and then just stack mastery” that we have right now. The passive stat racial is likely to affect your warlock to a minimal degree.

Dwarf maybe got improved with Stoneform, but as warlocks are a hale and hearty class anyway, I’m not sure it compares to EMFH’s get out of jail free card. Orc and troll are still very strong for PvE damage throughput. I’m not sure they’re done looking at the active abilities, but they’re still keeping to the factional theme: Alliance gets strong defensive abilities (Stoneform and EMFH) while the Horde gets strong offensive abilities (Blood Fury and Berserking).

The conclusion is that we’ll likely see the higher raiding warlocks stay Horde for the DPS output racials, but the Alliance warlocks won’t be completely SOL because of their newly reworked passive stats. In PvP, we might see more dwarf warlocks, but I imagine the same old trinket replacement argument will live on for human/undead.

Racials for warlocks in Warlords of Draenor