Healing Theory, Part 4: Comparison of No-Cooldown Heals

Previous posts in this series can be found here.

The background and motivation for the heal comparison I did here is in the previous post; I recommend you read that if you haven’t. The goal is to compare all of the “filler” heals of various classes, although I may be defining that a bit differently from what you expect. It basically means, anything you can keep casting nonstop. The idea is is that you’re limited in how many Wild Growths or Penances or Healing Rains you can cast due to their cooldown. So if we put aside all of those spells, what we’re left with the spells that are primarily controlled by mana–they’re what to use to turn most of your mana (everything that’s not used on your rotational spells) into healing. The efficiency with which those spells can lever a point of mana into healing output is largely what determines the marginal value of mana for your class.

The goal is to focus on a specific set of heals that’s somewhat limited in scope (i.e. not building a full spreadsheet for every class’s rotation) but is informative on an important cross-class issue, one that’s motivated by the other recent posts I’ve written. The primary result of this is finally having some concrete sense of the relative value of Spirit, but there are other conclusions to be found as well.

The Data

It’s a bit awkward to present spreadsheet computations in the post. I’m going to give a link to the sheet I was working in, and a screenshot (but you can’t see the formulas), along with a lot of explanation. I’m definitely interested in feedback from class experts on whether I left out anything important. It was a lot of work to try to look up everything I had to take into account for each healing class to give a meaningful and accurate account of the efficiency of a single spell. I tried to pick up the most prominent spells for each class (and some that I was just curious about), but should be able to add anything people suggest that fits the concept.

Spreadsheet download: Dropbox (updated 5/23)

Screenshot (5/23):Heals

The basic methodology is to find the efficiency of a heal by dividing its spellpower coefficient by its mana cost. Mana cost is expressed as a %, with Priest and Monks being corrected down by a factor of 5 afterwards to account for their different starting mana pool. Spellpower coefficient is a good metric for relative heal strength since, as discussed in Part 2 of this series, the base heal and the coefficient are closely proportional on all these spells. For the moment I’m not including set bonuses anywhere, since only some classes get bonuses that affect the heals I’m looking at here.

The “coefficient” and “cost” columns are taken from the spell data, with various factors added in as needed, as discussed below in the class-specific notes. “Ratio” = coefficient/cost.

The “class” column is for passive class-based modifiers, as well the Priest/Monk mana correction.

The “C/M” column is a multiplier to account for crit and mastery. If you download the sheet, you can change the crit/mastery values at the top and this will change appropriately. The Monk mastery is handled in the “coefficient” column instead since it’s spellpower-based.

The “Total” efficiency = Ratio*Class*(CM Adjust)

Class-Specific Notes


Note that since Rejuv is a HoT, it has significant overhealing, and also requires you to have sufficient targets that need healing in order to chaincast it. The overhealing issue is somewhat ameliorated by the fact that it charges Wild Mushroom, giving a second crack at making use of the same healing.


Fully understanding the Monk methodology requires reading the last post. But it’s a slightly unusual case. I know Monks don’t actually use Surging very often, but this mostly because they don’t generate off-cooldown Chi very often (Surging is somewhat interchangable with Jab for this discussion). Chi comes in a mostly cooldown-determined rate from Renewing Mist, Expel Harm, and filler use of Soothing Mist. But if you want extra Chi or Uplifts, you’re going to pay for expensive Surging Mists or Jabs, and so the efficiency of that option is quite important.

I almost didn’t want to include the Soothing Mist row, but I think people would be confused by its absence since it’s, commonly speaking, the Monk “filler”. Soothing Mist doesn’t precisely fit the paramaters of this study because it generates Chi in a way that’s very rate-limited, even if not literally a cooldown. So it’s misleading (in fact, incorrect) to read the high efficiency of Soothing Mist as saying that Monks can efficiently turn excess mana into healing. Soothing Mist use is limited only by time/mobility, but the marginal point of extra mana is going to be spent on instant Chi generators (the next revision of this sheet should include a Jab-Tiger Palm entry).


