Life in Group 5 – A Resto Shaman Blog
A resto shaman perspective on raiding


September 11, 2012

Resto Shaman Mastery – Pandaria Style

More articles by »
Written by: Vixsin
Tags: , , , , , , , ,

When Resto Shamans’ Mastery was first launched in the raiding scene, in the patch before Cataclysm release, it was difficult to view it as anything but a passing fancy, a distraction in our ongoing love affair with Haste. Little did we know at the time, but Tier 11 was about to convince us of exactly the opposite, and over the course of the expansion, we’d learn just how powerful Mastery could truly be given the right triage environment. In the end, Mastery became the go-to stat for Shaman in hard modes, and our assignations with haste were left to normal modes and those times at the end of the tier, when fights were shorter and mana less constraining. But, is that going to be the case in Mists of Pandaria or are we wearing blinders once again?

In light of the changes and the revamped triage environment waiting for us just around the corner, today I wanted to take some time to revisit Mastery, its effects, and its interaction with the secondary stat that’s been waiting in the shadows all time time—Crit. Because, I have the sneaking suspicion that Resto Shaman are once more on the verge of another unexpected romance.

(Feel free to skip to the Mastery v. Crit section if you don’t need the background on Mastery’s calcs and application).


How Deep Healing Works

Although Mastery’s contribution may seem mysterious because it’s not discernable in WoW’s combat logs, in concept, Mastery is fairly simple. For Resto Shaman, every 600 points that we invest into Mastery, we receive a +3% bonus to Deep Healing (which is to say that every 200 points nets us 1%). Combined with a base value of 39% at level 90, which is comprised of 24% base plus 15% due to shaman’s passive Mastery buff, your percent Mastery indicates the maximum amount of additional healing that your spells will do.

Because of its linear application, this means that you’ll receive the maximum benefit of your Mastery when a target is at 1 HP and no benefit being applied when the target is at 100% HP. Your contribution between those two points can thus determined by the following equation:

Mastery bonus on heal = (1 – (% HP of Target/100)) x Maximum Mastery contribution

So, for a heal cast target who’s at 70% HP by a level 90 shaman who has only the base +39% Deep Healing, the calc would be:

Mastery bonus on heal = (1 – (70/100)) x 0.39

Mastery bonus on heal = 0.117 or 11.7% bonus healing


What Deep Healing Affects

Since 4.1, Mastery has operated under the following conditions:

  • It will affect all single-target heals, including HW, HS, GHW, and UL
  • It will affect all mutli-target heals, including CH and HR
  • It will affect all of CH hits independently
  • It will affect BOTH the initial hit of RT and the subsequent hot ticks (which will vary based on target HP at the time of the respective tick)
  • It will affect HST and ES
  • It will affect Earthliving’s ticks
  • It will not affect the heal from Glyph of Cleansing Waters
  • It will not double-dip on AA, meaning that AA’s heal will be 30% of the initial heal with no additional Mastery bonus applied.
  • It will not affect the “healing” done by Spirit Link Totem

In Pandaria, it’s also important to remember:

  • It will affect Healing Tide Totem
  • Its affects will be transferred via Ascendance, but Ascendance’s healing will not be subject to an additional Mastery bonus. (For example, a base heal of 10,000 that becomes an effective heal of 10,500 subsequent to Mastery’s effects, will be duplicated by Ascendance as 10,500 and not subject to an additional Mastery bonus when the Ascendance heal is applied to a target).


Mastery v. Crit – The Great Debate

Although Resto Shaman (and theorycrafters) can delight in analyzing the benefits of three secondary stats, when it comes to a triage environment, there are really only two to consider—Mastery and Crit. Because of haste’s negative impact on Mp5, its HPS gain is offset by an increased demand for Spirit, something which will have a finite limit in every tier. This means that haste will continue to be limited by our available regen (and the duration of the encounter in question). So, with only two secondary stats to play with, the question then becomes, “which is more powerful—Crit or Mastery?” This was a question that I addressed at the start of Cataclysm in the monster post: Resto Shaman Mastery: The Complete Story. So, last week I dusted off the spreadsheets that I used in that analysis and started through the same process once again.

