Jov sez: That’s great, now work on your overheal

June 9, 2009

Back in Vanilla-WoW, the dividing line between the “good healers” and the “bad healers” wasn’t usually so much a matter of keeping people alive (since, apparently, even bad healers could be carried for that) but that little overheal number.

TBC, with it’s limitless regen and bottomless mana pool seemed to take the stance “overheal doesn’t matter as long as your assignment lives and your mana holds out.” The TBC attitude has definitely carried over into Wrath raiding, and with encounters like Naxx, it was easy to slip into a zoned-out buttonspam while you smacked around the various loot pinatas for purples.

In short, it made us sloppy, where we’ve a lot less room for sloppiness in Ulduar and, presumably, moving forward.

What is Overheal?

Overheal is any healing which occurs over the healing needed by the target (effective healing).  This can be caused by spell selection (which is pretty easy to control if you’re paying attention) or crit (which is somewhat more difficult).  So if Roguechick is down 4k health, your flash hits for 6k and your greater hits for 12k, your flash will have lower overheal (6k heal – 4k deficit = 2k overheal vs. 12k heal – 4k deficit = 8k overheal).

Why Overheal?

Overheal is symptomatic of several things, meaning there’s no one answer.  It could be there are too many healers for the encounter (leaving 8 healers fighting to get their heals off first because the fight only needs 7).  It could be your healers aren’t familiar with an encounter so are reactively healing (4 healers going “oh crap, that mage took a ton of damage *healheal*) rather than proactively (I know Mage is gonna take a big clump of damage here, so I’m going to pre-shield/ProM/get flash ready.)  It could just be the spammy nature of some of the fights (And in P2 Mimiron, I do nothing but spam PoH on g2 and 3).

Some classes and roles are more inclined toward overheal in general.  Tank healers, specifically (pallies and Disc priests) tend to face situations where tanks can be gibbed at a moment’s notice, so tend to adopt the spam-heal-and-let-it-land approach, and often see very high overheal numbers (over 60%)  It’s currently a fact of the game, but at the same time, that’s 2/3 of your mana wasted.

Why not overheal?

As I somewhat touched on above, mana is the primary issue.  Through potion sickness, and the regen coefficient nerfs of early 3.x content, to the more recent hit to OoC regen outlined in Zusterke’s posts of last week, we can’t count on the bottomless mana pool sticking around.  I’m not meaning to turn into a “Blizz is hitting healers with the nerf bat” poster, but I get the strong feeling that this is a trend which will be continuing for some time.  I feel this is backed up, in part, by a post from Ghostcrawler just 3 days ago.

We want healing to be less spammy and more deliberate, but that won’t work until overhealing matters. To get to that point, mana regen has to matter but the risk of the tank dying in two boss hits also has to be chilled out.

In other words, tank healing is likely to get less finicky, but the nerfs to regen aren’t finished.

But how do I stop?

First of all, I don’t want to come down unilaterally hard on overheal and say “all overheal is bad!!!11”  For the most part, as the game stands currently, any overheal that your gear can support isn’t by itself bad.  In some situations (the above-mentioned tank healing, for example) it’s a requirement.  What the overhealing metric is good for is giving you a direction you can improve on.  Yeah, it’s great that you’ve officially gotten Freya on farm; now work on better spell selection and less zoning out to whack-a-mole and getting your overheal below 40%.


  1. The main question is: can damage be healed more mana efficiently?

    In TBC (and pre-tbc) we had downranking and mana efficiency was a matter of chosing the right rank of the spell “while sticking with the exact same spell mechanic”. Mana efficiency was a separate concern, handled after chosing your spell-type. It was controllable and thus overhealing was a good measure for the quality of your spell rank selection.

    In WotLK we do not have downranking. Mana efficiency is mixed with spell selection and spell mechanics. Therefor it becomes an infinitely more complex concern since we have to mix spell mechanic and mana efficiency. Overheal does not reveal whether you made a good choice in spell mechanics. That is why I deem overheal an outdated metric.

    Healingmeters fail to capture the data we need to measure our mana efficiency. What if I get a 50% overheal on GHeal, but it was an ‘acceptable’ choice or risk to save a tank in trouble? What if I used it very mana efficiently in other situations? What if my spell was free (SoL, inner focus)? A free Binding Heal still has a bigger chance of proccing holy concentration than FHeal or GHeal. The information we need to measure our efficiency is not available in any meter I have seen so far.

