Priest Theorycrafting for 3.3 – Summary (Part 6)


Let us gather the various parts together and have a look at some of the numbers. This post also includes LootRanks and Pawn values.


  1. Introduction
  2. Mana regeneration including Spell Power and Stamina
  3. Mana pool
  4. Haste rating
  5. Critical rating
  6. Summary including LootRank and Pawn values

Summary of stat weights


  • Discipline = 1.00
  • Holy = 1.00


  • Discipline = 0.4272
    • MP5 value from mana regen = 0.4272
  • Holy = 1.0516
    • MP5 value from mana regen = 0.3818
    • MP5 value from Holy Concentration = 0.0924
    • MP5 value from Spiritual Guidance = 0.58


  • Discipline = 1.1551
    • MP5 value from mana regen = 0.1552
    • MP5 value from maximum mana = 0.1977
    • MP5 value from Replenishment = 0.1518
    • MP5 value from Shadowfiend = 0.0988
    • MP5 value from Rapture = 0.1581
    • MP5 value from critical healing = 0.1562
    • MP5 value from Divine Aegis = 0.1969
    • MP5 value from Inspiration = 0.0404
  • Holy = 0.8162
    • MP5 value from mana regen = 0.1632
    • MP5 value from maximum mana = 0.1719
    • MP5 value from Replenishment = 0.132
    • MP5 value from Shadowfiend = 0.0859
    • MP5 value from critical healing = 0.1443
    • MP5 value from Surge of Light = 0.0405
    • MP5 value from Holy Concentration = 0.0395
    • MP5 value from Inspiration = 0.0389

Spell Power

  • Discipline = 2.0
  • Holy = 2.0

Critical Rating

  • Discipline = 1.1315
    • MP5 value from critical healing = 0.4493
    • MP5 value from Divine Aegis = 0.5661
    • MP5 value from Inspiration = 0.1162
  • Holy = 0.7397
    • MP5 value from critical healing = 0.4772
    • MP5 value from Surge of Light = 0.1339
    • MP5 value from Inspiration = 0.1286


  • Discipline = 0.2
  • Holy = 0.2

Haste Rating

  • Discipline = 0.6309
    • MP5 value from hasted healing = 0.6309
  • Holy = 0.6227
    • MP5 value from hasted healing = 0.6096
Source Discipline Holy
MP5 1.0000 1.0000
Spirit 0.4273 1.0516
Intellect 1.1551 0.8163
SP 2.0000 2.0000
Crit rating 1.1315 0.7387
Stamina 0.2000 0.2000
Haste rating 0.6309 0.6096

Gear lists

Below are a pair of links to LootRank gear lists which are a good place to start with looking at the application of these stat weights.

Pawn values

Pawn values for these are:

( Pawn: v1: "Discipline33": Stamina=0.2, Intellect=1.1551, Spirit=0.4273, CritRating=1.1315, MetaSocket=104.257, HasteRating=0.6309, SpellPower=2, Mp5=1)

( Pawn: v1: "Holy33": Stamina=0.2, Intellect=0.8162, Spirit=1.0516, CritRating=0.7397, MetaSocket=97.1392, HasteRating=0.6096, SpellPower=2, Mp5=1)


Spell power is still king for both specs. The hunt for spell power, especially when gemming and enchanting will continue.

For discipline spec priests critical rating and intellect continue to be primary stats. Although critical rating vs haste rating is a personal preference critical rating outshines haste rating completely on a theorycrafting basis. Even in an unlimited mana scenario (which is actually feasible at T9.5), where discipline haste rating rises to 1.1621, ciritcal rating (1.3610) provides better value for itemisation. Spirit continues to be the poor cousin to intellect.

Edit: For an interesting read regarding crit vs haste check out Zusterke’s PlusHeal post on the subject.

For holy spec priests the stats are more evenly balanced. Spirit, intellect, critical rating and haste rating (in an unlimited mana scenario – 1.2344) are all evenly valued. In my experience holy priests continue to have significantly more mana issues than discipline priests decreasing the value of haste significantly.

How much MP5 are abilities worth a T9.5 gearing level?

  • Replenishment about 172-237 MP5
  • Shadowfiend about 129-155 MP5
  • Rapture about 247 MP5
  • Holy Concentration about 100 MP5
  • Divine Aegis about 909 MP5 at 35% crit
  • Surge of light about 215 MP5 at 35% crit
  • Inspiration is about 186-206 MP% at 35% crit

Thanks for reading and your constructive comments.

Gobble gobble.


71 Responses to Priest Theorycrafting for 3.3 – Summary (Part 6)

  1. Cínna says:

    Thank you for all the hardwork you put into this, its invaluable for people like me with poor math skills. Your time and effort is greatly appreciated 🙂

  2. Piru says:

    Thanks! I always look forward to your lists, though it doesn’t seem to have changed much since 3.2 in priorities.

  3. BobTurkey says:

    @Piru – priest mechanics haven’t changed much so the stats ony change slightly. 4.0 will be comletely different though.

    Gobble gobble.

  4. Virika says:

    Great work, Thanks Bob!

  5. Arrelliana says:

    I myself was really debating this issue over the past few days while going over the new fights and gear for ICC. I felt like I needed more throughput to hit the damage coming out. I came to the point I was trying to figure out how to get more burst. Crit is a great stat to make those spells hit much harder. I am gearing of course towards hard modes as my guild is very progression oriented having downed all content but yogg 0 (on phase 3). Mana doesn’t seem to be the issue as much as raw healing power.

    Ty for all the great work and I will be looking at the gear with these new stat weights and will forward any feedback I have. Looks like if you follow this line of thinking all the crit on gear in 3.3 is yummy.

  6. Thomas Jespersen says:

    Thanks! Note that the gem-system has become better in Pawn lately. It will derive scores for gems based on the other scales. If you look at the gem tab in the Pawn interface you can select which quality gems to use and it will show which gems is suggested

  7. Steve says:

    Thanks Bob 🙂 So how does lootrank take the set bonus into account? Now that you’re showing the emblem gloves as BiS, but you also show the dps T10 better than the healing T10, so clearly that can’t be accounting for the bonus?

    I’m sitting on 75 frost, waiting for 2 pc T10. I think I might start with shoulders/legs, but now I am thinking about shoulders/chest and hope for legs drop in VoA by then.

    Then at some point picking up Volde’s and False Gestures. Need many frost 🙂

  8. Adam says:

    Hey Bob,

    I’m still a little confused – haste value has been cut in half (and then some) from your 3.2 numbers, whereas crit has nearly doubled. Has there been some change in the underlying calculations? I don’t quite understand how they can switch positions so drastically, especially when you consider (even during 8 minute fights) the non-issue mana is. On a purely throughput basis haste always wins, even after typical lag is taken into account.

    Has the increase in fight length in the 3.3 numbers changed things that radically?

  9. tuloa says:

    I am new to your blog and hate the fact it has taken this long to find this. Thank you for your theorycrafting, it’s not something i’m personally very good at but it helps a hell of a lot in what to gem for and prioritys.

    Interesting how crit appears to be superior to haste and even MP5. I had recently started thinking about stacking haste to get those heals out faster but in ICC there doesn’t seem to be all that much fast damage.

    Thanks again!

    From a holy priest

  10. Tunga says:

    The crit numbers are so huge because you calculated them based on spamming FH non-stop. But nobody is doing that so the results are meaningless in any real situation.

