deleted

I am not so sure sustained hps and peak hps work any better than longevity and throughput. Peak hps is ok, although it is similar to throughput. However, I do not see how sustained hps addresses a specific practical problem well.

My rather simplistic view is that gear can only help address two very specific problems you encounter while healing:

1. If you simply cannot churn out enough healing to keep people, that is a throughput problem. Increasing either your peak hps or your throughput seems to address that issue fairly directly.

2. If you can keep people alive for while but then run out of mana, that is a longevity problem. Looking at mp5 addresses this problem fairly directly, but sustained hps addresses it only in an oblique way.

It seems awkward and contrived to address what really comes down to a mana problem in a metric that uses hps as its unit of measure. While I'm sure that sustained hps is highly correlated with longevity and could be used as a substitute, I wonder why we wouldn't just look at mana directly, since that is what we run out of.

In other words, it makes sense to quantify a healing output problem with a metric in healing per time. By similar reasoning, it makes sense to quantify a mana output problem with a metric in mana per time units.

A lot of people have said things along those lines. Especially for discipine priests, I think random gear choices based purely on ilvl and brainlessly followed is a viable gearing strategy. It is probably better than those who come here looking to find out "What is better? Haste or crit?" and then stack one and neglect the other. Think of it as a strategy which avoids bad/misguided optimizations very well since by chance you are very unlikely to have a significant stat imbalanced.

Of couse I think it is possible to do better than random choices within the same ilvl. In a particular role, it might make sense to optimize a bit one way or another. But even then it won't make a huge difference.

However, even if what I'm saying is true and only gear ilvl makes any real difference, it won't stop people trying to optimize their gear. Based on all the discussion here and elsewhere about BiS items, it seems obvious to me that the quest for "the best" gear is an end unto itself. It doesn't matter if optimizing stats really helps because people just like doing it anyway...

And I don't want to be a grinch and spoil everyone's fun, so please optimize your gear if it pleases you. We do all this for fun, right? I just don't want anyone thinking he/she HAS to worry about that in order to play well.

Exactly. Your situation is constantly changing as a healer, so maximizing towards a particular common circumstance does not make much sense. It's a brittle strategy -- not robust. Instead, minimizing your chances of "breaking" in less common but possible scenarios may make more sense. Admittedly I am biased by being a biologist, since that is how nature operates (that is, "optimization" is attained by making extreme fatal traits go extinct more than by prospering extreme positive traits). Still, I would argue that as a healer your job is more to keep people barely alive under a variety of circumstances than to keep people alive "very well" under optimal circumstances.

Exactly. Once you hit a few key limits, the differences start getting smaller and much harder to quantify.

However, I really enjoy the metagame of gear optimization. I love all the tools and spreadsheets. I know at the end of the day that it might not make any difference, but I love my lists...

While that is true, it on the other hand makes passing on gear that much easier - if the item has the same ilvl and no extra gemslot its very safe to pass until absolutely no one else wants it.

Very well worded.

Rawr has a custom role that allows you to do exactly this. Use your WWS/WMO(preferred) and enter in relevant information and it will do calculations to show how you can improve yourself gear wise for the next time you do this fight.

We agree about the peak hps / throughput part.
Considering the longevity / sustained hps part, when you go oom, there is two ways to address the problem. Either you can improve your mana regen, and spam more heals, or you can improve your hpm, and use less frequent harder hitting heals. Of course, with the current itemization, increasing regen is likely to be the "best" way. But we still should not neglect the impact of crit / spell power for the hpm ratio, and therefore for the longevity. This is precisely what "sustained hps" measures : how much total healing can we give till oom in a fight. If you are lacking mana, you will see high weights on regen stats. However, you will still benefit from crit / spellpower for longevity.

One first point is the trouble to weight longevity and throughput stats is a global ranking. A 60/40 (often read) split has no meaning, because we have no obvious correspondance between mana regen and throughput. Is one MP5 worth 1 sp ? 10 sp ? 0.1 sp ? You need to precise this.

One way to consider this point is the comparison of haste and crit. We all know that haste is "globally" a better throughput stat than crit. We also all know that haste increases our mana expense, whereas crit does not. How to "modelize" it ? You can put a negative weight on longevity for haste, but that's a little bit crazy. More haste doesn't mean that you have to spam more, just that you have the ability. You also can't ignore the question : if you're lacking regen, between haste and crit choices, you should be choosing crit, because it's good for your hpm. That's why you need to give a positive weight on crit and spellpower for longevity ranking. How much ? This is precisely what sustained hps computes.

Knowing you, I thought you might bring this up, because it is a good point. Changing your heal per mana will affect your mana by decreasing the rate at which you spend out your mana. This would affect your longevity without changing your mana regen. Longevity is actually proportional to:



So I am a bit uncomfortable with using mana regen as a surrogate for longevity since then you are implicity assuming your mana pool and your mana burn rate (the rate at which you spend out mana) are fairly fixed. That is fairly true for mana but your mana burn rate is affected by a lot of things, which does make me uneasy. Also, mana regen is related to longevity via a hyperbolic transform, which is very non-linear and tough to wrap your brain around.

Still, sustained hps has even more nebulous meaning to me. So if I have X sustained hps, what does that tell me about my chance to run out of mana on a particular fight? I don't know. "More is better" is about all I know. At least with mana regen, I have some idea about how quickly I am spending mana, and so I can estimate how a certain amount of Mp5 might slow that and how that ultimately will affect how quickly I run oom.

Also, I don't think it is all that crazy to assign negative Mp5 value to haste. The practical problem you're trying to get at is how the denominator is affected in the above longevity formula. There are a few alternative ways to approach that problem, but if you assume a spam healing (i.e. near max throughput) situation, then spell power and crit do not a affect that denominator and haste does (in a negative way).

Although some may disagree, I think that is a fairly intuitive approach. The "feel" of haste is that you can heal faster, but also that you run through your mana faster. With spell power and crit, I do not have a sense that my mana lasts longer -- only that I can heal for more. Maybe others end up feeling "Wow, this spell power really makes my mana go farther", but that has not happened to me, even though I know I can heal more with the same mana.

Alternatively, you could reject the "spam healing" situation and postulate that you have a fixed amount to heal. This is essentially what sustained hps does. In that case, your mana burn rate is unchanged by haste but goes down from spell power and crit. So an alternative to sustained hps would be to assign positive Mp5 values to crit and spell power.

Overall, I would like a better way to quantify longevity than what we currently have... but I really do not think any such metric should be measured in heal per sec. That just doesn't make sense to me. It is just the wrong unit of measure. I can live with mana per time or time, but heal per time is... odd.

Perhaps a more useful measure would be to know how long you can sustain your maximum hps. Sure, you don't usually operate at your maximum hps, but that would give you some useful information nevertheless. For instance, if you can sustain your maximum hps for 5 minutes, you know you are not going to run oom unless a fight lasts longer than that. For longer fights, you may need to slow down to last. That is useful to know, but I do not know how to easily translate sustained hps into such useful information.

Ideally I want metrics that make no assumptions about what the healer might be asked to do, including assumptions about how much healing needs to be done or how long a fight lasts. In other words, I want to know general bounds on what the healer is capable of, regardless of the particular encounter, raid makeup, or whatever.

For instance, peak hps is one such limit. That does not depend on who else is in raid, nor does it depend on the encounter. It is just a useful, general property of the healer. Even though that healer may never have to operate at max hps, it is still useful to know his/her peak hps.

Along a completely different dimension, an analogous measure on how long a healer can heal would be nice, although it is not nearly so obvious how to measure this best. I do not expect healers to be spam healing most of the time, but it still may be useful to know how long healers can maintain their maximum hps.

In theory, I don't object to following a "fixed amount to heal" scenario. But in that case I would want to specify a fixed rate of healing (instead of a fixed time) and ask "How long can you maintain X HPS?" And that is really the critical point for me: knowing how long a healer can last under some standard conditions is useful, but knowing how much hps you can maintain over a standard time interval is not. I'm not sure I'm getting that across well.

I guess another way I could say it is that sustained hps really answers the wrong question for me. I'd prefer to know how long I can heal at a certain rate rather than knowing at what rate I can heal for a certain length of time.... Hope that makes more sense.

Incidentally, you could specify an infinite time interval for sustained hps. This would answer "How much can I heal from my regen alone?", and it sort of gets around the problem of specifying a time. However, it still does not really answer the questions I most want to answer.

I agree with your idea of general bounds.
The throughput bound is easy : it's the peak hps, and correspond to the maximal hps you can provide in heavy damage phases.
The regen bound is less obvious. Basically, it depends on two main factor :

1. The needed hps you want to sustain
2. The duration you need to maintain it

Knowing both parameters, you can have an answer "Yes we can" or "No, you'll die".
Knowing only one, you can compute the limit for the other parameter (ie determine regions with answer yes and regions with answer no).

I don't have any theoretical reason to favor one approach or the other. Both makes sense, and are based on the same mathematical model. You just change the parameter you set up in the model (either fight length or needed hps).

I prefer to set up the fight length, but that's a personal opinion. The first reason for this is that I have a better estimation of fight length, which I also find more stable between fights that for needed hps. The second one is that I don't need to "change" the needed hps when my gear updates. Fight length is about the same order for every boss in LK (longer bosses have usely more regen phases). Starting Naax in blues, you want to heal for 6 minutes, but light damage. In hard mode Ulduar, you want to heal as mental for still about 6 minutes. I just know that if my peak hps is way above my sustained hps, I could do with more regen. If my sustained hps is close to my peak hps, it means I don't need a lot more regen.
But this is more practical reasons that theoretical ones. Last one I already indicated : there is a natural way to merge sustained hps and peak hps metrics. You need to be more clear about hidden weights when merging longevity and peak hps : you're adding apples and oranges.

I think you are right that to desire very general bounds. As you said, peak HPS is one such bound, and in my opinion quite useful to determine the abilities of one healer(however, one might have to differentiate between single target and raid-wide healing, especially if one wants values which are comparable between classes). Obviously the other important question is sustainability, however it is in no way obvious how to measure this.

Your first idea was to measure how long a given healer can sustain a certain rate of healing, a value one could measure in minutes per 1k hps. However, the fundamental problem is that time and hps are not anti-proportional to each other, so one has to fix an arbitrary hps value and just assume it is a good estimate. Again this is completely arbitrary and depending on class, spec and encounter may be wildly off.

Personally I had this idea as well and I think it is better than most other metrics. It is independent of raid set up(under the assumption that buffs like replenishment or paladin blessings are present), which is already a big bonus. You said it does not answer the questions you would like to have answered, and I agree. However I think you want something which just is not independent of the raid and encounter, so you cannot find an answer to this. So lets look at what information such a sustained hps value would yield. Beside the obvious interpretation(you could heal at this rate indefinitely), it is a measure of optimal sustained healing, in the sense that in practice your value will be lower due to overhealing, suboptimal choices of spells etc. Since it is class and spec independent(again, perhaps split in single target and raid-wide), one can simply add the values of all healers to get the raid sustained healing ability. Compare it with the similar sustained incoming damage of a specific encounter and one has an estimate of the number of healers necessary.

An additional advantage is that it is directly comparable to peak HPS. Peak HPS is the upper bound of the healing, infinitely sustained HPS is the lower bound. Actual healing will be somewhere in between. Also, by not using the mana pool size for sustained HPS(except for regeneration), there is this buffer to use for peak HPS without reducing your sustained HPS.

As a sidenote, I like both values because I have emotions attached to their interpretation. The infinitely sustained HPS is the "completely bored threshold", while the peak HPS is the "frantic healing goal", both of which I have experienced and can interpret.

After writing this wall of text, I have realized that the biggest plus in my book (beside the above intuition) is the fact that you get valued which might be comparable between different classes.

I've been following up on the discussion with haste. One thing I've found people aren't mentioning to often is Prayer of Healing and haste.

Lately, the more we do hardmodes, the more I find myself helping raid heal with Prayer of Healing. And even though I've got 420 haste in normal gear, I do feel like I could use more haste to power my Prayer of Healings.