Nailer

No it won't. Right now you have the fixed cost of applying 3 stacks once, and then you just keep refreshing it every 9 or 10 seconds. If you apply 3 stacks every time it blooms, you are paying the cost of 3 LBs instead of one. Not to mention you spend 3 GCDs instead of one whereas Blizzard stated they do not want us locked up too much in rolling.

 

Rijndael

You didn't understand what I am proposing. I am not proposing to put up 3 stacks at once. I am proposing to put up 1 lifebloom per 9 seconds, for 3 cycles and then letting it bloom. This means for 9 seconds you have 1 stack, for the next 9 we have 2 stacks, and the next 9 we have 3 stacks. Then the cycle repeats. On average, then, we have 2 stacks on the tank instead of 3.

It is true that the main value of lifebloom is tank stability, and tank stability will go down 33% with this. BUT expected HPS will stay the same. Yes I realize the HPS variance (rather than HPS expectation) of 2 stacks + 1 bloom is not as good (e.g. larger) than the HPS variance of 3 stacks. (This is what people are saying when they say the 'model is not realistic' if it claims the expected HPS is staying the same. What they mean is that the variance becomes much higher -- but expectation does stay more or less the same, by definition of expectation). Rolling lifeblooms were too good, so something had to be nerfed. They decided to either make you pay a lot of mana for existing low variance, or pay existing mana cost for much higher variance.

Now, I think it is a very good argument to make to Ghostcrawler that since what is being nerfed is the variance of the HPS of the lifebloom stack, a question should be asked: what low variance is worth in terms of mana consumption. Right now the nerf claims it is worth 100% mana cost increase. I don't think that is reasonable. I think it would be more reasonable to do something like this: increase lifebloom cost by 10%-20% across the board, and increase the cost of keeping a stack continously (via existing mechanics of a bloom refund) by 20%-40%.

 

Paininabox

Slow-stacking lifebloom is not a good way to cope with this change. If you're that crunched for mana that you can't handle rolling it, then something other than the spell is wrong. Even if this slow-stacking model to bloom proves to be the best way to handle it, then Blizzard will probably unnerf it a bit because it has been stated that they still want us rolling lifebloom.

In times like this, I wish we had a way to measure HoT overheal and efficiency.

Level 80 Restoration Druid Spreadsheet here.
(v1.41)

 

• Playered

For PvE we could endure a cost increase of 75%~ if there was some added value on the heal over time aspect which was shifted over from the bloom or something. If the intent is to make rolling on multiple tanks too expensive then that is fine (no matter how much we dislike it) but honestly it should be bumped up to being better on that one tank but not high enough to be so amazing on the second tank rather than being left at the current values.

I would also try and ignore the bloom aspect for PvE too as you aren't meant to suddenly try and time it and no designer is trying to make you do it either so spending any effort on trying to work out best ways to do it or anything seems pointless.

We should try and work on values for their intended mana cost increase which is 75-80% at best and 100% at worst. What we need to find out is how changing the spell by shifting some bloom healing to hot ticks will make it more efficient (keep in mind this cannot be too much as the bloom needs to remain a solid aspect of the spell) for PvE or at what levels of output the ticks need to be doing at the intended cost to keep it solid for the intended purpose. This being that on the MT who takes a lot of sustained damage it is good and if there is a second tank taking similar levels it we can fully commit to rolling it twice but it is exhausting on our mana to do much more than that.

However situations like Patchwerk we should not honestly be rolling more than 1x LB because it isn't needed (RJ/Nourish on OTs should be better) and encounters like Sartharion +3 we should have the choice of 2x LB/RJ rolling on tanks and almost nothing else or 1x LB/RJ rolling and then helping raid healing compared to right now where it is 2x LB/RJ rolling + raid healing + other stuff.

 

Rijndael

Well, I mean we can still roll lifeblooms on one tank without problems, and we can still slowstack lifebloom on multiple tanks without a mana penalty, so I don't see what the problem is. You can no longer stabilize multiple tanks at once, but that's the point of the nerf! You can do everything else.

It's not difficult to get a first approximation of lifebloom overheal, btw. Look at a Patchwerk parse, and compare actual lifebloom healing to what it would have done had it healed the entire fight duration. Dividing one by the other gets you the overheal %. Using my last Patchwerk parse, I obtained that lifebloom overheal is about 60% (actually 62%). This number should go down the faster a boss is hitting.

Re: what Playered said about timing blooms -- even if you completely ignore timing them, they will still land a heal the same percentage of the time that the tank is below full hp (which coincidentally is the same percentage that governs overhealing of all spells that go on a tank, be they HOTs or direct spells). So for example, if I slow stacked lifeblooms on our last Patchwerk kill, I should expect the bloom to land a heal about 38-40% of the time, and overheal the rest of the time. Incidentally, as a sanity check, all direct healers on that fight had about 60-65% overheal, which supports my model that a single number governs all overheal of tank heals.

 

Mbuzi

By slow rolling you still incur a mana penalty after bloom. Instead of wasting 400 mana every 10 seconds reapplying a 3 stack you're wasting 400 every 30 seconds. So you've reduced the mana nerf to 66 m/5 at the expense of stabilization HPS as you slowly ramp up the stack.

Regardless of which way you slice it you're either losing straight m/5, some amount of HPS, a GCD to reapply the stack because we have to let it bloom now to achieve similar mana efficiency, or some combo of the three.

 

Rijndael

I don't understand your reasoning. The patch notes say a bloom refunds half the mana cost PER APPLICATION OF LIFEBLOOM.

So here is what happens:

You slowstack 1 lifebloom (you have paid 2x current cost). In 9 seconds, you slowstack 2 lifeblooms (you have paid 4x current cost). In 18 seconds, you slowstack 3 lifeblooms (you paid 6x current cost). In 27 seconds, you let 3 lifeblooms expire (you are now refunded half of what you paid, which is 3x current cost, so you have now paid 3x current cost for 3 lifebloom spells. So you are breaking even).

 

Mbuzi

What do you do after the stack blooms? You cast 3 LBs again instead of the 1 you would have previously used to get the stack rolling again.

Any time you are going to reapply the stack of LB after letting it bloom, you are suffering a mana hit. The question is what period do you choose to suffer that hit over.

I suppose in a perfect world your stack blooms just as the boss dies, you therefore have no need to restack LB. In this case there is no loss of mana efficiency, but you have lost HPS as you slowrolled.

 

Rijndael

No, I start the cycle at 1 lifebloom again -- that is the whole point. Please read the earlier posts carefully, I don't want to crap up this thread explaining this over and over again. Earlier posts explain why the expected HPS of slowstacking is (approximately) the same, why the mana cost of slowstacking is the same, and why the HPS _variance_ of slowstacking is higher (higher variance is bad). If you want more explanation, please PM me.

 

Mbuzi

No need to explain. I reread your older posts and didn't previously understand you were including the bloom effect in the HPS calculation. I understand what you are saying about variance.

In your post yesterday you say (which is not the model I am using):
(*) This assumes that when the tank is not full, the bloom wouldn't overheal. This isn't technically true, but I am ignoring that as a first order approximation of what's going on.

In my mental model I am assuming the bloom portion is much more likely to be wasted which is why I'm coming up with lower HPS.

 

Rijndael

I have this theory that the overhealing of all tank healing spells is governed by (and approximately equal to) the same percentage -- namely the percentage of time the tank is below full hp. This theory is difficult to test because, while the overheal of direct healing spells is recorded by WWS, the overheal of HOTs is not. I tried to calculate the overheal of lifebloom from first principles in an easy to model fight like Patchwerk, and the theory seems to hold in that case -- e.g. the estimated overheal of lifebloom is about equal to the observed overheal for spells like Greater Heal and Holy Light. If the theory holds in general that means the bloom portion of lifebloom is no more likely to be wasted than the rolling stack portion -- they will both have the same overheal % on average.

This theory would fail if:

(a) The tank regularly dips below maximum, but not low enough to prevent some overheal. For 3.1 lifebloom this would be a dip of less than 9-10k hp. This happens on some bosses, but not on Patchwerk since Patchwerk hits really hard.

(b) If direct healing can be made reactive for some reason (like the boss telegraphs a big spike and direct healers can respond to that, like Murmur in Shadow Labyrinth). This would make direct heals pull away from HOTs in having less overheal, but it wouldn't make hots differ from the bloom portion of lifebloom, since you can't control the bloom realistically.

(c) If the damage is "rhythmic" and predictable, and a faster direct heal always beats a slower heal to the punch. That would make the slower heal have much higher overheal. Again this doesn't affect the bloom portion of lifebloom or hots.

(d) Some other stuff I haven't thought of .

 

Raised

While the slow-stacking strategy might be the way to gain back your mana efficiency on paper, in a real game setting it's not very useful.

-Your healing is completely arbitrary. You're just as likely to be rolling 1x LB during the high-damage breath/strike and 3x when your tank is topped up as the other way around, making the stabilization factor of LB (its main draw) very random.

-Your fellow healers are also effected. Having to adjust to the 1k+ HPS swings from 1xLB to 3xLB+Bloom will be noticeable from a healing standpoint, and will result in your other healers having to do more work to adjust to your LB cycles.

All in all, this is a nerf that's justified and meant to be a nerf. Changing your healing style to try to compensate for it outweighs just eating the mana loss and making more efficiency choices regarding Nourish vs. RG and LBx1 vs. RJ for raid heals.

 

Mbuzi

Don't get me wrong. I agree with you. I was trying to understand the model being put forth. As it stands now I will suck up the mana hit as stability and predictability are worth more to my guild than mana usage. Besides, there aren't many mana taxing fights in the game currently anyhow. If Blizzard chooses to make me think about mana regen I'll cope with that when the change hits.

 

Paininabox

That's not a truly good way to measure overheal, especially lifebloom. The stacking mechanic makes predicting how much healing you would have done impossible, because your stack falls off and then there's the time when you're actually stacking. For a fight like patchwerk, you are correct in that you can probably assume that a 3 stack will be up the entire fight and probably make a good estimation, but for 90% of fights your stack falls off for various reasons or you might swiftmend on the tank while you're restacking, etc. The only good measure of lifebloom I can think of that can be done with the combat log restrictions on hots is to count how many times you cast lifebloom, multiply that by the mana cost, and use that number to divide the total healing done by lifebloom so that you'd get a measure of how efficient it is to use lifebloom. Then again, that number would fluctuate wildly on a per-encounter basis or even every attempt.

Level 80 Restoration Druid Spreadsheet here.
(v1.41)

 

Fallenangel

edit : wrong calcs.

 

Rijndael

I agree, Patchwerk is a particularly easy fight to model. Obviously if you stop healing the tank, the numbers change (but overheal does not, notably, since it's by definition one minus the ratio of landed heals over total heals). This is true for direct spells and for HOTs. I guess what you are getting at is that if tanks take a lot of disparate damage, and receive a lot of disparate heals, it becomes very difficult to predict what happens to his health. This is why I was treating this as a binary random variable with two states: full hp, and less than full hp. The distribution of this random variable will, you are right, change from attempt to attempt, but to a first approximation it will determine the overheal %. In practice, overheal % for spells we can we measure tends to stay relatively constant, so it appears the "randomness" you are concerned with isn't a big deal in practice.

I agree that effective heal per mana is a good measure of efficiency also, although it will be a little harder to calculate it, now that lifeblooms have this weird mana refund mechanic. It would be hard to figure out, for instance, how useful the extra mana spent on rolling actually was (since it's confounded by other uses of lifebloom without rolling).

 

Paininabox

I heartily disagree with this statement. It has been my experience that some spells perform better in specific situations, and the presence of these situations is not a constant. Rejuv, for example, is phenomenal in situations where a large number of people are taking heavy damage that won't necessarily kill them quickly. In situations where few people take light to moderate damage, Rejuv becomes less great. You may say that, following this, that Rejuv would be constantly more effective in some boss fights and constantly less effective in others. I also disagree with this because random happens, such as a group of players accidentally shocking each other in Thaddius, but that doesn't mean they make that mistake all the time. Even guilds that are clearing farm content occasionally wipe on bosses they have cleared many times, and raid makeup will change from one week to the next. The overheal % also varies from player to player, as some people will use their hots to more effect than a less experienced player, or a more experienced player has a bad night and a less experienced has a really good one. In tank healing, you are likely correct that the hots you cast on a tank will likely have a constant % of overheal as all variance of a fight will converge on constant behaviors, but no one will heal only tanks and even tanks will make mistakes, such as messing up a rotation, so that even in the best situation the hots won't essentially overheal a constant % of the time. All of these wildly varying circumstances suggests, to me, that it would be false to assume much of anything as a constant from boss to boss or even so much as attempt to attempt.

I do not think that I couldn't be wrong, but I do think I would need some more true mathematical analysis and WWS reports to confirm what you seem to base largely on conjecture and speculation.

Level 80 Restoration Druid Spreadsheet here.
(v1.41)

 

Rijndael

I am going to test slowstacking on our next Naxx 25 and may have something useful to say then in terms of lifebloom overheal and effective healing over mana spent. In the mean time if you have a good refinement to help model HOT overhealing, by all means let's put our heads together to figure it out -- it's a useful problem to solve and not just for druids.

[edit: removed the rest of the post as it wasn't useful. ]

 

Paininabox

I look forward to seeing what you find out about the slowstacking method and how lifebloom measures out with total fight efficiency paticularly because it seems the only way to gauge how good the spell is. As to a HoT overhealing model, that would be extremely difficult to do, and my advocating the randomness of a fight makes it even more difficult. I do, however, have a few ideas that could help at least approximate HoT overhealing. I'll have to play around with it to see if my thoughts even hold any water, however. I wish Blizzard would simply have the HoT's tick on overheal so we could parse them from the combat log without having to take a more indirect approach.

Level 80 Restoration Druid Spreadsheet here.
(v1.41)

 

• Playered

For what it is worth Replenishment never gives 100% of the effect as right now we should be having around 210~ in best in slot gear fully buffed but we get around 170-180 from what parses show.