    If I would like to measure my mana efficiency, an extremely simple meter should just measure our total mana consumption.
    Raw Healing done / Mana consumption = raw mana efficiency (did you chose efficient spells?)
    Effective Healing done / Mana consumption = effective mana efficiency (did you apply these spells properly?)

    But many meters even fail to count the amount of spells you cast and only show the spells that hit! Yet it would be far more sensible than overheal. That is what I think we need. Untill then, we’re stuck with overhealing as sloppy metric. Self-analysis based on overheal will not push a healer to the limit.

    That said..
    I aim at low overheal myself 😀

  2. This keys in to tanking style too. If you know a little about your tank you can figure out what amount of overheal is to be expected. A higher avoidance tank is going to result in more overheal. A tank with more Effective Health is going to have less overheal, but might still require more mana to keep up.

    I’m not sure how GC think we’re going to make healing more deliberate than it is now unless more healing classes get something like PW:S.

  3. Paladin here, but I think this is something that needs addressing for all healers.

    Without the ability to downrank, overhealing becomes a huge problem, especially for classes with only a few healing abilities.

    Choosing a smaller spell for more mana to reduce my overheal is dumb. And there’s a HUGE difference between a Flash of Light and a Holy Light in terms of amount healed – and not that much difference in casting time.

    As the saying goes, when you only have a hammer, every problem starts to look like a nail. There’s no longer decisions about how much healing you need to do – especially on the tank. Since 90% of the time Flash of Light just won’t cut it? Welcome to overheal, population me (especially when you add in splash heals). I think Zusterke said it best that overhealing is a stat that doesn’t really have a lot of application in today’s raiding environment.

  4. Jove… =(

    Yes, there are fights where you have to watch your mana. Vezax of course comes to mind. But even on that fight, you want to provide as much healing as possible.

    Without going into tons of math and analysis, on any fight where healing actually matters, we cannot reactively heal. If we can’t reactively heal, then overhealing is not only necessary but beneficial, as it potentially reduces the chance of our target dying, and is not actually harmful, as long as we don’t run out of mana.

    So the better way to look at this is, heal as much as you can, without running out of mana, and don’t worry about overhealing at all. Overhealing only matters if you’re running out of mana, and then you should just heal less, or more efficiently.

    This is the short, no-math no-analysis version, but that’s the general idea.

  5. Regarding your tweets:

    I believe overhealing is not something you want to “work towards” reducing. I realize this is something that matters more for paladins than for priests, but think about how you’d evaluate the skill of your healers (especially output), and decide if that’s compatible with trying to reduce overhealing.

    It may be that that alters your evaluation instead.

    Finally, if the crash and burn tweet is about me, let’s talk.

  6. Hmm, I may not have made that clear in my first comment but I do not agree with the “overheal doesn’t matter” logic. It is, undeniably, at all times, sensible to work on your efficiency. If only to prove yourself you have too much regen and could stack more throughput or train yourself for the fights where it matters or have that reserve of mana when shit hits the fan.

    The most prominent example that comes to mind is a T5 Leo fight in TBC where an idiotic priest spam flash healed the entire fight. Halfway the fight 2 healers die and taking the healing up a notch. Despite his innervate, his spriest, fiend, shammy and chainpotting he ran oom.. which essentially meant another healer down and left us in an unsalvageable situation. Ofc.. he made sure everyone else was to blame. It may help to know that I healed as much as he did… without pots, fiend, spriest, shaman, innervate. If the priest would have worked on efficiency, we could have made that fight.

    Efficiency must, at all times, be a goal to work for. Not doing it… is doing exactly what Jov sais: ‘becoming sloppy’. The problem I have with overheal.. is that it fails measure it accurately. A BHeal with 30% more overheal than FHeal is more mana efficient. A PoH with 50% overheal is more efficient than FHeal with 16% overheal. Does 40% overheal on PoH mean I used my PoH badly or does it mean I managed to ‘outscore FHeal’?
    That is why overheal fails.. not because “it doesn’t matter is your mana can sustain it”.

    I may sound a bit radical about it I suppose.. I just have seen so many priests lavish in their regen when the raid is stable that they forget how to keep up when trouble is around.

  7. My real position is slightly more nuanced than “overheal doesn’t matter.” My position, summarized, is “when reactive healing is not possible, you should overheal as much as your mana allows for.” This generally is a tank healer position, as I can’t actually think of examples where raid damage can’t be reactively healed (and, consequently, where you would want to do things like pre-cast group heals, or pre-hot raid targets.) On the other hand, I would like to see raid healers helping on the tank when they have no raid targets to heal, so I feel this applies to every class.