The most important note when interpreting this is that Spirit has a very different meaning to Discipline due to Rapture. With Spirit giving up to 1.5 mana every 12 seconds (0.625 MP5) in addition to its usual 0.56 MP5, the marginal value of a point of Spirit is roughly doubled. If a Disc Priest is crisp with her Rapture use, that added mana return can be cashed out into healing spells like Prayer of Healing, no different from any other mana. A strong case could be made that it would be most informative to multiply Disc Priest values by roughly 2 to account for this in the chart (but currently I have not).

Smite is worth highlighting as the only heal in the entire list which is smart. So its practical efficiency should be higher than it appears due to low overhealing.

I included the pre-5.2 PoH, just for context–so people could compare current versions against the spell that upended the healing world for two patches.


Nothing too complex here. The Holy Radiance combo is very strong, unsurprisingly, but only works if you have a clump of people to heal. Otherwise, worth noting that a Divine Light on the tank and an Eternal Flame on another target is much better than a Divine Light on the off target and Beacon healing on the tank.


  • Class modifier is 25% from Purification.
  • Chain Heal gets an extra 25% for Riptide.
  • All spells have a chance to proc Earthliving. 4*10% for Chain Heal, and 20% for Riptide and Greater Healing Wave. The Earthliving HoT uses 5 ticks.
  • All three spells are discounted for the chance of Resurgence.
  • Riptide and GHW add 30% to crits for Ancestral Awakening. Technically a slight overestimate right now since the Earthliving HoT is included in AA.
  • Riptide uses 7 ticks.
  • Deep Healing is valued at 0.75% per point of mastery, corresponding to an average heal target at 75% HP.

Despite the mess of various procs here, not too much to add. Best I can tell, Glyphed Riptide comes out on top of a somewhat lackluster slate of things to spend mana on in between Healing Rains. (Healing Wave would be higher but I generally didn’t include the weak heals for now).


I’m going to keep the interpretation sparse for now, although the numbers will hopefully inform any upcoming discussions. I also want to see if any significant corrections or additions come in from readers. One interesting note here is simply how much work goes into taking into account all the various procs and passives that apply to even simple heals so that an apples-to-apples comparison can be made across classes.

This is our first in-depth look at a balance-related topic in this series. Where it might be most informative though is in helping understand gear choices within a class. Going back to main point of Part 3, the more efficient your spammable heal is, the the more Spirit is worth to you. One class has always been known to be a Spirit lover: the Disc Priest, due to Rapture. However, an important piece of the puzzle is their strong spammable heals even before taking Rapture into account. Beyond that, Druids and Holy Priests have solid uses for Spirit, as well as Paladins who are really in a position to leverage Holy Radiance. Shaman and Monks seem to have little use for it; in fact its utility looks somewhat dubious for Shaman even if you consider a significant % increase to the value of Spirit due to Mana Tide.

But for now I mostly want to put the numbers out there and see what people think of them.

11 thoughts on “Healing Theory, Part 4: Comparison of No-Cooldown Heals

  1. Question for you – when dealing with smite, how did you account for divine aegis? Ditto illuminated healing. Is that simply factored in as crit/mastery?

    Tanks focus on effective health. I’m starting to be more interested in effective healing. Not all heals are created equal – absorbs get first crack at the incoming damage, and smart heals target the lowest raid members. (Have you seen anything anywhere about which absorbs get prio? Is it time based or spell based? divine aegis vs illuminated healing vs sacred shield vs trinket shields)

    This is outside of the scope of your interest in this post, but the mana used in dispelling disease/poison/curses/magic effectively heals the damage the debuff would have caused.

    For example, as a resto druid healing with a disc/hpallie combo, my overhealing is astronomical. The 26.76 for rejuve is closer to 15 or 10 effectively for me, because my hots get clipped most of the time in almost every fight. Stacking more spirit to cast more rejuvenations isn’t as effective as casting fewer rejuves that hit harder in the first 3 ticks over the course of a fight for me.