  11. Arrelliana says:

    The numbers would support almost any spell you would use flash heal is just used as a base spell to get the relative weights of the stats compared to each other. Each spell uses a base amount of mana and has a cast time or cooldown. ( a cooldown is also a cast time even if it is gcd since it is a period in which you can’t recast the spell). Pretty much every casting class uses similar modifiers like HEP (Healing Equivalency Points or compares to spellpower.

    What the calculations are used for is simply how much HPS for a set amount of mana can I put out. If you increase spellpower the HPS goes up. If you increase haste the HPS goes up and so does mana usage. If you increase crit the HPS goes up but mana usage does not. If you increase your mana pool through int, the amount of time you can do the said hps goes up and the amount slightly based on int giving crit. ETC all of values are interrelated.

    Bob’s guide is a way to value the stats on different items to be able to compare a piece with all crit/haste to an item with just spellpower or just crit etc. They are not set in stone that if you put on all the “BiS” items you will heal the most. Bob and I both advocate balance and the gear that is out there seems to have a mix of stats so you will always have all the stats we need.

    But Tunga, specifically what he is showing is point for point crit gives you more HPS than haste at this gearing level. The numbers were different @ t7, t8, and t9 if you look at the same math he used there. Talents also modify the value of crit if you didn’t take talents like inspiration or holy concentration or divine aegis (which we would never do but).

    Your playstyle also will affect what stats you prioritize. Disc generally favors crit and holy haste. Holy wants the faster casts to push out the spells. But there is an argument that a crit loaded holy priest on a lull burst lull fight will be better suited than fast smaller heals. If you run with a group of aggressive healers you better be quick and crit would hit your instant cast spells harder than haste in that situation. Haste also has no effect on circle of healing and Prayer of Mending which are 2 of the top spells we use.

    So look at all of the information and if you want to weigh your stats and like the logic use the numbers. Compare the gear and make informed decisions. But the results are meaningful if you understand what they are being used for and blanketly dismissing them is a fallacy in and of itself. Your point would be better to disagree or post opposing views.

  12. Adam says:

    @ Arrelliana

    Per percent, haste is always more throughput than crit. This is because an increase in 1% haste means you did a full extra heal over 100 spells, whereas in increase in 1% crit means you did .5 to .8 of a heal cast (depending on talents). You are correct to say PoM and CoH benefit more from crit than haste, but incorrect to say that they do not benefit from haste at all. These spells should be treated as spells with a 1.5 second cast, since that is the length of a non-hasted GCD. Every point of haste that lowers this GCD is useful for healing throughput for any spell.

    It’s difficult for me to make a decision on these stat weights, mainly because they have changed so drastically from 3.2 numbers for unclear reasons. Another difficulty is that I switch from holy to disc so often that it’s hard to know which stat weights to focus on :).

    Thanks so much for your hard work, Bob, it’s excellent to have a resource like you in the priest realm. Keep up the awesome work! Is there any chance you could email me the spreadsheet you used so I may check it out?

  13. BobTurkey says:

    Thanks for the comments all.

    Thanks Thomas Jespersen. I’ll have to have a look at it.

    @Steve – No LootRank/GuildOx doesn’t include set bonuses as far as i’m aware. I don’t think the bonus for T10 will be that great so the DPS gear might still be on top.

    I’ll be buying the head for my first tier piece and shoulders second I think. More on that soon.

    @Adam – Haste value has been reduced because I wasn’t including the additional mana cost of the spells haste allows you to cast. This is described in the haste post (i’ll cross link them when i get a chance).

    Haste doesn’t always win on a throughput basis, well not for dsicipline anyway. Your example works if you ignore overheal, but doesn’t otherwise.

    No the increase in fight length from 6 mins to 8 mins has slightly decreased the value of intellect mainly.

    If you read the preceeding articles about the calculations or each stat you will see in detail how it was done. The calculations actually ahve barely changed at all. The most significant was haste and that follows extended discussion on Elitest Jerks.

    Nice comment Arrelliana. I couldn’t have put it better myself.

    Gobble gobble.

  14. Steve says:


    Am looking forward to hearing why you don’t think the bonuses will be all that great. I think a critted flash heal of 9K giving a free 3K hot will be very nice for those of us who are struggling to two heal some of the harder fights.

    I do flash heal more than other disc priests I bet, but it is my play style. I penance to get Grace up and for any big heals, but otherwise I have shield, pom and a flash heal coming in (canceling if needed) esp on marrowgar 25 with those large hits when a tank is not quite properly positioned.

    My concern about theory crafting is that it looks at best-case scenario, and not the real world for most of us.

  15. John says:

    besides the mana restore meta. for more output, is the 25SP+2% int or the 11mana+3% increased critical healing better now, that the critical values shifted.

    i find more updated itemlist on wowhead as on lootrank. maybe link gearlist on that. like example this:;minrl=80;maxrl=80;ub=5;gb=1;gm=4;wt=123:96:24:61:23:103:22;wtv=2:1.0547:1.0516:1:0.9114:0.6096:0.2

  16. Adam says:

    “Haste doesn’t always win on a throughput basis, well not for dsicipline anyway. Your example works if you ignore overheal, but doesn’t otherwise.”

    I don’t understand this statement. A crit heal is more likely to overheal than a hasted non crit heal, and even with DA, you have to assume not all the shields will be absorbs, and you are still only at 80% of haste’s value. I am also assuming that your calculations take into account that it easier to get 1% haste than it is to get 1% crit from the standpoint of itemization points. If they do not, then this is another reason haste is the stronger stat (for throughput).

    I see your argument from a mana usage perspective, but I really don’t see that being meaningful at this level of content. The regen we get from gear alone should be enough, and everything else should be viewed from a pure throughput perspective, IMHO. Now, this doesn’t mean I get 1,000 haste and 0% crit, but it does mean that I would value haste slightly better than crit when it comes down to “that item vs this item” comparisons.

    Is the EJ thread you are referring to the updated 3.3 priest healing compendium?

  17. Arrelliana says:

    The one thing about haste is it does seem better for holy than discipline. The one thing haste does get hurt by is latency. I think the main argument would be on offset pieces or the 5th piece of non-set gear you pick up.. Do you grab a helm with crit or haste if you are using the other 4 pieces of tier.. or do you replace tier gloves with a crit piece instead of haste. I generally go for the haste+crit pieces without spirit if I can get my hands on them. I generally always switch specs depending on which spec favors the fight and I rarely get the luxury to fully itemize both specs so I go with balance.

  18. Adam says:

    Point well taken, Arrelliana; however, I rarely notice latency affecting my ability to chain casts together. As long as something is being spammed, latency is a non issue, in my experience (granted, I have a 40-50ms ping most of the time).

    Interestingly, raid healing (my primary role) as disc when 50%+ of heals are PWS benefits very little from haste OR crit.

  19. Steve says:

    @Adam: A fair amount of my direct heals from from the glyphed PW:S, which I think crits.

  20. Adam says:


    Glyph of PWS is generally something like 4-5% of my healing, not really much to argue that it benefits a whole lot from crit. Now, if PWS could crit…drool :).

  21. Arrelliana says:

    Yea pw:shield once the first one procs borrowed time haste is a non issue and is only affected by your raw spellpower then. I think haste is more effective in steady damage fights or high damage fights..but lull burst lull fights crit is probably a better throughput stat.