Now, there are several important factors to consider when analyzing Crit and Mastery. First, it’s important that we recognize that we’re dealing with two data sets—heals that can trigger Ancestral Awakening (ie: HS, GHW, HW, RT, UE) and heals that can’t trigger it (everything else). For heals that cannot trigger AA, the critical strike bonus is limited to 200%, but for heals that can trigger AA, the critical strike bonus becomes 200% + (2 x 30%) = 260%.

Second, it’s important to remember that Crit and Mastery are two factors which interact when contributing to the total value of a heal. In other words, if you are ever healing someone below 100% HP (which should be the case, with rare exception) you will be dealing with a contribution from both Mastery and Crit rating. This is why the analysis of Crit and Mastery isn’t simply setting Mastery = Crit, and solving for the HP % at which the contribution is the same. You have to include Mastery in both sides of the equation.

Which leads us to the third factor in this analysis—the amount of stats you have available to allocate. Because Crit and Mastery scale differently and are unequal in terms of conversion, we need to identify a budget of stats to work with during the course of our analysis so that we have a reasonable frame of reference. (Which is to say, it wouldn’t do us much good to try and model a Crit rating of 95% or 0%, nor would it be appropriate to model Mastery in the same way; we want to identify a reasonable range of values that we can expect to see in MoP). So, while I used 2000 in my Cataclysm analysis, it seemed more reasonable to work with a value of 4000 this time around to account for the stat inflation we’ll see on Pandaria gear. This equated to a range of 6-12% for Crit and 39-59% for Mastery.

So, where did all of these factors get me? To the point where I could, for a given base heal value, % HP, and ratio of Crit to Mastery, determine an average expected heal. Plotting those expected heals netted me the following graphs:


MoP – Output Scaling for Heals that trigger AA, by Mastery-Crit ratio and % HP

MoP – Output Scaling for Heals that do not trigger AA, by Mastery-Crit ratio and % HP

What does the two graphs tell us? A couple things:

  1. The HP values where the lines intersect are those points where Mastery has a greater impact on HPS than Crit. For spells that do not trigger AA, the data suggests a value of approximately 64%. For spells that do trigger AA, the data suggests a value of approximately 39%.
  2. The point of intersection between all these lines is NOT a constant.

This is, in fact, the same data behavior that was observed in Cataclysm. Whereas a theorycrafter might be inclined to think that the data would behave as if it had a single pivot point, in fact, it shows us that as you begin to stack more and more Mastery, the point of rotation (the point of intersection between that line and another) actually decreases. This behavior is something that can be illustrated if you take the data above a step farther, and look at the Mastery per Point of Healing. And that’s exactly what I did below, by dividing the total Mastery investment, in points, by the healing gain (expected heal – base healing value).


MoP – Mastery Invested per Point of Healing, for Heals that Trigger AA


MoP – Mastery Invested per Point of Healing, for Heals that Do Not Trigger AA

If the point where Mastery overtook Crit in terms of HPS benefit was a constant, the asymptotes implied by the above graphs would be stacked. Instead, you can see that they shift downwards, as I increased the percentage of my available stat budget that I invested into Mastery. Alternately, you could invert the data from the above, and look at the Healing Gain per point of Mastery, which is illustrated below:

MoP – Healing per Mastery, for Heals That Do and Do Not Trigger AA

Likewise, if Mastery versus Crit had a single breakpoint, you would get the same value when you compared the point of intersection between two lines. But, as I discovered through my analysis, the point of intersection for the Mastery-Crit combinations that I looked at varied:

  • For heals that trigger AA, the breakpoint range was: ~40 – 46% HP
  • For heals that don’t trigger AA, the breakpoint range was: ~64 – 66% HP

If you’re interested in looking at the actual data, in Excel format, it can be downloaded here. It can also be viewed on Google Docs here.



The relevant and actionable information that you can get from this look at Resto Shaman Mastery in Mists of Pandaria is twofold:

  1. The breakpoints have dropped, significantly, from where they were in Cataclysm. Mastery is not the standout king any longer.
  2. The breakpoints where Mastery contributes a greater amount of throughput than Crit are: approximately 43% HP for heals that can trigger AA, and approximately 65% for heals that can’t trigger AA.