    Here’s a example. You have a boss mob like Steelbreaker who’s swinging for 25k (or more if you take him down later) once every 2 seconds on a 40k tank. You can’t see the first swing land and start casting, because either your spell will land and it won’t be enough (no class has a consistent ~1 second (given lag/reaction time) 10k heal, and you’d have to heal for at least 10k to move your tank into safety range.) or it won’t land fast enough (IE Holy Light/Unglyphed Healing Touch/Greater Heal)

    The only solution is to pre-cast heals. Priests have often used cancelcasting to achieve this end, which is a better solution. However, it’s still not quite as good as letting the heal land, assuming you have enough mana. (because there is a chance that the hit will land between when you cancel and when the heal would have landed)

    Now, however, I also have to point out that there is a point where overhealing is useless. Theoretically, it’s where the amount of overhealing done is such that there is a 0% chance of your target dying. For example, if the raid lands 25k of healing every 2 second period on the tank, there is no point in healing more than that. But rarely can you calculate periodic healing in exactly that way. Generally, it is safer to say “heal as much as possible.”

    It all comes down to probabilities. You want to minimize the chance that your target dies. In general, you do that by healing more.

    The “what if other healers die” argument is an interesting argument, which I mostly refute by “tell your healers not to die.” 😉

    More seriously, it’s again a matter of probabilities. If you’re holding back mana in case a disaster happens, and by holding back mana you increase the chance of your target dying, this may not be a good tradeoff.

    Your argument that overheal is not an accurate measure of efficiency is an excellent argument, and something I hadn’t considered before. I’ll put it to you another way though. Is overheal reduction something you want to work for if it increases the chance that your target dies? This is the real metric, imo, not some 1 – (effective healing / total healing) measure. You’re trying to increase the chance your target lives, not improve a score on the meters.

    Unfortunately, “chance your target survives” is, well, difficult to measure.

    • Hmm, it sounds to me like your position isn’t “when reactive healing is not possible, you should overheal as much as your mana allows for.” but rather “preheal (with risk of overheal) as much as is needed to ensure survival”. That makes perfect sense! And this applies on both tank healing as well as raid healing. Even in raidhealing we precast heals:
      – XT Tantrum
      – Loatheb
      – Vortex
      – Grobulus Injection
      – Decimate
      Heck.. I even precast Abolish Disease at Heigan. It only becomes more apparent in Tank Healing because it requires more ensurance… a hell of a lot more.

      Efficiency is not the healers only concern. If that were the case I would just heal using bandages and be the most imba player around (not that anyone said so, I’m merely trying to acknowledge your point).
      The main priority of a healer still remains to ensure survival. I don’t think anyone disputes that. But can the same level of ensurance be delivered more efficiently?
      I recall in TBC that I spammed-and-let-go GHeal rank 1 on my tank. The heal was relatively small and cheap, so my overheal was small. IF the tank would take a heavy blow, rank 1 was enough to ensure survival of a 2nd blow. This would buy me the time to recover the tank in time for the 3rd blow. Ranks 2 and 3 would not make the difference in ensurance but they would in manacost. The extra mana saved being my ensurance that I would last longer or have enough backup to rectify a bad situation.

      Ofc, this scenario does not translate so well to WotLK. But the goal is: provide the level of ensurance you need as efficiently as you can. This isn’t just a healer thing… but applies to pretty much everything in life: deliver the same job (ensurance included) with as little resources as possible.

  8. When I’m running full raid buffed (Paladin) my HL is about a 1.6 sec cast – maybe less (with Lights Grace and LotW up)- on a crit I’m getting back 60% of the mana cost of the spell, but due to gearing am only paying 80% on a non-crit. My FoL is about 1.1 sec – and returns a slightly more efficient ratio, but when you compare 15k crit to 5k crit – and I’m MT healing – I’m going to overheal.

    My overheal really goes overboard when I’m the only pally in the area tossing up JoL.

  9. Aaaand pallies fail at reading comprehension. Stop getting your metal panties in a bunch and work that big int score before coming at me with torches and pitchforks.