  22. BobTurkey says:

    @John – I have no plans to change from the Earthsiege meta ATM.

    I’ll look at the WowHead links. I think GuildOx (formerly known as LootRank) use the

    WowHead database as their data source. So it should be the same.

    @Adam – Elitest Jerk’s thread is here:

    Gobble gobble.

  23. Faust says:

    Thanks, man. This information is invaluable.

  24. Rosslyn says:

    Is there a typo for MP5 = 2.0 at the beginning of the summary?

  25. k07n says:

    Thanks Bob! ❤

  26. Faust says:

    Rosslyn is right. There’s a discrepance between the MP5 values at the beggining of the sumary and the ones in final table, right before “gear lists”.

  27. BobTurkey says:

    Yes that was a typo. Fixed.

  28. Grouikette says:

    Once again, for such a great job accomplished, Bob, thanks a lot, your work is great.

    However, at least for Disc spec, I must say that it would be better to switch to the Ember Skyflare Diamond as the best metagem.
    The IED is worth 105 points according to your theorycrafting, while ESD is worth 95-100 points at the 3.2 BiS gear level (and ESD>IED if int>3200 if I didn’t do a mistake).
    But the important point is that >50 of the 100 points of the ESD are throughput points, while only 10 of the 105 points of the IED are throughput points.
    So we have two metagems that worth almost the same points, but one of them (ESD) is clearly better for hps.

    And since mana is NOT an issue for disc priest for now (provided you have a correct raid composition), 85 points of mana regen proccs of the IED are not so useful (if not useless).
    So if mana is not an issue (which should be the case for 99% of us), I really think that ESD has became superior to IED since the 3.3 (and maybe before, but let’s skip that)

    As a side note, Bob, I think that if you could create a “hps only” lootrank, it would be really useful for priest that do not have mana issues and seek hps-oriented gear.
    I think that a great idea would be to determine the max mana regen (and mana pool) we need to heal through the hardest encounter (for example Valkyr 25 hero), and then calculate if our avatar has reached this value, and then use the “hps only” lootrank for further gearing. Can we create a “mana pool+regen cap” like dps have a hit cap ?

    (and accept my apologies for the writing, this is my best english ^^)

  29. BobTurkey says:

    @Grouikette – IED vs ESD depends if you value throughput or regen more. I wrote these stat weights for balance so I still go with the IED with the higher stat weight total. I still get low on mana on some fights, especially if i’m using PoH a lot. Mainly hard mode fights or with reduced numbers of healers.

    “Can we create a “mana pool+regen cap” like dps have a hit cap ?” This is an interesting question. You probably could, although you’d have to make assumptions about raid composition and fight characteristics (where the damage to the raid happens and how fast). If you were going to do this you would be best to base it off the hardest fight (healing-wise) that yo could think off and work from there.

    A combined mana pool/regen cap would exist, it would just be a matter of figuring out what it is.

    Gobble gobble.

  30. […] — new 3.3 theorycrafting from BobTurkey– Part 4: Haste Rating Part5: Critical Rating Part 6: Summary Hunters — Jaedia’s ICC raiding loot list.  Warlock–Oath’s 3.3 Wish List.  […]

  31. […] Bob Turkey vient de terminer une série de 6 articles de theorycrafting sur le poids des caractéristiques pour les 2 spécialisations heal : […]

  32. Steve says:

    @Bob: Any new thoughts on the T10 bonuses for priests?

    I just picked up my 245 T9 pants, so I am thinking I can swap out the 232 T9 hands for the Gloves of False Gestures and forego the T10 bonus for awhile, if it turns out to not be all that great for disc.

  33. barnacletubber says:

    Thanks for this latest work. Ultimately, WoW is a game of increments, and although the healing Priest changes at gear level 245 are not as great as those at level 226, the adjustments are important — perhaps nowhere more than in mindset. I want to respond with my thoughts to the following comment you made In Part 1 of your summary:

    “Healing priests can choose to just aim to gear at the highest item level possible and provided they choose appropriate gear (stuff with spell power, intellect, spirit, haste, etc on it) they would be competitive with a similarly skilled and experienced priest who selected gear based on these stat weights. This is a bit of a slap in the face for someone who has spent hours fiddling with spreadsheets, but at the end of the day stat weights are about fine-tuning gear selection and not the be-all-and-end-all of gearing. “

    So why bother with theorycrafting at all? – Because it represents the science behind the game, the “why.” It is the vocabulary without which you cannot have the thoughts; those who have been exposed to it are thinking in terms that those who have not simply cannot. It provides the knowledge that leads a player to develop a deeper general understanding and intuition about the game and character management, which in turn allows for the highest level of fine-tuning.

    The more I have learned about and paid attention to your theorycrafting, the better my understanding and intuition has become, and the less I have to constantly rely on your theorycrafting!

    If it’s not the be-all-and-end-all, it is most certainly the means to an end. Without your work, I wouldn’t have the understanding that I do. So I thank you, once again!

  34. BobTurkey says:

    @Steve – Good for tank healers, pretty average for everyone else. I only tank heal about 10% of the time so currently I have no interest in the T10 set bonuses.

    Actually i’m more likely to get the T10 shadow gear and heal in that. 😉 I spend 50% of my time DPSing ATM anyway.

    @barnacletubber – I like numbers and spreadsheets? 🙂

    But yes you are right. Learning how things works makes you better at it, but you can survive just fine without knowing how it works.

    You are welcome. Much of the stuff I post up here is stuff that I’m think about for myself anyway. Writing it up here is just a convenient place to keep it and it also allows me to share I with others who might find it useful.

    I’m glad you have.

    Gobble gobble.

  35. Arrelliana says:

    Not sure where to post this but wanted to start getting some community feedback as these are some fresh numbers on 2 set bonus…

    The hot does refresh and does continue to tick.. it also builds a small amount the inital tick was going around 600 up to 2-3k. I haven’t got to get parses together to see the effect for sure and understand the mechanic but it continually ticks.

    The real question is how effective will this be. On my initial recount parses for 10 man icc with trash and lord was between 1-2% of my effective healing with a high overheal component. If you want to see the proc it is called Blessed Healing and the icon is the same as flash heal. But if you have the badges buy the shoulders and hands equip them and spam yourself if it isn’t what you want sell it back to the vendor. I am witholding judgement on the 2 piece bonus atm as I think the stats are worthy of serious consideration as you move to 264-277. (And I am a fan of tier looking the part) But I regularly play disc and holy so I keep my gear very balanced.

    Few more numbers (all of these are just trash and lord marrowgar (holy on trash & disc for marrowgar (healing with a paladin)

    Effective Healing 30 ticks @ 26365 total
    Min tick 572 Avg tick 879 Max tick 2135

    Overhealing amount 40639

    I can’t compare to flash heal yet till my parse is posted to see the actual % that it was proccing or the amount compared to the flash heals.

    I hope this helps.

  36. Vamptessa says:

    @Arrelliana – Interesting data! It would be interesting to see on Saurfang, where I end up flash healing a fair amount it seems, but still mainly shield/pom/penance/renew.

    My numbers broke down to 1.3 million in heals (including guessed absorbs)

    43 Penances (25%), 36 PoMs (21%), 29 flash heals (20%), 63 shields (saved mainly for boiling blood) (15%) and 40 renews. (9%)

    While the druid healed for 1.7m total, he over healed for 63% to my 10%.