With this information, it then becomes a question for each Resto Shaman to identify which heals make up the majority of your effective healing, and if encounter mechanics are hitting hard enough to drop your raid consistently below those Mastery v. Crit thresholds. In Cata, we absolutely saw fight mechanics and triage states where the raid simply could not be topped, and from what I’ve seen from testing thus far, there’s compelling enough evidence to suggest that that will be the case in MoP as well. But, with the breakpoints dropping to much lower levels than we saw in Cata, I don’t think Resto Shaman can continue to ignore Crit to the degree we did before. We may not be eloping with it any time soon, but it’s certainly worth rekindling a friendship with just in case Mastery starts to seem a little stale.


  1. Interesting work. I have one, somewhat unrelated question. You state:

    Because of haste’s negative impact on Mp5, its HPS gain is offset by an increased demand for Spirit, something which will have a finite limit in every tier.

    In Holy Paladin theorycraft (last I checked) Haste is actually considered a stat that *reduces* mana consumption, because it allows to shift your mix of spells towards the cheaper spells while still maintaining higher throughput. High haste allows one to cast more Holy Lights (cheap) and fewer Divine Lights (expensive) for the same throughput.

    Is there something about Resto Shaman spells that this strategy does not apply?
    Rohan´s last post ..Rotations

    • Excellent question; I had to take some time to think about this one.

      I would agree with Therya that part of the difference is due to the design of paladin throughput and the “buffer” that Holy Power introduces into the equation, especially given that your HP finishers in MoP look like they’re be contributing a significant portion of Pallies’ HPS.

      But an additional part of the differential between Resto Shaman and Holy Paladins is that while your haste and mastery work collaboratively, for Shaman that relationship is independent. More casts do not necessarily equate to the same percentage contribution of Mastery. Even more importantly, at low levels of HP, Mastery outperforms Haste significantly. For example, on a target with 80% HP, a badly-geared level 90 Shaman could expect to see HW have an average HPS of ~25k. Given the same stats, on a target with 20% HP, that HW is going to have an HPS of ~46k. So, even if I was able to stack haste to get my HW down to a 1-sec cast, it still wouldn’t be able to match that HPS; it would come up just shy at ~40k HPS. (Again, rough numbers, but the proportion is what’s important).

      And that variability is what has pulled shaman away from haste in cases where Mastery is a big player. But (and I think this is the really interesting implication of your question) in cases where Mastery doesn’t stand up well against our other secondary stat option (Crit), and that’s exactly what I’m talking about about, haste is an excellent complement. So it’s very possible that as our valuation of crit increases, we will be using Haste as a way to “save” mana to be used on higher HPS spells.

    • Pirjo

      I did a lot of theorycrafting for haste/crit build as a 4.2 pally when crits got 100% and HL got 100% beacon transfer (basically that more spirit was silly). So let’s convert that argument to shaman’s…
      -You are chain casting HW, maybe a set number of GHW’s in there, riptide is in there.
      -You have just enough mana to do this at your current level of haste.

      You could make the argument that between:
      (1) Gaining X spirit to get Y more healing by converting HW to GHW
      is less effective than
      (2) Gaining M haste and N offsetting spirit to get Y more healing done by casting more HW’s.
      Because (X)>(M+N).

      While this may be true for beating spirit, this isn’t necessarily true for beating out either crit or mastery. Actually since you are dealing with single target heals only in this situation, “crit” would decimate “haste+offsetting spirit” due to AA and resurgence. It would be a butchering really.

      Leaving theory behind… when I think of a fight where all I feel like I do is single target heal like this situation, I think of Rag10H. And honestly, I’d rather have the extra spirit so I can choose to drop a GHW on the tank than have a stronger time-averaged healing done given by a little more haste. Once we leave the single target healing realm though, the mana cost of Healing Rain removes our ability to chain cast and really makes the haste argument mute.

      • shammypie

        however in the latter situation where you’re only dropping healing rain and a heal or 2 here and there the haste feeds into healing rain and greatly increases tick rate and earthliving proc rate.

  2. Therya

    With resurgence returning us 148% mana f every healing wave that crits, you could argue that haste in that manner would do the same for us if not even more. Remember though that Holy palas get many free procs also. While they are able to maintain a decent throughput using cheap heals to sustain their mana and free spells to sustain their output to some extent we cannot do the same. It’s a long discussion that involves exactly what Vixsin posted above however. And given the fact that we only need around 1194 haste to reach our first important breakpoints and then 5676 just to add one more tick on our riptide, you can see how mastery and our new favorite pet crit will get more love than haste.
    Vixsin, excellent detailed and comprehensible post, I will use your info (with my many thanks and linksies) to make the mastery section in EJ more complete since your breakpoints have more numerical depth and take into consideration more variables making them incredibly accurate.

    • shammypie

      I don’t think you can really justify stacking haste to healing wave spam since the mana is only 148%. this means we are only up by 48%. therefore you need at least 52% crit for this to be a mana positive strategy. You could get better mana regen by not casting. The point of casting healing wave is that you quickly reach a point where you regen becomes greater than the (cost/cast time + crit returns+ mps). As you increase haste the cost actually increases and the formula returns less. furthermore looking at present values in game healing wave only returns 133% mana. finally due to the way mastery works, a GHW on a low hp target is more healing than two healing waves since the second healing wave will be devalued due to new hp% mastery bonus.

  3. Mazhug

    Great theorycrafting. I’m loving your series of articles before MoP.

    Just a question and a comment.

    The question: where it says “This equated to a range of 5-12% for Crit and 33-59% for Mastery.” I thought that base mastery was 39%. Why is that you are trying to predict below the starting point?

    The comment: I think where it says:

    “Mastery bonus on heal = (1 – % HP of Target) x Maximum Mastery contribution”

    The “1″ should be a “100″ or “(1 – (% HP of Target/100))” so we don’t get a negative value.

    Just being a little smart-ass here =)

    • No worries at all–those are great catches. The 33% should have been 39%; that’s definitely a typo.

      The implication on the % HP was that it was in decimal form, but I could understand how that might lead someone astray if they didn’t pay attention to the example. I’ve updated both accordingly.

  4. Xico

    Thanks for the interesting post, but I think I have a small correction to make.

    I can’t really confirm this as I’m no longer subscribed, by I’m pretty sure that crits for spells that benefit from AA are 260% of the normal amout: a heal that hits for 100 will crit for 200 and proc an AA heal of 60 (30% of 200), for a total 260 healing (or 260% of the normal heal).

    This is further increased by the metagem, but I’m not sure about the exact values. For damaging spells, the crit metagems used to make crits do 209% normal damage, but this was changed when they changed the crit damage talents in Cata. It seems it’s now at 206%. I’m not sure how it works with heals, but it should be taken into account. If it’s similarly at the 206% mark, the total healing with AA will be 267.8%. This may shift your number a bit towards crit rating.

    • Spot on; typos be damned! 260% was used in the calcs, but the narrative incorrectly cited 230%. That’s corrected.

      Re: meta gem inclusion, I did exclude it from the analysis for simplicity’s sake. But you’re correct in that Revitalizing (Spirit + 3% Crit effect) will be a strong contender early on in progression. And yes, it would drop down the breakpoints by around 2% on the high and low ends.

  5. You must have an insanely high iQ Vix. I don’t have english as my first language, and I must admit that I fall short on understanding the entire meanings sometimes, but generally it’s fine :-)

    Thanks for another awesome blog.

  6. Rust

    I’m glad someone else came to this conclusion – every since the Crit healing buff and the Burning meta change Crit has been very competitive with Mastery (farm status aside) but largely ignored by the community. I wanted to point out something though:

    The breakpoints should be (theoretically) lower when looking at full fight throughput because of assumptions we have made and can make. The assumption made is that the triage healing model precludes people from being topped off. The assumption we can make off of that is that as a result encounters are designed with more frequent, smaller damage packets than under the assumption that damage comes in one huge burst. I think most of the fight mechanics for T14 we’ve seen bear that out.

    As a result of that, Crit’s value goes up because any healing done over the breakpoint reduces the chance of reaching the breakpoint, and it also means that the value of healing when HP is high is closer to the value of healing when HP is low.

  7. Bindura

    Bear in mind that at higher HP levels the chances that a Crit will over heal increases, thus reducuing effective healing done and diminishing the value of Crit at higher HP levels.

    • That’s very true, but then consider that every stat is devalued at high levels of HP. Because, if you’re at the point where critical hits are consistently recording high overhealing, then you’re in one of two types of situations:

      1) You’re mistiming your heals and/or spamming.
      2) You’re overhealing the encounter.

      In the case of the former, proper stat prioritization isn’t going to help you. And in the case of the latter, your answer is haste, and all you’re accomplishing is beating other healers to the punch.

      • Taughror

        Bindura’s point is important to take into consideration because it could, in some cases, affect the break points for target hp. If you are above the break point (i.e. crit is providing more benefit than mastery) and the bonus healing from a crit would cause over healing where the bonus healing from mastery would not, it would mean that breakpoint is actually being set too low.

        This can be accounted for by taking into consideration the ratio of the healing value of the spell compared to the total health of the target.

  8. A very good analysis, indeed. Top quality.

    If I might add a comment; since Crit is more attractive now than ever before due to **pure healing throughput** and Crit ALSO gives us mana regen, there may be a point where Crit becomes better to stack than Spirit because it’s giving us more in combination; extra mana to spend on healing AND more healing throughput. That might make it even more attractive, though my own analysis has hit the problem that it is now very similar to what I do for my day job (hence I can’t ever motivate myself to do yet more bloody work).

    Anyway, excellent post. V useful. Ta! :)
    Stoove´s last post ..First

    • As a financial consultant who often gets tasked with writing executive briefs, I can completely relate.

      I look at spreadsheets and databases all day, and I’m oftentimes asked (because our VP likes my writing style, rofl) to distill that information into easily-digestible brief. And then I go home, and look at spreadsheets and databases (argubly, much nicer ones), and sometimes distill that information into what I hope is an easily-digestible post. Which explains why some nights I just go home and spend the evening screaming at people in Call of Duty! :-P

      • Ehehe, well as a scientist, plotting obscure graphs and trying to work out what the hell it means in practice is what I do at work. Almost entirely in Excel. Then when I get home, all I want to do is forget about the bloody program. Seriously, that green icon is invading my dreams now…

        But my chosen escapes are Minecraft or Skyrim. CoD hasn’t been my thing since it got all fancy and modern on us ^_^
        stoove´s last post ..Science, or; “we will devour your soul”!

      • Shamanhands

        I’m not sure the mana return from crit can be modeled and converted into an (e)HPS value, anyway. I’m foresee quite a bit of hand-waving and unsatisfactory simplifications. Regardless, I have to agree with Stoove that there should be SOME mention of crit returning mana. After all, crit’s contribution to mp5 means spirit can be traded for some other useful stat…like more crit…which increases hps.

  9. Rodrigo

    U have a Contact Me Link but never reply my things =(

    • Hrm, that’s odd. I don’t see an email from you in my Inbox. Try sending it directly ( and I’ll work on making sure that the Contact Form isn’t dumping my messages out into the middle of space somewhere.

  10. I’m a bit surprised no one else has pointed out that your second set of graphs are wrong. Your graphs for “Healing Gain per Quantity of Mastery Invested” fail a very quick and basic sanity check:

    Your graphs indicate that Mastery is more valuable at 40% health than at 1% health.

    Your graphs indicate that Mastery is more valuable at 100% health than at 60% health.

    These are both clearly incorrect.

    The problem is that you’re calculating and graphing “Mastery Per Healing”, rather than “Healing Per Mastery”, as you describe. Your formula starting in “Crit v. Mastery” AP19 should instead be “=(C19-$B19)/($L$5*AP$18)”, and similarly “=(C90-$B90)/($L$76*AP$89)” for your AA affected table starting in AP90.

    The breakpoints you listed (40-46% and 64-66% for AA and non-AA heals respectively) are accurate.

    The correct graph for this data, according to your spreadsheet, should look more like this:

    It’s a far less exciting graph, to be sure, but at least it’s accurate :)

    • You, sir, are absolutely correct–those charts are misrepresented! (/hangs head in shame) Somewhere between making the graph, labeling it, and then writing the narrative, things went awry! But, I’m happy to know that the data is solid.

      So first things first, I’ll get those graphs correctly labeled and described. They are Mastery per Healing, not the other way around. That way they can return to being relevant to the conversation.

      And, I’ll add a second chart to illustrate what you linked above–the Healing per Point of Mastery. (I like that you combined both AA and non-AA on the same graph, so I’m going to stick with that).

      Thanks again for the peer review and the eagle eyes. I do appreciate it. :-)

  11. Aanzeijar

    I’m glad you finally stop treating crit like the ugly stepchild. It has been so annoying arguing against the “mastery is everything” mantra.

    And for the ” Whereas a theorycrafter might be inclined to think that the data would behave as if it had a single pivot point”

    • Yes, I’m aware of the discussion that went on at the start of Cataclysm re: the interplay of Mastery and Crit. I’m even familiar with the thread where the “math teacher” that Jynus is referring to presented his original analysis:

      … because it’s what I built on when I started writing my Cataclysm post on Mastery ( Drhay was an incredible resource and very patient with my attempts to wrap my non-math focused brain around why the breakpoint was moving.

      In the larger perspective though, what makes the difference now is that those “large quantities of Mastery” that Jynus is talking about in Cataclysm are now a reality in Mists with stat inflation. So looking at 4000 available stat points is actually relevant to our discussion this expansion, where it wasn’t last time around.

  12. Mario

    Very interesting Vixsin! And the site is all shiny and fresh, nice layout too. I love it.

  13. Healakong

    I’ve been preaching this to my guild and other Shammys since the very beginning of cata.. Mastery/crit is the way to go because of riptide and hst

  14. [...] stats are good under different conditions. Vixsin did a good analysis of the idea by comparing Mastery and Crit in early Cataclysm (linked: updated version for Mists of [...]

  15. Firestyle

    I just wanted to put in a word of thanks for all the hard work you put into this. It really is a blessing for the community. Being able to come here rather than try to reproduce this thought myself is absolutely wonderful. Much appreciated!

  16. Raahe

    Thanks for the informative post! In all honesty, the math is a bit above me, but I’ll bite and comment anyways :)

    So, if I’m understanding this right, then the correct stat priority (secondary stats) should be something along the following lines –>

    Spirit (to a level of comfort)*
    Haste (to easily achievable break point)**

    If that’s the case, this will make gearing and stat allocation a very interesting exercise.

    *With the interplay of crit and regen, would high levels of crit somewhat negate the priority of spirit?
    **Depending on gear inflation, this may change and if we’re just short of a break point it would raise the priority of haste.

    I’m sorry if I missed the point. After 6 years of playing a druid (mostly feral, some healing) I’ve just recently switched to a shaman.

  17. [...] Life In Group: Crit vs Mastery – A discussion on benefits/drawbacks of crit vs mastery. [...]

  18. Embermoon

    Vix, I love your analysis, sir.

    Ever since Blizzard made the mana-regen proc from water shield a baseline ability, I figured that the intention had to be for crit to play a fairly prominent role in shaman mana replenishment.

    I’ve seen some folks who are absolutely insistent, that 10k spirit is necessary, for resto shamans (in entry-level raid gear) to heal through mogu vaults content. My suspicion is that too many people are likely taking too much damage, and pulling on the healers to compensate, relegating the healers to a state of constant heal spamming.

    Also, @ Firesytle – /salute to a fellow Duskwoodian!

  19. Nedezak

    Even though crit and mastery are very close in throughput, i still believe mastery is quite favorable over crit. The reasons for this are quite simple:

    1 Crits often cause overhealing, decreasing the effective healing of a crit

    2 Crits are not reliable, Example: Crits are chance based, even if you have say 10% crit chance it does not mean on a fight you will crit 10% of the time. Theres a good hance you crit 8% or 12% of your heals.

    3 Im not sure on this one, but I am not sure ALL hot ticks can crit, though all hot ticks do benefit from mastery for sure.

    Further i would like to note that spirit is in 25mans always preferable over crit and mastery, because of the mana tide totem mechanic

  20. [...] I will not get into our mastery too much here, because if you would like a more comprehensive look, Vixen does a great job here. [...]

  21. I’m having some difficulty finding a “mastery soft/hard cap” I’ve heard it’s useless to have your mastery any higher than 50% as a restoration shaman… Between AskMrRobot, Noxxic, and Icy-veins I’m at a loss as to what to do… Each site tells something different.

    I want to increase my healing output… not sure where to look/turn. Any advice would be most appreciated

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>

CommentLuv badge