One of the reasons that healing done is very poor performance metric is that healing is limited by damage taken, not by how quickly you can heal. So if you take two identical priests and both have "enough" throughput to heal an encounter, then whoever ends up having more to heal will end up having the most "effective" HPS. Unfortunately, this concept is not well-understood by the wow community, but it is very important to understand if you want to make sense of healing done numbers from WWS or elsewhere, especially when comparing completely different raids.

One of the things I noticed looking over the WWS logs you linked was that your raid was taking less damage per second than the other priest's, so it isn't surprising that his effective HPS is higher -- there was more incoming damage per second to heal than you had. Proportionately, you were each healing a similar percentage of the incoming damage, so I see no reason at all to think there is any significance difference in your healing.

The only mistake you are making in my opinion is believing that the difference in HPS you are seeing in these WWS parses implies a difference in healing performance. 7100 might seem a lot bigger than 6300, but do you have any evidence that you could not have sustained 7100 if you were in his/her situation? I don't think the difference means anything. If you really want to have more HPS, you could tell your raid to screw up more often so you could heal their mistakes.

I've been struggling with this idea recently. Up until last week, I was modelling crits at a value less than 150% of base heal because more crit leads to more overheal, but when I was trying to figure out a good way to model that, it dawned on me that spell power *also* leads to more overheal, as do talents that increase your healing. Intuitively, I can see why you would think that small heals benefit more from crit than big ones, but what I'm wondering is if anyone can actually model mathematically 1) that they do in fact benefit more from crit and 2) how much more do they benefit.

Doesn't this depend entirely on the shape of the incoming damage? It might feel right to say that a 20k heal doesn't benefit much from crit, but if the boss hits your tank for 25k (heroic hodir), then you land your 20k heal, then he hits for 25k again, couldn't a crit easily mean the difference between life and death there? (sure this is pretty specific, but since noone actually has a 20k heal I don't mind getting a little specific) If you are on a fight where you are spamming AoE heals trying to keep everyone at full life because there are harsh damage spikes to go with raidwide damage clocking in at about 2k per second (hardmode 10 man mimiron phase 2), then aren't your small heal crits often going to overheal just because you want to keep everyone at 100% so it's worth it to throw away the extra 1k healing to make sure they don't get spiked out?

The only argument that is working for me right now that crit leads to more overheal is that crit is random. If we know our FH hits for 5k then we can cast it on someone who is down 5k and expect to heal them to full (or close enough) and not waste mana, but the crit mades us retroactively waste mana in some sense, and given the way some fights work, it might not be okay to just leave the person down 5k and wait for them to take a little more before topping them up. Healers who leave everyone at 90% to increase their efficiency aren't anyone's favourite healers. If fights had no random element then we'd win more, so random elements seem bad, but how can we quantify the badness that random adds to crit, and is it really better or worse for some spells than others?

Really this isn't stressed enough: you cannot evaluate a healer with numbers only!

It has been stated and repeated before... There's too many factors, including incomming damage, healing assignement, specs, other healers, healing style, etc...

I am just tired of all those people swearing by the meters... That's only the uncovered half of our job.
So, I'll throw my rant, just skip if you're not interested...



...HPS is nice and all, but if I wanted I could get awesome hps, and still let some dps die in the mean time... Then clueless people in the raid would look at recount and think: "Wow, this one healer's good, it must be the others' fault".
Of course, that's not what we want. Our job is to make everyone stay alive. Producing high output is only one tool among others to reach this goal.

Personnaly, I'm more of a trigger-happy healer. This means I often top the meters, because I heal a bit faster than my teammates. But, be it my heal, or someone else's heal who lands last and overheals, unless the target is in dire danger, it's wasted throughput from your healer team as a whole. (Assignement, crosshealing, blabla, that's not my point...)
So, understanding how your healing buddies play their part is a great step toward your personal efficiency.
but it involves knowledge of all healing classes, as well as an UI that shows a whole bunch of incomming heals/hots/shields... (and of course reliable teammates).

Although, the good side of playing the meters is that it emulates everyone to be at the best at their job.
Still, I know some healers who usualy hold back, and shine when the shit hits the fan. They appear at the lazy spot on meters, but that doesn't make them less skilled nonetheless.

Of course, sometimes all you have to do is to top off people. I find that boring ^^.
As I understand it, it's when:
- Your raid is skilled in avoiding damage. (Tanking, position, cooldowns, reactivity...)
- And/Or, other healers are doing the ungrateful jobs of saving lives. (This mean casting renew/FH on that lone guy who has the dot during tantrum, instead of throwing a PoH over a wild-growthed group... I'm exagerating a bit.)

Now please, don't feel insulted in case you already know how to play a priest, and just want to maximise your throughput. If you're curious about some other healer's data, look at the spells he cast, who he heals, and his overhealing.


To l337n00b:
Crit's randomness is even less of a bother when you aoe heal. Since the smart targetting from CoH, PoM, CH, and WG will mostly hit the people your PoHs didn't crit on.
And about FH/renew, I want the crit for the proc, not really for the throughput.
Strangely enough, the only drawback I found on crit is I when shield spamming as disc.

On the other side of the argument though, if someone is seriously lagging on the healing meters, that can be a sign they are doing something wrong. If you bring three holy priests to the same fight and one of them consistently heals 20% less than the other two with similar gear, that's a huge warning sign.

Effective healing isn't as perfectly correlated with skill like DPS is for damage classes, but it's still the best metric we have. Even if there's only a 50% correlation between healing meters and healing skill, that's really meaningful. In practice I suspect the R2 coefficient is much higher. Lets not throw out this useful metric just because it's imperfect.

This is a significant point.

Another difference in evaluating healing is your healing assignment and the other healers in the raid. No healing is done in a vacuum, you've got 4-6 other healers in a 25 man who each have their own class, assignment, style of healing and gearing. Add the reaction time of the DPS and tanks to avoidable damage, the particular tactics used for each fight, and other variables and comparing healers off a single HPS number isn't realistic.

Personally, I like haste. I like the fast GCDs and fast casting. I push to over 900 haste in a set I use for trash. I heal many fights in a 700ish haste set. Haste is purely a throughput stat though and gearing like this for long, mana intensive fights is not feasible. I often drop back to a balanced BiS gearset like Constantius describes to get the hybrid longevity/throughput component that crit provides. Fortunately, my guild has a loot system that allows me to build multiple gearsets for different fights and approaches to gearing.

I aim to end each fight low on mana, allowing for a margin of error depending on the fight. I aim to save the potion cooldown for recoverable emergencies.

I don't even begin to think my approach is the one true way though. My understanding of attempts to quantify the throughput value of haste and crit put them close to each other. This approach works for me but I know plenty of priests who favor crit.

I really don't think we'll be able to come up with a good model. Sure, we can use one model, plug numbers, a nice spreadsheet, and get some nice curves and answers. Problems will be that they wan't be accurate.

In very short, I agree with the intuition that crit on small heals is less overheal than in big heals. Because when you're healing a 4k, you're still likely to find someone with 6k to heal. When you heal a 20k, it can still happens, but it's less likely that your target needs a 30k heal.

However, spellpower leads about to the same overheal ratio. Yes, it is more reliable. But anyway, when a target needs 4k healing, and if we have a GCD free, we'll flash heal it. And since there is no more downranking, we don't have any choice, we flash heal. Or more precisely, we use a lot of different spells, but they depends more on the situation than our spellpower. POM if not on CD, COH if raid heal is needed at one point, POH/HN for groups heals, FH/Renew as filler or for single target.

The main reason one could prefer spellpower to crit, and I'm not sure it's a real one, it's because of it's realiability. We are healers, our job is to keep people alive, even when RNG is bad. So, when I have several choices, I tend to take the one with the lowest death probability. Crit in this way is more random, and we may want to avoid it (or more precisely, not actively stack it). But in most cases, there is anyway randomness for such "prediction". The death probability depends on future incoming damage, that are often. And there are scenarios where a crit could save someone, and a non-crit "bigger" heal would not be sufficient, as well as opposed scenario favoring spellpower. There are also other healers in your raid, that may (or not) heal your target. And even when everything is predictable, we quite never plan to heal someone just at the surviving limit, and even less frequently to let them low on health. As holy, I really don't think there are better stats than others. You just need a good mixed stuff, with pure regen, mixetd and pure throughput stats, and adapt it to the needs of the fight.


@Tedv: There are cases when different people with the same spec, both raid healing, can have different throughput. One of the holy priest of my guild makes a heavy usage of flash (up to 50-60% of his effective healing), and low usage of POH (usely 20-30% at most in POH-friendly fights). Knowing that he'll be catching low health target, I become more a POH spammer, rotating targeted groups. I end up with more hps, but if he was not filling the "low health responsability", even if he's not asked specifically to do it, I would use more fh to save them, and less POH to produce the raw need of hps. And I'm not sure that our added throughput / death rate would be better. We just informally splitted the raid healing task.

This was about SoL procs....
There may be an mistake in my understanding, but I thought that flash heal was a 1.5 sec cast. If the GCD (no haste) is also 1.5 sec then your throughput is always reduced by using an SoL proc versus casting flash of light as it is the same amount of time, same amount of base heal, but SoL can't crit.
I still love the talent, but more for instant casts (which don't actually save time, but can be cast on the run) and mana savings.

The question is how much Surge of Light reduces your healing. My point was that if you aren't casting the standard Flash Heal anyway, then the option of occasionally casting a free, instant cast version will only increase your healing.

Alternatively, you can think about it like this. If someone is hurt and you don't have surge of light up, you probably won't use Flash Heal on them. You'll probably just cast Prayer of Mending or Circle of Healing on someone. Maybe cast a Renew or a Prayer of Healing until you actually get a Surge of Light proc, and THEN give someone a free flash heal.

I'm suggesting that a Flash Heal that costs mana and can crit is not a very good use of a global cooldown. If that's not how you're using Flash Heal, then Surge of Light doesn't hurt your throughput. For example, on Freya 3 attempts last week, I cast a regular Flash Heal less than once per pull. Same for Thorim, Mimiron, and so on.

I don't follow your logic. If you don't feel that flash heal isn't worth casting without surge of light, why would it be worth casting with surge of light when it takes the same amount of time and does less average healing? If you prefer renew when you don't have surge of light, it seems like you'd also prefer renew when you do have surge of light.

Is it just a mana issue? If you prefer renew to flash heal for its HPS advantage, it seems like you wouldn't want to use SoL flash heals either, since they're even lower HPS than a normal flash heal.

Anyway, I disagree with your spell priority. If someone is hurt (rather than multiple someones standing together) and not poised to take automatic non-lethal damage soon (say, by tanking something or having a dot), than I'm much more likely to flash heal them than to renew or mending them. If a ground tremor could happen soon, for instance, they need to be topped off now, because a prayer of mending won't save them when it triggers from lethal damage, and if they might take damage or might not (say, on Thorim hard mode), I'd much rather get them topped off asap rather than moving my mending from one person to another or hoping my renew tops them off before they get nova/bolt gibbed. I usually bounce mending off the tanks on Thorim.

CoH/PoH are saved for when multiple people need healing, not just one. They don't really compete with flash heal in the decision making process at all. If they're appropriate, you cast them. If they're not you don't.

I find it really odd that you did Freya+3 and didn't use Flash Heal. How did you heal people with Nature's Fury or Iron Roots? What did you do (as happens often) with the one person who didn't get hit by your PoH due to range issues?

Wreath typically casts anywhere from 10-20% of his total healing as Flash Heal; some of those are obviously SoL procs, but he's definitely using it as a top-up spell. I use it extensively as Disc when I'm not PWS-spamming the raid in preparation for the next Tremor.

Sorry to keep nitpicking, but I again don't understand your logic.

Flsh heal does hurt your throughput. It is a very precise amount of healing with no chance for increase aside from slight scaling with spellpower. Regular flash heals that cost mana have greater throughput (via criticals) but cost mana; surge of light is purely a regen talent.

Can I see some parses when you cast virtually no flash heals that cost mana? I'm very interested to see how you heal.

I'm going to argue semantics here, but simply because it's mostly a regen talent doesn't mean that's the only bonus you get. There is a difference between instant and not instant cast heals.

SoL Flash Heal's and regular Flash heals take the same gross time of casting; but yes, you're correct as the health gets to them quicker. However, your HPS is still lower because they have no chance of a critical (no holy conc, no inspiration). The best benefit to a SoL proc is a free serendipity stack, in my opinion.

The mana, health arriving 1.5s (minus haste) faster, and serendipity aren't the only benefits. Instants can also be cast while moving.

Definitely, though a shield is a stronger *heal* while on the run. Don't get me wrong, I love Surge of Light and I won't spec without it. But it is a HPS loss, no matter how good the other benefits are.

Sure a shield is stronger, but you have pre-existing weakened souls to consider. Or the fact that you're generating a weakened soul when there's a disc priest in the raid and you prevent his bigger shield. Or situations where you need to move for more time than 1 global (which is sort of the same thing as a pre-existing weakened soul). Or situations where a 2nd target needs a heal during the 4 second PW:S cooldown, because you can't realistically get soul warding in a normal holy build.

In those cases, SoL is a HPS gain. I don't have numbers to back it up, but I would guess that these HPS gains outweigh any HPS loss from forced non-crits.

Depending on your spec/glyph choices then renew would be a better choice while moving.

Anyways, I wasn't trying to start an argument just wanting those who forgot to remember that Surge of Light is a regen talent choice.

Let's presume a situation where a healer would ordinarily toss a small, single target direct heal (a Flash Heal, a Nourish, a Lesser Healing Wave situation).

Holy Priests have five possible options for dealing with this situation. We'll discard two of them - PW:S and Greater Heal - out-of-hand. We'll note that a third - Surge of Light-fueled Flash Heal - is superior to other options, but may not be available.

Which leaves us either Renew or non-Surge of Light Flash Heal. No matter what choice we make here, we get the same number of Holy Concentration procs.

Which is why I now put 3 points in Mental Agility (18/53 spec) instead of HC. MA supports CoH, glyphed Holy Nova, PoM, and Renew -- the spells I actually want to cast when raid-healing Ulduar (the others being SoL, PoH, and an occasional BH). HC depends on frequent use of single-target spells with cast-times, which works directly against what I use Holy for in 3.1. On encounters where I am asked to perform a healing role that would require the type of spell mix that HC encourages, in my experience I can perform that role much better as Disc than Holy anyway.

Maybe this all changes in 25-man hard modes (which my guild is just starting) but I don't see why it would.


I agree. HC was designed for a 3.0 world that presupposes you only have one healing spec and that you thus will be Holy-specced in all encounters, and sometimes single-target healing. That was also a world where a 3.1 Disc priest's ability to spam PW:S/FH/Penance was undreamed of--Holy FH/GH spam was pretty much a priest's best option for single-target healing. HC needs a redesign to harmonize with what Holy has actually become. It could be as simple as letting CoH crits trigger HC (while maybe dropping FH from that list).

One could also add that each spell adds a Serendipity stack -- but SoL does it with no mana cost and as an instant. That is a significant benefit. If I have a 3-stack of Serendipity going in anticipation of something like a Tantrum, and it is about to expire, sometimes I will cast a normal FH when no one really needs it just to keep the stack going. I.e., it is worth ~600 mana to me just to keep that 3-stack going. Frankly the main reason I invest 2/2 in SoL is the free cost of building Serendipity stacks for PoH, not because I see a lot of GCDs where SoL or FH is the ideal healing tool for the situation.

A few idle observations from the last page or two:

1) Between PoM, FH, CoH and Empowered Healing, a Holy Priest can certainly find a higher percentage of their crits being effective healing than some other classes, especially since PoM and CoH have some smart logic.

2) Every now and then the raid will be mostly topped off and I'll be cast cancelling a Flash Heal on a tank anticipating damage (not spec'd for faster GH and target is PoM'd, etc). Without realizing it I end up using a SoL proc with 100% overheal. That is kind of frustrating Blame it on a stupid user I guess.

3) I have found the following priority when Spec'd for Renew to be optimal so far:

1) PoM up at all times
2) CoH when raid wide damage (get the CD clock going)
3) PoH if there is still 3 people in a party who can use most of it
4) FH if there is still someone left at 50% or less
5) Emp Renew if there are people at 50% or more and not in immediate danger of getting 1 shot

That results in ER being cast about 2/3 more often than FH due to targets quickly being brought over 50% from Priest AOE and Druid HoTs, especially with Test of Faith. With ER being critable and proc'ing other things as well, it's been better than expected and possibly optimal over a FH + Empowered Healing spec. /shrug

First off I must appologise if I'm posting this in the wrong thread. But I have had questions about CoH and PoH. I use alot of those while raid healing. With a double cast it can take at least half of our raids from about 40% down to full. But I've still been having problems fully utilizing these abilities. I find myself going OOM too quickly for comfort especially if the fight is taking more then 5-7 minutes. Right now I've been working with a hybrid spec and it's worked rather well for me so far but I keep getting nagged at for it because I didn't take a couple of the tallents here or there. I am dual spec Healing/Shadow and I am on armory on scarlet crusade under the same name I have on here. Unfortunately i've been at a loss as to where I could start with revamping my talent tree. Mostly cause I really like it. If anyone is kind enough to check it out and make some suggestions. Feel free. But until then I'd really like some help with my CoH and PoH problem.
thanks all.