    Only 10 of those would put the hot up. It doesn’t seem like all that big a deal, does it? However, if that hot lands on a marked person, it sure might make the difference.

  37. Arrelliana says:

    I agree with you and the mana free hot is not a bad thing I just see this as being a very small percentage of our healing once the parses break down. To me it is gonna shake out to a 1-2% boost. Maybe this goes up as you spam flash more..but rarely will we be in that cycle unless we are dedicated to a tank..once you start throwing the flashes around the stacking effect will be a no issue. But it does appear that it follows the philosophy as the set bonus is just that. A bonus..not something you gear directly for. Also the stats on the gear kinda support this as they are comparable across all of the comes down more to which stat you want to favor//haste vs. crit vs. regen.

  38. Arrelliana says:

    quick math here on your numbers…
    10 30 ticks of 650 each avg…so a boost of 19500 healing.. your total was 1.3 million so this would have worked out to .02% of your total healing with no overheal at all.

    Those numbers make me sad panda.

  39. Steve says:

    @bob – A number of people are reporting < 2% for 2pc t10.

    So, what do we do about regen? I'm building a spreadsheet in Apple's Numbers which lets me compare raw stats on 3 sets, and then calcs out of the crit, haste, etc. My mp5 modeling is wrong atm.

    It currently looks like some pure upgrades are in fact downgrades. I can't figure out what two pieces of T10 healing I'd take and then I guess forego spirit/regen on all of the other pieces?

    Meteor Chaser's Raiment is fairly attractive given that the only better chests require heroic 25 trophies, correct?

  40. Hanthion says:


    You said…

    “I think that a great idea would be to determine the max mana regen (and mana pool) we need to heal through the hardest encounter (for example Valkyr 25 hero), and then calculate if our avatar has reached this value, and then use the “hps only” lootrank for further gearing. Can we create a “mana pool+regen cap” like dps have a hit cap ?”

    That’s something I’ve considered very much…

    The variables involved would be absolutely colossal but it would be really neat to see some kind of mana cap.

    Every time I get a new piece of gear I try to test out my mana regen capabilities to make sure I’m not ending up w/ a ton of mana at the end of fights. I’ve been using Lord Marrowgar in ICC25 and Twin Valkyrs TotC25 as my base references (for me these are the most healing intense fights I’ve encountered). I’m sure Twin Valkyrs in ToGC25 would be a great reference as well but I’ve never even stepped foot inside ToGC25.

    Anyway, I make sure to run these bosses every week and pay very particular attention to how close I get to running out of mana. Then, by using the wonderful information from Bob’s blog, I decide whether I need more regen or more throughput. If I run out of mana early, I’ll socket gems and/or swap trinkets for more regen. If the fight ends having never used a mana potion or hymn of hope then I’ll socket gems for throughput.

    My point anyway, is that I don’t think you can really do anything more than that. I really don’t think it would be worth the time and effort to figure out some sort of mana cap because there is just waaaaay too much involved.

    I’m positive that it can be done…but the solution will ultimately be scenario specific. And then you’d have to repeat the process for the other thirty gazillion raid setup scenarios that could occur.

    You just have to find your niche in your raid group, and then from there determine what kind of stat setup they need from you. If your raid group seems to take a lot of small sustained damage, you might wanna be more mana regen oriented because you’ll be casting a lot more spells. If they take a lot of high burst damage, then you might wanna be more throughput oriented.

    Do you run w/ a disc priest? If so, disc priests mitigate a lot of burst using bubbles and other cool stuff so your throughput stats can become less valuable.

    Do you have a druid tank? Druid tanks rely a lot on having a huge health pool at the cost of having low mitigation causing them to get hit really hard so your throughput stats can become extremely valuable.

    The list goes on…and on…and on…

    This information on Bob’s blog is just a guideline, it’s really up to you to determine how to use them most effectively.

  41. Hanthion says:


    I noticed that the above EP weights you’ve used to determine your gear lists are the same throughput EP weights shown for your trinkets in 3.3.

    This leads me to believe that you’ve determined a more throughput based stat setup will take precedence over a mana regen based stat setup in ICC.

    If I’m correct in my assumption then I’m wondering how you came to this conclusion. I’m not trying to say that you’re being biased in your EP weights, but I’m wondering if you’re anticipating certain scenarios in ICC25 that would require a more throughput oriented stat setup.

    Above you said…

    “No the increase in fight length from 6 mins to 8 mins has slightly decreased the value of intellect mainly.”

    This confuses me. Doesn’t an increase in fight length mean that mana regen stats become more valuable since you need more mana to heal for two extra minutes?

    I guess what I’m trying to say is that after reading your blog today, my whole world is turning upside down. I was expecting to see more mana regen required for ICC since the fights are longer but I guess I’m just confused now…

  42. BobTurkey says:

    Thanks for the 2pcT10 information. Its about as bad as I thought it might be 😦

    @Steve – I’m not sure what you mean by this, “So, what do we do about regen?”

    Meteor Chaser’s Raiment is very nice. When I get a chance i’ll get the gear lists up and my personal list of target upgrades.

    @Hanthion – For Lord Marrowgar in ICC25 get your raid to stack under his butt. Its actually a very easy fight to heal.

    The trinkets post has two sets of stat weights used. These from this post and a set with the SP:MP5 ratio set at 0.6.

    Longer fights mean that base mana pool and Shadow Fiend are less effective (Assuming same number of uses of SF). This means their contribution to stat weights decreases. Regen components like Rapture work the same regardless if the fight is 6 or 8 minutes.

    I’m finding no more mana regen required in the first four bosses of ICC. Its if anything less intense than ToC, a little. I assume it will get harder (although who knows with the normal version).

    Gobble gobble.

  43. Vamptessa says:

    @Bob – If we went all BiS, we’d have to stack int gems like crazy to make up for the lost in mana regen, no?

    I think I am going shoulders/pants and just saying screw it, this min/max is no fun 🙂

  44. BobTurkey says:

    No. There is heaps of int on the BIS pieces already. If as disc you went all BIS you might have low spirit, but I don’t think it would an issue as you’d have a pretty huge mana pool.

  45. […] list uses the stat weights that Bobturkey has come up with.  He has solid math behind it and I very much liked the initial stat weights he […]

  46. Zusterke says:

    I’m a bit puzzled that your introduction seems to hint at haste: “Healing priests can choose to just aim to gear at the highest item level possible and provided they choose appropriate gear (stuff with spell power, intellect, spirit, haste, etc on it)”
    but in your stat weights crit beats haste by a mile for holy priests. I did a lot of work around crit for holy priests myself lately and there is a rather steep diminishing return. The diminishing return starts as early as 20% crit but it only weighs heavily beyond 30% crit. I fear that getting more crit than that (raidbuffed) would not be a good strategy for Holy Priests.
    For Disc priests, it would be quite good, ofc.

  47. Alhena says:

    Question about the gems… according solely to the pawn numbers, Potent Ametrine is better than Luminous Ametrine for a Holy spec (even the spirit one is better)

    Is there any consideration for choosing INT in that case all things being equal?

  48. BobTurkey says:

    @Zusterke – Yes. I feel that once you get to about 30% unbuffed crit adding any more starts to have a reduced impact.