Your concept of how to break down mana usage is something I initially didn't think of either (and its a good way to go about it) but you forgot to include the mana pool which can be equated down into mp5 depending on fight duration and that innervate fluctuates between fight length. It also ignores OOC procs which in a rotation like that shouldn't be hard to model.

 

Nitz

The only problem I see with the Lifebloom nerf is how triple stacks + RJ + RG will swing compared to 1 Lifebloom + RJ + RG then using Nourish for tank healing. I definitely see me still triple stacking LB on top of other HoTs unless mana really becomes a problem (and it might become, from what I've seen of Ulduar, the fights are long and we all know Innervate is heavily nerfed).

 

Tatsumakisen144

Sorry if I am mistaken but aren't many of these arguments based around LB's application on tanks only? You have to remember this:
I for one like to cast one stack of LB on the non-tanks once they take minor damage in raids so as to provide some sort of cushion should they take any further spike damage as well as to help out the other healers. Having examined this play habit further ive come to the conclusion that apart from the re-rolling of LB on tanks, the subsequent blooms and mana returns on (lets say 5) of the non-tanks will give me sufficient mana return per application on tank.
So while I am waiting the 9 seconds or so in order to re-roll the 3 stack on the OT, I get 5 (at 50% its original cost) mana returns from the blooms that come off the non-tanks. That's 2.5 LBs back to me in mana costs that will be used on the OT or while I wait for mana regen through the increase in Intensity talent.
Ofc the only way to find out whether the mana gained from each bloomed non-tank target is mana efficient so as I may re-roll a stack of 3 on an OT remains to be seen.

 

Odoacer

Most of the discussion is based on tank rolling due to the fact that lifebloom is generally not a raid healing spell. It has been proven in previous responses and debates (as well as general logic) that regrowth/rejuv outweigh lifebloom in terms of raid healing. In most cases, you're lucky to only get a few ticks off and then the bloom is just overheal, because your raid members should top off the raid member(s) before it even comes close to finishing its duration.

 

KrinKer

I have been wondering a lot about both lifebloom changes and Nourish suposed mana eficiancy.

Blues have said that the refund would be based of of base mana (if I remember correctly). So assuming the spell costs 800 mana outside of tree form (which would make it 560 mana in tree form because of the 20% mana reduction on hots, correct me if I'm wrong) we would get 400 mana back when it blooms. So it would cost 160 mana to cast lifebloom if it blooms, to me this looks like a pretty nice buff for both raid healing and pvp. If only we had something that would allow us to "bloom" it on demand it would become so much better. Imagine If blizz added something like: "you can't have more than three target lifebloomed at the same time" it would nerf the rolling lifebloom and allow us some control over the bloom. I know I'm probably dreaming but still.

As fasr as Nourish goes, I'm wondering if it's as mana efficiant as people have it to be. To me the hots that we have to keep up on the target needs to be added to nourish's mana cost to make it closer to reality as what it really costs to have the number nourish pushes out. In order for Nourish to be better than Regrowth atm in need 3 hots on the target ( Lifebloom, rejuv and regrowth) SO in order to cast 1 nourish (and for it to be more efficiant than grlyphed regrowth) you need to have those spell casted.

So for 1 nourish to be cast and to be more efficiant than Regrowth (meaning 3 hots +++ on a target) it costs 509 (nourish) + 366 (lifebloom) + 466 (rejuv) + 719 (regrowht). So to cast nourish (the first one) costs 2060 mana. So assuming lifebloom is the last hot casted you would have the time to cast 5 Nourish would mean another 4605 mana. So in a perfect world when u need to spam Nourish a druid would have to spend 6650 mana to make nourish most efficiant and needing to refresh hots again, meaning nourish has an actuall cost of around 1110 mana.

I know it is probably complicated to understand (not sure I understand my own self right there) but to me we have to take account of the mana cost of the hots in the mana cost of Nourish. Even if the hots are already on the target, those hots have a mana cost and without those Nourish is completly bad.

Someone please correct me if I'm wrong or tell me If I'm right but to me it sound right.

Thanks and sorry if I'm completly wrong.

 

david0925

To make the most out of Nourish you would need to cast those spells, but you're also assuming that the HoTs themselves don't heal, which isn't the case. Regardless of using Regrowth or Nourish, you should be keeping Regrowth, Rej, and Lifebloom on the tanks anyway to provide a stream of healing. If you are referring to raid healing (which I don't think druids are good for aside from Wild Growth and maybe Regrowth+SM), then Regrowth will definitely be better than Nourish. The buff to Nourish made it a very worthwhile spell to cast on targets you already have hots on, which is just another tool for Druids to be efficient Tank healers. If you need to cover raid healing outside of wild growth, then Rej and Reg will both be better.

Maniq is my hero