    The effectiveness of my holy paladin raid member’s spammable spells is based partly on direct heals and partly on illuminated healing, and the overhealing on absorbs is very low. If I compare divine light (non tank) to regrowth, if we both have 40% overheal on the direct heal portion, but his illuminated healing has 0% overheal, and my living seed has 62% overheal, his effective mana to results ratio will be a great deal better than mine.

    Wandering off topic, I think that the discussion of healing best practices needs to include the mix of healing classes and tanking classes.

    • “(Have you seen anything anywhere about which absorbs get prio? Is it time based or spell based? divine aegis vs illuminated healing vs sacred shield vs trinket shields)”

      Hi Berry! :)

      For most shielding effects, shields are prioritised based solely upon their size. So if a prot pally, for example, has a 20k Illuminated Healing shield, a 30k trinket shield, a 40k Divine Aegis, and their own 100k Sacred Shield, and takes a 150k hit, the damage first wipes out those three small shields, and leaves their Sacred Shield with 40k remaining absorption.

      If that prot pally took a 40k hit, that would leave them with a 10k trinket shield, a 40k Divine Aegis, and a 100k Sacred Shield still active.

      So the smaller the shield, the more likely it is to be fully effective, because it gets prioritised first.

      The only thing I’ve found so far where the order changes is with Blood DKs’ Blood Shield; it always gets used first regardless of what other absorption effects are on them. If it didn’t, I guess Blood DK’s active mitigation would feel less useful in raid compositions that are heavy on absorption healers.

      • Buh, wtb ability to edit comments. When I wrote “If that prot pally took a 40k hit…”, I meant “(instead of 150k)”, I just didn’t say it. :)

    • DA and IH are both just handled as additional healing from crit/mastery, yes.

      I definitely do want to figure out how to handle this issue of shields and competing overheals more. I’m not sure at the moment exactly where I’d be going with it though.

  2. Hamlet, interesting numbers there, and as I said on Twitter before, it’s nice to see a more mathematical approach to something I have felt intuitively for a long while (that Resto Shaman has very little ability to turn spare mana into additional healing, since our spammable heals are so weak compared to our cooldown-limited heals).

    I’ve been thinking about working on a post about the legendary meta gem and how it provides additional benefit for classes that can either convert overhealing into meaningful healing (Holy Paladins with IH@~30% overheal recovery, Resto Druids with Rejuv>Wild Mushroom@150% overheal recovery) or into secondary resources that can be used to power further mana-free healing. The analysis is pretty similar.

    However, I’d quibble with you about the value of Resto Shaman Spirit. Yes, it doesn’t affect *me* very much to have that extra mana. But look at how much more healing my team can pump out if they drop their own Spirit and put it into throughput stats, and I keep a high Spirit and give them more mana to spend on their (now harder-hitting, because they have more crit and mastery or perhaps replaced spirit-mastery with int-mastery gems) already very efficient filler.

    I give each healer on my team an average of 2700 spirit if I use my MTT on CD. That’s 2700 stats they can shift into a few more percentage points of Crit or Mastery, and they can rely on it because I drop the MTT for them, not for me.
    They can turn that mana into far more throughput than I could provide if I had dropped my Spirit for throughput stats – because they can convert overhealing to meaningful healing, or they can generate secondary resources, and I can’t.

    Furthermore, on fights where I am limited by raid positioning, or the damage profile of the encounter, that Spirit gets put to even better use because these other healers don’t share my limitation, and I wouldn’t have been able to match their throughput even if I had dropped all that Spirit into throughput stats. (Heroic Dark Animus comes to mind here. I surely could have done more healing myself if I’d ran less Spirit/more throughput, but I don’t think even at perfect play I could have matched our Disc Priests and Holy Pallies.) And even running with a high Spirit build, on fights where significant damage occurs while we can be stacked up, I do very well, so I wouldn’t say that my high-Spirit focus is hindering my performance in those situations.

    Granted, they have to maintain enough Spirit to be able to stay afloat outside of MTT windows, since I’m not *truly* providing that 2700 Spirit 100% of the time. So they might elect not to drop the full 2700 into throughput stats. But even still, because other healing classes can turn extra mana into healing more efficiently than I can turn extra Mastery/Crit/Haste into healing, I find that it is more useful to my raid as a whole if I keep my Spirit high.

    I’m not willing to say that this is the way to go for every raiding Resto Shaman; I raid in 25-player format, with 2 Holy Paladins, 2 Disc Priests, and a Resto Druid. I suspect a high-Spirit build might cripple some 10-player Resto Shamans, especially in groups that 2-heal or that do not push new content as quickly as I do, or may be less of a benefit to a raid that has fewer Paladins and more, idk, other Shamans or Monks. But I just wanted to make the point that I don’t think Resto Shaman Spirit can be properly evaluated in a vacuum.

    • Good point about the legendary and classes that can store healing. I can see looking into that some.

      On Mana Tide:
      At the outset, I’m not totally happy if the conclusion is that the best allocation of a Shaman’s secondary stat scaling is to divert it towards being a mana battery. If it turns out that’s the case, it’s a very good argument for nerfing Mana Tide–at the very least, make it non-Spirit based like most other regen effects. That way Shaman could still provide that unique utility without an enormous distortion on their stat choice.

      But, moving from how it should be to how it is. The analysis in this post is actually quite informative on how to evaluate MTT, as it answers exactly the question you want to answer: how valuable is an added point of Spirit to the other classes relative to you.

      Marginally, 1 point of Spirit gives, on average, 16/180*2=0.177 Spirit to each other healer in the raid–a little over 1/6 of a Spirit. So say (just to make the thought process clear) you’re in a 10-man with you and a Disc Priest. Added mana is worth about 10 to you if you’re typically filling with Chain Heal (in this post’s slightly odd metric: %spellpower worth of healing / % base mana spent) and around 30 to the Disc Priest (who doesn’t get Rapture benefits from the MTT Spirit). So you can estimate MTT adding rougly (10+30)/6, or around 7, to the value of Spirit. So the value of Spirit has increased noticeably (from 10 to 17), but still not as valuable as it is to a Priest or Druid.

      In a 25-man, of course, the addition can be much higher. If you are powering multiple Priests, Druids, or AoEing Paladins, then the extra PoH’s, Rejuvs, and Holy Radiances that your Spirit allows “you” (your raid) to cast can result in it having a value of 30 or higher in the above framework. That makes you about as Spirit-loving as a Priest (although not dramatically higher than a Priest).

      The main point here is: the value of a Spirit fed into MTT is a fraction of the sum of the marginal value of Spirit for all other healers in your raid. Since this isn’t (yet) a full-blown Shaman spreadsheet/model, I reach no conclusion on how that ranks the stat next to crit or mastery in various situations. But the analysis here should give a good ballpark of the relative increase in the value of Spirit due to MTT, which should help you estimate how to correct for MTT in any Shaman model.

      • I was asked to stack more spirit to be a mana battery, but MTT is only 200% of your spirit now and not 400% as it used to be and I do not see the point of why I have to forego healing throughout for other healers. I think you need to also factor the less healing throughout you are doing, meaning the other healers have to heal more and thus consuming more mana. And then of course they will want your mana tide totem. However even if my logic is flawed and I am no math genius to work this out, in reality people do look at HPS and healing done, and raid leaders quickly forget you have less value in your healing as you forfeited that better stats to be avmana battery. I think Dayani comment is ‘nice” but I have faced criticism of my lack of heals as I have stack more spirit to appease the raid leader. I believe other shamans do not want to be a mana battery just to compound the fact that we are seemingly weak at the moment. In reality people do judge by HPS and healing done especially when people die. They equate low HPS/Healing done with “the healer suck it is why we died”. Theory crafting is nice, but I tried three raiding guild in 6 weeks, and guess what? Each one looked at “numbers”. Shamans can not afford to be a mana battery – we are already weakish and really do not need be in a position to be more so.

  3. Pingback: HolyCalcs beta, 5.2 PTR - Page 3 - Elitist Jerks

  4. What I found interesting about this post was the conclusion that Resto Shaman don’t have much use for Spirit given the relative inefficiency of our spammable spells. It’s odd because it’s entirely contrary to the gemming and gearing position of Resto Shamans since the start of the expac.

    It wasn’t until I went through your post a couple more times that it dawned on me why that contradiction exists, and it’s due to the one spell that you discounted at the onset–Healing Rain. HR is different from Penance and WG, in that it is NOT a cooldown-constrained spell, but rather it is a mana-constrained spell. Your supposition is, in lumping it in with WG and Penance, that HR is cast every ~12sec for the duration of a fight. But, in fact, this wasn’t/isn’t the case. Because of HR’s incredibly high mana cost (43.1% of base) it becomes the determining factor of our regen needs. Early in the expac, HR wasn’t nearly as sustainable as it is now because of available regen (from passive mp5, Resurgence procs which notably scale with crit, trinkets, and Mana Tide). So, the natural process was to stack Spirit (and Crit thereafter) so that we gained more ICR, and thus more casts of HR, which resulted in more throughput.

    But now, with gear and the introduction of the meta gem, Resto Shaman are in the situation of having enough mana to make the cooldown on HR more of a constraint than the mana cost. (Something Dayani talks about above–shaman don’t really have a spell by which to turn excess mana into healing). So, what you see from players in top ranks is … more haste. (Which fits with Shamans’ historical path to greater throughput at the end of an expac, as seen in ToGC/ICC, DS, and even in micro-increments each tier). This ties into something you mentioned on Twitter–the use of Glyph of RT. Prior to the meta, Shaman didn’t have sufficient regen to allow for pure RT filler. But as we add more haste and reach a point where we’re trying to burn as much mana as possible, Glyph of RT becomes a more viable tool because we can afford to use it a la Rejuv.

    Unfortunately, the imbalance between HR and our other spammable spells was further exacerbated in the recent buffs, which is why I think there are a growing number of Restos who have been arguing for future nerfs to the spell (along with other Resto Shaman cooldowns–something you broached a question about a week or so ago).

    • Yeah, that’s interesting. I was also just talking on twitter–probably just as you were typing this–about different rotations between HR and Unleash Elements, which is all generally predicated on a current milieu of constant HR use.

      A more detailed way of thinking about this that I’ve never laid out is looking at the value of Spirit as roughly piecewise linear. Imagine starting with 0 mana and doing 0 healing. As you start to be allowed more mana, you spend it on your most efficient spell, until it’s limited by cooldown, then you spend it on your next-most efficient spell until you can’t do that any more, and so on. And in then in practice there are rough edges when you’re near the boundaries, and for using other spells for practical reasons. Another approximation is that “on cooldown” often doesn’t mean precisely on cooldown, but rather, “the first time after the cooldown finishes that the spell would be effective”, which varies a lot.

      So really when I talk about cooldown vs. non-cooldown, it’s a slight abbreviation for: what spells that you cast with your marginal point of mana in the current gear/resource regime? A lot of the time that doesn’t change much. For Druids it will basically always be Rejuv–that’s what it is at 0 Spirit, and that’s what it will always be since it’s very efficient and has no cooldown. For Shaman, perhaps you do hit a cusp sometime early in this’s expansion’s gear range. But that’s why I emphasize so much the marginal value of stats at current gear levels. If Shaman hit a very meaningful Spirit breakpoint sometime last tier, the results as expressed in this post are still the most relevant right now.

      But the issue would still be interesting to look into more. Note that the end of Part 3 ( http://iam.yellingontheinternet.com/2013/03/15/healing-theory-part-3-mana-and-other-resources/ ), the two analyses I thought would be interesting were 1) the efficiency of spammable heals (this post), and 2) the mana upkeep required to maintain the rotational spells, which goes directly to this issue.

  5. Pingback: Healing Theory, Part 6: The Mana Economy | It's Dangerous to Go Alone

Leave a Reply

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