    The ultimate solution would be something like the shaman HEP programs written to calculate stat weights specifically for your toon as they are geared at that point-in-time. Unfortunantly I don’t have the interest or the time to do so.

    These stat weights are a set of generalised weightings to give priests an idea of whats good for them and whats not. It doesn’t take into account things like haste soft caps or the diminshing returns of high crit values.

    However to be perfectly honest, at teir 10, you can pretty much gear however you like. All priests will have resonable mana regen, haste values and crit values and its more a case of adding more whatever you like.

    Personally I am completely comfortable healing T10 in T9 gear. I’ve got everything I want (mana regen/haste/crit/SP) already and T10 is just adding a bit more of each.

    @Alhena – I haven’t looked how these stat weights effect gems in 3.3 yet. As I suggested to Zusterke above, you can pretty much choose whichever stats you feel you need most. Or you can just stick with +23 SP gems in every socket.

    Gobble gobble.

  49. Avalonna says:

    Thanks again Bob – You’ve saved me alot of time with your great work again. This will come in handy with some posts I’m writing, and I’ve been putting off for lack of time to do something similar.

  50. Zusterke says:

    Bob, I certainly won’t waylay your point about T9 providing ‘enough of everything’. It isn’t the first tier we had comparable gear lists/weights even though the underlying assumptions and work is different 🙂
    I guess it proves you statement that weights and gearlists are more about refinement than about getting decent gear.

    I’m still amazed at your results with Haste though. Where I have been blinded by not bringing its manacost and throughput together… this work clearly indicates the downside of the associated manacost.

  51. BobTurkey says:

    I didn’t include mana cost either until there was discussion about it on Elitest Jerks and it got me thinking that I needed to include it.

  52. Karrah says:

    I find it interesting to break up your stat weights into a regen part and a throughtput part so one can easily get a new set of weights for an infinite mana scenario (which happens to be the case for me in 95% of the fights). For a holy priest, this leads to:
    Throughput Regen
    MP5 1
    Spi 0.58 0.47
    Int 0.18 0.63
    SP 2
    Crit 0.61 0.13
    Haste 1.23 -0.62

    Maybe you should include that break-up into your next theorycrafting post because that gives a much finer understanding of what’s happening with your model: if you need more regen focus on MP5, spirit and intell and if you never face any mana shortage just go straight for SP, haste and crit. Those are already well-known points but the numbers prove that as long as you don’t care about mana, haste is definitely better than crit for a holy priest, a point that was a bit confusing for some people when they saw our global stat weights.

  53. Karrah says:

    Once again the throughput/regen break-up for a holy priest as it’s a bit hard to get a table-like style in a comment:
    stat throughput / regen
    MP5 0 / 1
    Spi 0.58 / 0.47
    Int 0.18 / 0.63
    SP 2 / 0
    Crit 0.61 / 0.13
    Haste 1.23 / -0.62

  54. Thorrion says:

    Hi, Bob! I’d like to thank you for the great contribution you make to the priest community! I find your disc priest theorycrafts to be amongst the best if not the best out there. However, I’ve found a serious flaw in your calculations that leads to biased throughput stat weights. Let’s look at the Critical Rating calculations.

    Flash Heal Crit MP5 value = (base heal + (SP * SP co-efficient)) * SP MP5 value * proportion from crit * (1 – over-heal %)

    1. You assume 1.0 SP to be worth 2.0 MP5.
    2. You calculate the effect of increasing Crit Rrating on the spell output (i.e. “healing done”).
    3. You multiply the calculated healing by SP-to-MP5 ratio and take the result as the MP5 weight of the Crit Rating.

    There’s nothing wrong here. However, instead of using the same approach to calculate the effect of increasing Spell Power on the spell healing output, you just assume that SP = 1.0 (or 2.0 in terms of MP5 budget). That means that you have your Crit and Haste ratings measured on a different scale than Spell Power. As effect of Spell Power on Flash heal output is less than 1.0, this leads to Crit and Haste being underestimated.

    Now I’m working on the disc priest stat simulator, and from my preliminary tests, for roughly the same gear level as in your calculations, I have the following weights:
    Stat Weights (Regen):
    Stamina: 0.00
    Intellect: 0.82
    Spirit: 0.44
    MP5: 1.00
    Haste: -0.54
    Crit: 0.00
    Spell Power: 0.00

    Stat Weights (T-put):
    Stamina: 0.00
    Intellect: 0.14
    Spirit: 0.00
    MP5: 0.00
    Haste: 0.76
    Crit: 0.39
    Spell Power: 0.55

    Stat Weights (Weighed Throughput/Regeneration = 2.0):
    Stamina: 0.00
    Intellect: 1.31
    Spirit: 0.44
    MP5: 1.00
    Haste: 2.22
    Crit: 1.40
    Spell Power: 2.00

    I hope that will help you and everyone who’s interested. I’m open for discussion and will check this post regularly.

    /Thorrion @ Dunemaul-EU

  55. BobTurkey says:

    Good to see others taking an interest in theorycrafting. That post on Simcraft seems to have stirred some interest I see.

    Actually i’m using 2 spell power = 1 MP5.


    Flash Heal Crit MP5 value = (base heal + (SP * SP co-efficient)) * SP MP5 value * proportion from crit * (1 – over-heal %)

    The “SP MP5 value” part is where it gets multiplied by the spell power value of 2.0.

    How did you calculate your “Stat Weights (Weighed Throughput/Regeneration = 2.0)”? I can’t figure it out.

    Gobble gobble.

  56. Thorrion says:

    Thanks for a swift response!
    Actualy, I’ve googled your blog back when you did the 3.2 disc priest theorycraft. Since then I’m reading the rss feed.

    Ok, back to numbers. I guess we’re talking about the same things but in different terms.

    “Actually i’m using 2 spell power = 1 MP5.”

    Look, if you say 2 SP = 1 MP5, then dividing by 2 we get 1 SP = 0.5 MP5.
    It’s just a matter of what you mean by “equals”.

    Let’s just say that you value throughput twice as much as mana regen, right?
    I just find it more natural to say that 1 SP = 2 MP5 in such case.

    Flash Heal Crit MP5 value = (base heal + (SP * SP co-efficient)) * SP MP5 value * proportion from crit * (1 – over-heal %)

    The “SP MP5 value” part is where it gets multiplied by the spell power value of 2.0.

    Yeah, I know, that’s exactly what I was talking about in points 2 and 3.

    Again, in my original comment I haven’t said that it’s wrong. I just mentioned it to refresh the base assumptions.

    What I’m trying to point you to is that you are measuring crit and haste ratings on a different scale than spell power and thus they are incomparable. Let me try to explain this in couple of different ways.

    First, let’s look at this from the pure math point of view. Here’s how we can calculate the basic throughput (healing per second). We won’t look into secondary effects such as Divine Aegis, Inspiration, etc, and also don’t account for overheal.

    SP – spell power, CR – crit rating, HR – haste rating.
    Base_Heal = 1887, Base_Cast = 1.5, SP_Coeff = 0.8068

    Healing = Base_Heal + (SP * SP_coeff);
    Crit_Healing = 1.5 * Healing;
    Crit_Chance = CR / 4591;
    Combined_Healing = (1.0 – Crit_Chance) * Healing + Crit_Chance * Crit_Healing = Healing * (1.0 + 0.5 * Crit_Chance) = (Base_Heal + (SP * SP_coeff)) * (1.0 + CR / 9182)
    Haste = HR/3279
    Cast_Time = Base_Cast / (1.0 + Haste) = Base_Cast / (1.0 + HR/3279)

    HPS = Combined_Healing / Cast_Time = (Base_Heal + (SP * SP_coeff)) * (1.0 + CR / 9182) / (Base_Cast / (1.0 + HR/3279))

    CR HR
    Base_Heal + (SP * SP_coeff) * ( 1.0 + —— ) * ( 1.0 + ——)
    9182 3279
    HPS = ———————————————————————

    [hope that you can see the improvised formula drawing!]

    We can see that HPS = HPS(SP, CR, HR).
    To calculate the effect of each of the stat on the throughput, we calculate the derivatives:

    d(HPS) d(HPS) d(HPS)
    —— , ———, ——–
    d(SP) d(CR) d(HR)

    Which are going to be the weights of the corresponding stats.

    While you do calculate the derivatives for CR and HR (though in terms of healing, not the HPS, but that does not matter here), you just assume

    d(HPS) / d(SP) = 1.0

    which is wrong.

    Now, let’s take a look from another point of view. Let’s write down the [units] for a several expression.

    SP_to_MP5 = 2.0 [MP5/SP]

    [MP5/SP] means that if have a value X measured in [SP] and we multiply it by SP_to_MP5 coefficient measured in [MP5/SP] then [SP] will eliminate and we will get an equivalent value in [MP5].

    Here’s the formula you use for the crit calculation:

    Flash Heal Crit MP5 value = (base heal + (SP * SP co-efficient)) * SP MP5 value * proportion from crit * (1 – over-heal %)

    Let’s write down the units for all the variables used:

    Base_Heal [Healing]
    SP [SP]
    SP co-efficient [Healing/SP]
    SP MP5 value [MP5/SP]
    Proportion from crit []
    Over-heal []

    Flash Heal Crit value = ([Healing] + ([SP] * [Healing/SP])) * [MP5/SP] * [] * [] =
    ([Healing] + [Healing]) * [MP5/SP] = [Healing] * [MP5/SP]

    Healing and SP will not eliminate, because they are different units!
    If you are to calculate the flash heal [MP5] value, you need the [Healing/SP] coefficient. This exactly the same thing that we’ve seen in the previous example, where we lacked the d(HPS)/d(SP) calculation.

    And finally, we can come up with some example using the real numbers.
    Let’s take the Combined_Healing formula mentioned above for example.

    Combined_Healing = (Base_Heal + (SP * SP_coeff)) * (1.0 + CR / 9182)
    Combined_Healing = (1887 + 0.8068 * SP) * (1.0 + CR/9182)

    Let’s assume that base SP = 2774 , base CR = 1000 (about 22% crit chance)

    Effect from increasing CR by 1 can be calculated as:

    Healing_IncCR = Combined_Healing(CR+1) – Combined_Healing(CR)
    Healing_IncCR = (Base_Heal + (SP * SP_coeff)) * (1 / 9182)
    Healing_IncCR = (1887 + 0.8068 * 2774) / 9182 = 0.4493

    If you would multiply this value by your overhealing coefficient (1 – 0.5) and SP MP5 value (2.0) you’ll get exactly the same number as in your calculations on the crit rating page. So I assume that you will agree with this numbers.

    But we will move a step further and calculate the effect from increasing SP by 1!

    Healing_IncSP = Combined_Healing(SP+1) – Combined_Healing(SP)
    Healing_IncCR = (SP_coeff) * (1.0 + CR / 9182)
    Healing_IncCR = 0.8068 * (1.0 + 1000 / 9182) = 0.8947

    This value is the same thing as the missing derivative from the first part or missing conversion coefficient from the second part.
    You just assumed that SP value is 2.0 and so crit rating seems more than 4 times less valuable than Spell Power in your case (we’re not including Divine Aegis now, remember?)
    While actually it’s only 2 times less valuable.

    I hope you’ll see my point now 🙂

    About your question on my Weighed scale.
    My stat weights for the “Stat Weights (Regen)” table are measured in MP5 (hence MP5 has the 1.0 value)
    The “Stat Weights (T-put)” table uses HPS units.
    I then normalize Throughput values by Spell power, so that SP = 1.0

    Now we have two scales, one is regeneration value relative to MP5, and second is the throughput value relative to Spell Power.
    This two scales are unrelated in general case (the real value of regen vs throughput varies greatly between fights, groups and playstyle), so I just assume that SP/Mp5 = 2.0 based on my and your experience.

    So each stat in the final mixed table is calculated as:

    Weighed_value[Stat] = Throughput_Value[Stat] * 2.0 + Regen_Value[Stat] * 1.0

    Hence MP5 = 1.0 and SP = 2.0.

    My calculations are based on the flash heal spam, just as yours. Now I’m working on adding the spell rotation support. As soon as I’m done I will be happy to share it with you, so you can get your own char from armory and run the simulation for your own stats.

    I hope that my comments don’t sound agressive or rude, it’s what I’ve tried to avoid. I don’t want to say something like OMGFAIL! I just wanted to share my point of view which may not be correct of course.

    Hope to hear from you soon,
    /Thorrion @ Dunemaul-EU

  57. Thorrion says:

    Unfortunately, my math drawings failed..
    Hope you’ll be able to figure it out anyways.

    /Thorrion @ Dunemaul-EU

  58. Heals Mcgee says:

    I appreciate all your hard work! Keep up the amazing crafting.

  59. BobTurkey says:

    Wow Thorrion! That was huge. I followed most of it.

    The stat weights I calculated are measured relative to MP5. So Discipline Crit rating MP5 = 0.4493 (from

    With 2 SP = 1 MP5 (or 1 SP = 0.5 MP5, same thing), crit rating is just under half (0.4493 x 2.0 = 0.8986) of the value of spell power (2.0).

    You are saying because i’m using 2:1 throughput:regen ratio that I should be doubling the critical component also? Haste too I imagine?

    Gobble gobble.

  60. Thorrion says:

    Hi, Bob!
    First of all, thanks for a response!
    Second, I’ve made a typo in my last post:
    Healing_IncSP = Combined_Healing(SP+1) – Combined_Healing(SP)
    Healing_IncCR = (SP_coeff) * (1.0 + CR / 9182)
    Healing_IncCR = 0.8068 * (1.0 + 1000 / 9182) = 0.8947
    should be read as
    Healing_IncSP = Combined_Healing(SP+1) – Combined_Healing(SP)
    Healing_IncSP = (SP_coeff) * (1.0 + CR / 9182)
    Healing_IncSP = 0.8068 * (1.0 + 1000 / 9182) = 0.8947

    which means that this 3 strings are all related to the Increased healing due to spell power increase.

    Now, back to your comment about doubling your weights.
    Well, not exactly.

    Quote from the link you’ve provided:
    “To determine its value as MP5 we equate the total heal for the FH as MP5 based on the SP to MP5 ratio (2.0:1) and then calculate how much is additional healing is due to the critical.”

    You calculate “total heal”, but use “spell power to mp5 ratio”.

    But, TOTAL HEAL is not equal to SPELL POWER.

    Look, you have 3 throughput stats: Haste, Crit (rating + int), Spell Power.
    1. You are using “total heal” as a measure of throughput;
    2. You are using 2:1 ratio to convert throughput scale to regen scale;
    3. Stat weight then is the amount of additinal healing due to increased stat;
    4. You calculate the total heal due to critical hit effect;
    5. You calculate the total heal due to haste effect;
    6. You just take spell power to be 2.0 <— Wrong!

    You should instead do:
    6. Calculate the total heal due to increased spell power;

    Only then you can compare the relative contribution of Haste, Crit and SP to your total output healing.
    How would you react if for example I would calculate Haste rating value as you do, assume spell power to be valued as 2.0 and also assume the crit reting to have a value of 1.5 without any justification?
    You already have the formula in your hands to calculate the healing output bonus of spell power!

    If I still haven't convinced you, try the following:
    Take your stat weights as an input and do the reverse calculations – calculate the total healing done.
    Here I summed up your throughput parts of crit rating (output+divine aegis+inspiration) and took the throughput part of haste rating:

    SP = 2.0
    Crit Rating = 1.1316
    Haste Rating = 1.1621

    If you increase your stats by the inverse of their weights, you should get the equal amount of healing output increase.

    incSP = 1.0/2.0
    incCR = 1.0/1.1316
    incHR = 1.0/1.1621

    I think that you'll get more or less equal results for Crit and Haste Ratings, but your output due to increased SP will be much less.

    Also, without looking at your calculations, answer yourself this question:
    Does the SP contribution to the healing output scale with amount of crit or haste your charater has?

    /Thorrion @ Dunemaul-EU

  61. Firebert says:

    Hmmm. Interesting.
    As it stands, my stat weightings over-value crit, with the non-variables SP=2 and MP5=1. The values I have are such:
    MP5 = 1
    SP = 2
    Spirit = 0.3475
    Intellect = 0.9588
    Crit = 1.3626
    Haste = 0.9037

    But, however, if I halve these values, and put SP=1 and MP5=0.5, I get:
    MP5 = 0.5
    SP = 1
    Spirit = 0.3475 (no change)
    Intellect = 0.7471 (77% of above)
    Crit = 0.7059 (51% of above)
    Haste = 0.1373 (15% of above).

    These are based off my own current gear’s stats, and with very slightly modified assumptions to yours but totally comparable. Is dividing down a valid action to make upon those numbers?

    Also, Thorrion, I think that you’ve found a little hole. With the value of Spell Power overestimated in your Flash Heal Critical example, this should cut down the Crit MP5 value in DA in our weightings by at least little, or otherwise throw a new scaling on it. I enjoyed reading it and have started playing around with it a little more.

  62. Thorrion says:

    Hi, Firebert!
    Nice to see you interested.

    Well, if what you are doing is calculating the MP5 value of Spirit, then if you decrease the MP5 value (from 1.0 to 0.5) then your Spirit MP5 value should decrease by half as well, assuming that spirit only gives you MP5.

    I suggest you use HPS (healing per second) as a measure of throughput instead of SP (You can still scale it afterwards to make SP value to be 1.0, 2.0 or whatever you want)

    Have you calculated the actual Spell Power value as I suggested?

    As for my claculations dividing by 2.0 or any other number will keep the relations between the stats intact.

    Here’s the output for SP=2.0, MP5=1.0

    Stat Weights (Regen):
    Stamina: 0.00
    Intellect: 0.81
    Spirit: 0.41
    MP5: 1.00
    Haste: -0.54
    Crit: 0.00
    Spell Power: 0.00

    Stat Weights (T-put):
    Stamina: 0.00
    Intellect: 0.16
    Spirit: 0.00
    MP5: 0.00
    Haste: 0.85
    Crit: 0.46
    Spell Power: 0.57

    Stat Weights (Weighed T/R = 2):
    Stamina: 0.00
    Intellect: 1.38
    Spirit: 0.41
    MP5: 1.00
    Haste: 2.45
    Crit: 1.62
    Spell Power: 2.00

    And here’s for SP=1.0, MP5=0.5

    Stat Weights (Weighed T/R = 2):
    Stamina: 0.00
    Intellect: 0.69
    Spirit: 0.21
    MP5: 0.50
    Haste: 1.23
    Crit: 0.81
    Spell Power: 1.00

    In the tables above Regen Weight is how much MP5 will increase of the stat by 1 give, Tput weight is how much HPS will it give.

    Weighed value is just a mix of two weights. See my previus posts for details.

    And by the way, I don’t like mixing throughput with regen at all. I rather use two separate scales and have 2 separate values for each gear item I consider.

  63. Thorrion says:

    And by the way that big value of haste is because I use the Flash Heal spam model. We’ll see how it changes after I model the proper rotation with a lot of Borrowed Time and account for latency.

  64. Kenan says:


    I’m actually a druid, but your blog was returned in Google search results when I was hunting for an interesting question. Sadly, this seemed to be the most relevant thread you have that still has comments open, so I apologize for the (slight) derailment.

    As you clearly know, the tooltips for the Runed Cardinal Ruby indicates 23 spellpower gain while hybrid gems like the Reckless Ametrine show 12 spellpower/10 (insert stat of choice here). You’ve indicated that you believe the Runed Cardinal is underbudget, which I can agree with. However, it seems you’re assuming that the gem is *actually* providing you with 23 spellpower and not, say, 23.14159265 spellpower instead. Since Blizzard has been known to preserve decimals in stats (even though they are displayed in whole numbers), could it be possible that the gem is actually correctly budgeted but provides slightly more or less spellpower then actually indicated?

    I’ve been considering this possibility for a while now, as many people point out that if you gem two Reckless Ametrines, you’ll get more spellpower then gemming one Cardinal and one Quick King’s Amber (12+12 > 23). However, this never really sat well with me since I’d assume Blizzard would want to keep things as equal as possible. Would you have any thoughts on this theory?

  65. BobTurkey says:

    I’ve never heard of discussion of Blizzard preserving decimals, but thats not to say they don’t.

    Your suggestion about it being worth slightly more than 23SP is quite possible.

    Gobble gobble.

  66. Zusterke says:

    I know decimals are preserved in other situations. For example if you should obtain 26.4 mana from some effect, you’ll get 26 and 27 mana from the effect with a 3/5 and 2/5 ratio.
    I haven’t bothered checking out if this was correct for the gems.

    @Thorrion do your final numbers include DA? It would seem odd if they don’t.. but interesting none the less. Also, can you give the exact gear and situation assumptions you did your calculations with? Especially with respect to the throughput stats this would matter a lot.

  67. Thorrion says:

    Most of the assumptions are taken from BobTurkey’s posts.

    Messy wall of text incoming. Feel free to ask questions!

    Base stats:
    Crit = 511;
    Haste = 546;
    Intellect = 1412;
    MP5 = 162;
    Spellpower = 3352;
    Spirit = 770;
    Stamina = 1240;

    Fight parameters:
    i5sr = 0.95; (time spent inside 5 seconds, relative to the total time)
    length = 480; (in seconds)

    replenishment_return = 0.01;
    replenishment_uptime = 0.80;

    shadowfiend_cd = 360;
    shadowfiend_mana = 0.05;
    shadowfiend_strikes = 10;

    rapture_cd = 15.0;
    rapture_mana = 0.025;

    divineaegis_effect = 0.9; (just some empirical weight)
    divineaegis_bonus = 0.30;
    divineaegis_duration = 12.0;

    total_damage = 1000000.0;
    inspiration_duration = 15.0;
    inspiration_reduction = 0.1;

    arcanetorrent_cd = 120.0;
    arcanetorrent_mana = 0.06;

    hymnofhope_cd = 360.0;
    hymnofhope_buff = 0.20;
    hymnofhope_mana = 0.12;
    hymnofhope_effect = 0.90; (just some empirical weight)

    overheal = 0.30;
    crit_overheal = 0.20;

    Specialization (+5% holy crit)
    Strength (+15% int)
    Skyflare Diamond (+2% int)
    of Kings (+10% stats)

    Flash Heal spam is assumed.

    Total mana regen is calculated as the sum of the following:
    – Weighed Regen (Int/Spi regen, weighed by i5sr uptime and accounted for Meditation 3/3)
    – Mana pool divided by fight length
    – Replenishment returns
    – Shadowfiend returns
    – Rapture returns
    – Arcane Torrent returns
    – Hymn of Hope returns
    – Additional mana required because of Spell Haste
    – MP5 stat on gear

    Total throughput is calculated as the sum of the following:
    – Spell HPS
    – Divine Aegis HPS
    – Inspiration HPS

    Spell HPS is calculated as:
    FH_HPS = (Flash_Heal_total_amount) / (Flash_Heal_cast_time)

    Flash_Heal_total_amount = FH_normal_heal * (1.0 * (1.0 – overheal) + 0.5 * crit_chance * (1.0 – crit_overheal))
    Flash_Heal_cast_time = max(1.5 / (1.0 + Haste_percent), 1.0)
    FH_normal_heal = FH_tooltip_heal + (spellpower * FH_sp_coeff)

    Divine Aegis HPS calculation is not accurate, because incoming damage is unknown. I didn’t account for 10k DA cap, but instead reset DA stacks every 12 sec.

    DA_HPS = avg_stacks_per_duration * da_heal / divineaegis_duration;
    avg_stacks_per_duration = 0.5 * (casts + 1) * crit_chance;
    casts = divineaegis_duration / Flash_Heal_cast_time;
    da_heal = FH_normal_heal * 1.5 * crit_chance * divineaegis_bonus * divineaegis_effect

    the tricky part is avg_stacks_per_duration calculation.
    I thought of DA stacks as a sort of ladder, and then calculated the statistical average number of stacks on each timeline division, i.e. spell cast. Based on that, I’ve came up with the formula above to calculate the average number of stacks that can be obtained during the DA duration. It was actualy a sum of arithmetical progression multiplied by the crit chance. Then I’ve made some experimental simulations based on RNG to check if it was correct.

    Inspiration HPS is simple:

    INSP_HPS = FH_HPS * inspiration_reduction * insp_uptime;
    insp_uptime = 1.0 – pow((1.0 – crit_chance), hits);
    hits = inspiration_duration / Flash_Heal_cast_time;

    That’s it for MP5 and HPS calculations.
    To calculate stat weights, I’ve calculated MP5 and HPS increasing each stat by 1 (before applying the buffs of course) and subtracted the base MP5 and HPS values.

  68. Zusterke says:

    A few questions or notes:
    – By the look of your formula, I’m not seeing a healing bonus from Focused Power. While this bonus applies equally to crit as to haste values, it may impact crit differently due to overhealing.
    – It could be merely a symbolic simplification , but I get the impression that you calculated your total haste as (1+Haste%). This haste% is unclear. If multiple sources of haste are taken into account (ie: gear and talents), these sources multiply. IE casttime/[(1+Hastefromgear) * (1+HasteFromTalents) * (1+HasteFromBuffs)]. This is important for 2 reasons: your native haste increases a lot more, which will impact your hps outcomes and the balance between crit and haste. The multiplicative bonus behaves differently, mathematically, in a way that crit benefits more from it (which is exactly reflected by the native haste going up and thus making crit more attractive).

    I am a bit puzzled about how you calculated DA. To me it seems more reasonable to assume that the shield will be consumed. We are talking about a 2.5K shield. If even that isn’t consumed within the duration, then it is unlikely that hps matters a lot. So.. I would assume DA like a healing bonus on crits that doesn’t have the overheal factor as such:

    Flash_Heal_total_amount_with_DA = FH_normal_heal * ((1.0 – overheal) + crit_chance * (0.5 (1.0 – crit_overheal) + 0.45))

    Although, you could restructure the formula to:
    HPS = (FHeal / base_cast_time) * (1+HasteFromTalents%) * (1+HasteFromBuff1%) * (1+ Crit% * (45% + 50% * (1-crit_overheal))) * (1+HasteFromGear%)
    What’s interesting about the formula this way is that you single out some behavior and it should lead to simpler (partial) derivatives.

    Still, regarding your DA method:

    DA_HPS = avg_stacks_per_duration * da_heal / divineaegis_duration;
    avg_stacks_per_duration = 0.5 * (casts + 1) * crit_chance;
    casts = divineaegis_duration / Flash_Heal_cast_time;
    da_heal = FH_normal_heal * 1.5 * crit_chance * divineaegis_bonus * divineaegis_effect

    If I work this out, I get:
    DA_HPS = [0.5 * (casts + 1) * crit_chance] * [FH_normal_heal * 1.5 * crit_chance * divineaegis_bonus * divineaegis_effect] / [divineaegis_duration / Flash_Heal_cast_time]

    There’s an error there which becomes more apparent as you rejuggle the formula a bit:
    DA_HPS = (crit_chance * crit_chance) * [0.5 * (casts + 1)] * [FH_normal_heal * 1.5 * divineaegis_bonus * divineaegis_effect] / ….
    So you took the crit factor twice! And since crit chance is below 100% it’s likely to have diminished your DA HPS outcome.
    The error stems from inserting your crit chance in your DA heal, I think. Your crit chance is already included in the ‘average stacks per duration’ and your amount of crit does not impact how big the DA procs are, only the frequency.

    Another general note to your work is that you take into account overhealing, overhealing from crit but not ‘overhasting’. Your overheal is based on the assumption that the target does not need the extra healing from SP and crits. The same can be said about haste: if SP/crit/da is going to overheal, what good is a higher frequency of heals? Of course, this becomes valid where you consider healing different targets, but then a shieldspam /aoe heal is probably a stronger answer. So, if you would take into account overheal, I fear a note of ‘overhaste’ should be included as well. This would also incorporate the increased effect of latency with higher haste and you have a decent amount of haste from talents and raid buffs already. Note that this overhaste impacts crit and SP as well, since haste is less effective.

    Bringing this all together, I feel crit is a bit underestimated in your theorycraft.. provided I understood it correctly, and haste is a bit overestimated since it doesn’t seem to suffer from any downsides and is 100% effective.

    After having said all this, there only remains me 1 thing to say and that is: very nice work! It’s always a pleasure to read such good and detailed work. Thank you for the joy it gave me to read it and analyze it.

  69. Hey BobTurkey – I’m just want to say that I love your very detailed number crunching! In fact, it’s one of the most detailed guides that I can find.

    But I don’t agree with your Blogs subtitle – this stuff is not for casual gamers 😉 It’s great content for maxing your characters.

  70. […] for the reply Hegen. The stat weights were from here: Priest Theorycrafting for 3.3 – Summary (Part 6) � BobTurkey’s WoW Blog Where he ranks Crit @ 0.7387 and Haste @ 0.6096. I have seen similar numbers pop up elsewhere as […]

%d bloggers like this: