My other Tree druid and I ran rapid rejuv on the Lanathel fight merely because of the 2sec healing tick to counter her 2sec aura-dot. In theory it makes sense. We split up the raid, each of us covering two groups, with the last group covered by a chain healing shaman (group 1 is our tanks and two melee) and typical main tank heals.

I found it to be pretty difficult to maintain uptime on all targets throughout the fight, but at no point were we ever in danger of losing anyone due to the raidwide damage. Its hard to say that its necessary, but I can notice a significant amount of hps throughput difference when I did the fight without Rapid Rejuv.

With regards to the equation M = LN(B/A)*H, I would say that though this is a good attempt at calculating the effectiveness of heals (and would like to thank Rijndael and Paininabox for starting the groundwork on this), it overly favors direct heals by a VERY wide margin, to the point that I think it would make Druids appear to be sub-standard healers even on fights like Twins and BQL. I would love to see an effective weighing system developed for heals, and eventually even an addon that implements this system.

Perhaps the problem is that the starting equation, 1/x, has too steep a change in slopes. Maybe 1/[2sqrt(x)] would work better? The anti derivative over (A,B) is [sqrt(B) - sqrt(A)], and therefor M = H[sqrt(B) - sqrt(A)]. The alternative (M = LN(B/A)*H) exponentially favors heals cast as the target gets closer to 0. M = H[sqrt(B) - sqrt(A)] still favors heals that come when the target is close to 0, but not excessively so.

I was inclined to suggest a much more simple equation, M = C*H, where C represents the target's health deficit (as a percent), which would have a linear slope and would not factor in how high the target's health was after the heal. The only reason I consider the other suggested formula M = H[sqrt(B) - sqrt(A)] to be better is that it arbitrarily factors in a "minimum threshold" of hp by having a sharper curve when the target's at lower hp.

Last thought: a starting equation of 3/(4x^.25) would provide an M = H*[B^(3/4) - A^(3/4)], which is less steep than the other two options while still favoring heals coming in at lower health.

The thing is, you seem to have a particular outcome in mind, and you are tweaking the formula until the result you want (healer value approximately equal) comes out. But there's no reason to suppose healers are equal. Different healers are good at different things -- burst healers (shamans, holy priests) really are a LOT better than maintenance healers (resto druids) at saving people from death, whereas maintenance healers are better at keeping a ton of people out of trouble in the first place (as Carebare was saying).

This is why I wrote my post with a particular "objective" metric in mind (death probability). If it turns out burst healers come out better with respect to that metric, well we just have to take that at face value and conclude they are better at saving people from death in this sense. No one is going to stop inviting resto druids to raids if some death probability addon shows them as low on the meters -- because druids still have a useful job to do that no other healer can do as well as them.

Healing is more complicated than dps, so we need more metrics to evaluating healing in raids. Different healers will come out better or worse depending on the metric you use. With dps it's just a single number (although some encounters favor less tangible things like burst or target switching or mobility).

edit: Incidentally, death probability is completely self-calculating, you can start with a prior probability P(death|current hp, encounter), and simply update it as people die to obtain posterior probabilities (if multiple people run the addon, the death data can be shared as well to obtain a large sample size across servers/guilds).

Death probability also has a lot to do with the fight in question (and so would be different from fight to fight, guild to guild, etc.). All we can really hope to achieve is a "reasonable" estimation of death probability based on a very small sample of variables compared to the vast number of variables that actually affect it.

You are right that I was trying to tweak the values with an outcome in mind, but that outcome was not to make all healers equal. I was trying to find a "proper" way to value the preventative nature of hots in a system that would obviously favor direct heals. Ultimately I was trying to value heals based on their ability to maximize health pools rather than on their ability to reduce the probability of death - which perhaps was the wrong perspective to take.

Edit: The reason for this perspective is that I think such an addon/formula would be providing false calculations if it undervalues hots. In 10 man, for example, on Blood Queen Lana'thel, a druid can essentially solo heal the raid in a 2-healer group. I suspect a priest can too. But I highly doubt a shaman could. The shaman would be providing much larger heals than the druid, and yet the shaman's party would probably be wiping (if the shaman were the only raid healer). And yet I believe this formula would declare the shaman to be the more effective healer.

Edit2: Perhaps some other things worth taking into consideration in this formula would be:

1) Health restored relative to damage taken: (H/D)*H.
Where H = healing done, and D = health deficit. This would make small heals for heavy damage somewhat useless, but if the damage is small and the healing is small, the healing receives nearly full value.

2) Time between damage taken and healing done: 1/[sqrt(T)+1]*H.
Where T = time in seconds between the damage received and healing done. The longer it takes to receive the heal, the less that heal would be worth. I choose sqrt(T)+1 as a means to avoid dividing by values very close to 0 if the healing occurs nearly instantly after damage is taken and to avoid diminishing the value too much as time goes on. (This would probably be a somewhat complicated value to determine because of the effect that different sources of damage would have on the time intervals.)

I think the best way to view healing effectiveness would be to simply look at incoming damage. There's no need to derive a strict mathematical model and/or look at health deficit. We could simply assume the goal is to keep people at 100% as much as it is to keep them above 0%. Maybe we can't achieve 100% health 100% of the time, but stating it as the ultimate goal would probably make analysis easier. We don't easily have the information in the logs about health deficit, but we certainly have it for incoming damage.

You simply say "here is the damage profile" for a particular fight. Given the available tools, what is the best way to counter that damage profile? Obviously, this is what we actually do every time we step foot (branch, root?) into a raid. If you still want a strict mathematical model, then model the damage, not the healing. But really, there's no need because we already have all the data for the real thing, and don't need to rely on approximations.

If you wanted an addon to tell you effective healing, then it could (possibly for each fight) have an expected damage profile (and possibly also an expected healing profile). It would then tell you if the situation was or wasn't matching the profile, whether that be DPS standing in fire, or healers doing the wrong things. Then we could discuss the benefits of different profiles.

So many things related to comp and strategy come into play and being able to swiftmend anyone in the raid at pretty much any time is certainly very powerful.

I don't want to bog down the thread restating things I've said before much, but the short of it is that everything left that I quoted assumes that it makes no difference for the targets survivability whether they have a RR or a normal rejuv on them. In situations where that is the case, and there seems to be agreement that 25 normal BQ is one of them, you'd be crazy to run with RR. However, in a case where you'd have to cast some additional heal on someone with a regular rejuv but not on someone with RR most of those objections don't hold up.

Positioning is a huge consideration and the main source of the strength of RR. Looking over the fights in ICC sofar, in pretty much every fight where ranged don't clump up in melee range of the boss, they spread out trying to put 10 yards or so between them, making raid healing ranged classes much more difficult. Look at how you're using your spells in a 5x1 rotation too. Do you find you have a tendency to WG the melee because you know it'll hit 6 people there. Do those people really need rejuv on top of all the CoH and WG and Chain heal and Glyph of HL and JoL heals they're most likely already getting?

I feel myself fixing to go on for a while, so I'll cut myself off. Looking at the average HPS per target and evenness of the spread of healing casting rejuv x5 on ranged people followed by WG x1 on melee I think RR will prove to be very valuable in any situation where people need more hps than a single unhasted rejuv provides, which I expect to be the case in most hard modes.

P.S. A good way of clarifying my point using the terminology used recently in this thread: If I were designing a formula to normalize the value of raid healing I would include a term that significantly discounted the value of healing people who are standing close to other people relative to healing people who are 10+ yards from the closest person.

What are you guys planning to do for Valithria on 25? I'm not sure how our single target HPS compares to Holy Priests (Shamans & Paladins will be on Valithria), but I'm guessing we will be the ones that raid heal and stay out of portals (throwing spare GCDs on her ofc). Or if every healer can go in, spec and glyph for Nourish?

One other consideration and a question to be answered is whether she is affected by Tree of Life aura or not. If she is and you zone into the portal as the only resto druid, everyone outside suddenly loses 6% off of their heals, which is probably enough to mean that if you're only using 1 tree you'd probably never want to have them going into portals as a result.

Unless you have a Prot Paladin as well.

I missed the last round of testing her on the PTR, but when I went in on 10s we had three healers (two trees, and a shaman) even with 6-7 stacks of the buff from the portal I would notice very little change in my HoT ticks most of my Nourishs would crit around the 16-18k mark (as opposed to the normal 10k ish) The Resto shaman was hitting well into the 20-25k range. I would suspect a Paladin well up into the 30k

I have a feeling this will very much be a holy pally/priest fight with shamans after that. Us Trees and the Disc priests will be outside to hold things together there. I think it actually works out pretty well for us as there is a lot going on outside and movement is a very big concern. However if your raid is attentive a lot of the damage can be avoided.

Well, BQL and Twins aura damage is greater than an unhasted Rejuv. That's not really the point though. I've been saying for awhile, RR is an issue of focus vs. spread. Raidwide auras are pretty much the worst situation for increased focus. The obvious example is Saurfang, where all the damage is focused (tank, Mark, and Boiling Blood).