    1) Priest blog. My statements can be construed to apply to priest healers. This isn’t a general healing blog, this isn’t a pally blog. The spells I mentioned are priest spells, and there’s no reason whatsoever you all should be taking this as directed at anyone but priests.
    2) I specifically state in the post that MT healers are going to face overheal vs. tank being gibbed and SHOULD choose overheal. MT HEALERS ARE EXCLUDED.
    3) I also say that with the way the game works now, eliminating overheal isn’t going to happen.
    4) I only point out what GC said, and posit that if more mana nerfs are incoming to the point where we’re not gonna be able to maintain facerolling PoH mindlessly as we can now, maybe we better start thinking again about better spell selection.
    5) I don’t say overheal is bad, I say it’s something you can work on.

    Jeesh, lrn2read, people.

    (My comments not directed at people who have more to say than “I’M A PALLY AND YOU’RE WRONG WE OVERHEAL IT’S WHAT WE DO OMG”)

  10. *Hands Jov some Vicodin*

    Shhh . . . good Priest blogger! 🙂

    Clam, centered, karma, and all that junk.

    I liked this post. Overhealing is a good metric to look at, especially if you are a new priest (emphasis)! I never really considered it too much, I knew it was bad but now I see it is situationally bad. Sometimes good and the only option on occasion. Thank you.

  11. THANK you! Thank you thank you thank you. I am SO sick of people posting heals per second and asking me things like, “why are you only healing 1K per second?!” in a fight that wasn’t that tough. (um, because that’s what was needed?) I once had the other healer in a 10-man raid post a comparison of our total healing done, and I was way below his numbers. Translation = crappy healing.

    But looking at the overheals, he was about about 50%! Fully half of his heals were healing nothing. Wasting mana, cast time, and, and, and…. harumph. (I’m a shammy; he’s pally – so maybe he was fine with the overheals, dunno)

    I’ve actually asked a few veteran players (the 4-year types, vs. my 1 year of playing) about this and both said, “oh, don’t worry about overhealing. It’s fine.” I’m thinking, “that’s why you always run out of mana and I don’t, I guess.” Although, to be fair, one really was an awsome (shammy) healer.

  12. The problem with looking at overhealing in a vacuum like that is you don’t have the big picture (particularly with class differences Lantanasham). It goes back to the things Zusterke describes above, and actually to a conversation I was having with a fellow guildmate last night about DPS meters.

    Any of these stats in a vacuum aren’t going to necessarily give you the information you need on whether those were good choices or bad choices.

    From Vanilla to TBC, the face of healing changed a lot. People weren’t stuck in roles simply by being X class, and we all had a lot more spells and tricks at our disposal. The game also changed. Wherein I was watching my overhealing in MC/BWL/etc to make sure it was often under 6%, in TBC, if you held back your healing too much, without the buffer of 10 other healers, you ran the risk of your assignments dying. Regen also changed, so some of the things in Vanilla we did in order to keep from running out of mana simply were no longer necessary.

    The game also changed with Wrath. And it’s probably going to change again, because Blizzard is trying to design healing to be less “lemme hit X spell over and over” and be able to heal just fine. They don’t want priests casting CoH over and over and being able to cover a whole raid.

    I played a holy priest exclusively for Vanilla WoW and most of TBC. Wrath healing has returned some of that sense of urgency to healing, where you have to make decisions, as opposed to spam X spell as fast as possible, like some of the TBC fights became.

    We should be trying to be the most efficient healers we can. That’s an extremely good goal. Sometimes that means that we’re going to overheal a lot to keep our charges up. Spell choice, raid damage, etc., all play into that, where you will run the risk of overhealing by using the best spell for the damage that’s being done. There’s obviously a difference between the priest that uses PoH as their sole mode of healing versus someone spamming GHeal to keep a tank up with large burst damage.

    And yes, there are a lot of sloppy healers out there who think Naxx healing is the way it should be. But no stat within a bubble is going to solely make you a gold star healer or a fail healer, just like topping the DPS meter doesn’t necessarily make you a better raider. I can easily top a healing meter as a holy priest. I can also probably make my overheal very small. But overall, I want to make sure I’m healing the best way I can. From having to heal recently (I play shadow now), mana can be an issue again…and it’s exciting to have to think about healing again.

    • This is the perspective that I try to cling to, especially when I’m DPSing, but that is usually so tough in the face of, “DPS meter! DPS meter!” I was elated to really, truly realise that things like my buffs, Bloodlust, and actual *movements* and choices really did/do make a difference.

      It drives me crazy when people rail against those with lowe DPS, and use that, or heal numbers, as end-all-be-all measures of worth.

      Besides, for me, whatever makes the raid *fun* is really the point. Even if we do a terrible job, sometimes it can be really fun.

Comments are closed.

%d bloggers like this: