Elitist Jerks WotLK Healing Compendium v3.2: Same Old Thing

04/24/09, 10:54 AM   #61
Elimbras
Don Flamenco

Dwarf Priest

Eitrigg (EU)
 Originally Posted by tedv [...]Of course, I'd also like to note that there is a very strong correlation between keeping everyone alive and outputting a lot of healing. If the boss deals 12 million damage throughout the course of the fight and your healing team only heals 11 million damage, then people have absolutely died. [...] Total healing output won't tell the full picture, but it does tell the vast majority. We should not ignore this tool simply because it's only, say, 80% correlated with our goal. With that in mind, I'd like to reiterate that for most fights, the primary heuristic we should follow is maximizing the total amount of healing we can output. [...] For example, suppose you have three possible gear setups. You could get 7 mana and a healing efficiency of 3 healing per mana, or 3 mana with an efficiency of 7 healing per mana, or 5 mana with an efficiency of 5 healing per mana. The 5/5 split lets you heal 25 damage while the others only heal 21, so it's the best choice, barring strange fight mechanics where a 5 damage heal is too slow but a 3 damage heal isn't. I want to know if the typical gearing setup leaves us in the 3/7 state or the 7/3 state. It's highly unlikely that your gear is perfectly itemized such that getting either throughput or regen stats will give an equal increase in total healing, however.

As long as the raid is not decimated by the boss, it does have enough total hps. I really think that usely we are not lacking pure total throughput. If your hps is less than the incoming dps, your total raid life is a random walk with a negative drift, and then, you're likely to hit 0 really quickly ;-) In fact, to be able to survive with a reasonnable probability, and since our strating lifes are not such high, we need to have significantly more hps than the boss does damage.

Now, it's true, absolutely true, than increasing hps increases survivability. It does. It increases the positive drift of the random walk, and that decreases the probability to hit 0. However, I would not say that there is such an impact as you state. Imo, timing spells would have, in many cases, a bigger impact. We are speaking of, let's say, at most 5% more throughput, where we have more than 25% overheal, and we may not be casting full time ;-). I don't think that moving from regen flask to spellpower flask, or that changing gems from haste / crit to spellpower, will have more than 5% impact on total throughput.

I don't say that throughput is not good. We need throughput, and we need regen to be able to heal. I just don't think that optimizing the last percent will have any impact we can predict. Being able to chain cast in any situation, even if one / two healers dies, or if you shadowfiend dies, may be more productive than having more total healing in normal case. Might be...

Your example is typically the examples I don't think realistic, because you are too close of the edge of total healing.
And more importantly, you need to know how your 26 total damage (because it needs to be between 25 and 27) are partionned. If the scenarios 3*7 or 7*3 leads to any overheal, they are also no solution. Or if they lack the burst hps to heal spike damage ;-) You can even construct damage pattern where only the 5*5 solution can perform successfully (for example, a "tantrum" like scenario, with 4 times where you need to put at least 4 immediate heal : the 3*7 won't have the mana the fourth time, and the 7*3 won't have the burst hps (or even grouping two heals, it will fail the fourth time with only mana for one more heal). I don't think my scenario is realistic, but it's as realistic as yours, and it shows why going from 25 to 27 total healing is not always increasing survivibility.

04/24/09, 11:15 AM   #62
tedv
Observation: I am awesome

Goblin Priest

Mal'Ganis
 Originally Posted by Elimbras Now, it's true, absolutely true, than increasing hps increases survivability. It does. It increases the positive drift of the random walk, and that decreases the probability to hit 0. However, I would not say that there is such an impact as you state. Imo, timing spells would have, in many cases, a bigger impact. We are speaking of, let's say, at most 5% more throughput, where we have more than 25% overheal, and we may not be casting full time ;-). I don't think that moving from regen flask to spellpower flask, or that changing gems from haste / crit to spellpower, will have more than 5% impact on total throughput.
But we should do more than just share our opinions. Lets get real math on the total healing output we can do and then decide.

By the way, I don't buy the argument "more +healing just produces overhealing so regen is better". The argument relies on constructed examples and there are other examples where the +healing setup happens to win. Here's a typical argument.

 Originally Posted by Strawman Suppose you have a 5k heal and enough mana to cast 12 of them, but if you geared for spell power instead of regen, you could get a 6k heal and enough mana to cast 10. Both setups can do 60k healing. If you need to heal your target for 5k, then the regen setup can heal the damage from 12 hits (60k damage) while the spell power setup has overheal and can only heal 10 hits (50k actual damage). Therefore regen is better.
That is a bad argument. Because you could just as well say that the tank is taking 6k hits. Now the spell power setup can heal 10 hits (60k actual damage), but the regen setup now needs to cast TWO heals to heal up every incoming hit, resulting in 4k of overhealing for every 10k healed. So the regen setup only heals 6 hits (36k actual damage).

The point is that you can muck with the example numbers however you want to get a specific situation where a particular spell power / regen quantization is optimal or not. But real healing situations aren't so cut and dried. These examples aren't meaningful, and it's the primary reason we should focus on total healing output throughout. If one setup can do 10% more healing than another, then it's almost certainly an improvement, regardless of whether the heals happen to be a bit faster or slower.

 Originally Posted by Elimbras Your example is typically the examples I don't think realistic, because you are too close of the edge of total healing. And more importantly, you need to know how your 26 total damage (because it needs to be between 25 and 27) are partionned. If the scenarios 3*7 or 7*3 leads to any overheal, they are also no solution. Or if they lack the burst hps to heal spike damage ;-) You can even construct damage pattern where only the 5*5 solution can perform successfully (for example, a "tantrum" like scenario, with 4 times where you need to put at least 4 immediate heal : the 3*7 won't have the mana the fourth time, and the 7*3 won't have the burst hps (or even grouping two heals, it will fail the fourth time with only mana for one more heal). I don't think my scenario is realistic, but it's as realistic as yours, and it shows why going from 25 to 27 total healing is not always increasing survivibility.
Again, you can muck with examples however you want. The purpose of this example was to show that healing 25 total damage is almost always better than 21. It's not useful to try extrapolating anything additional from it though.

 04/24/09, 11:48 AM #63 Ehitos Glass Joe   Ehitos Night Elf Priest   Vek'nilash (EU) Hello guys, here are some of my thoughts... Almost every holy priest has Inner Focus, but why? Imo Inner Focus is just a waste of 1 talent point because there is only few fights when you could actually run oom. Ok I get that people take it so they can have +25% crit to a spell every 3 minute but still, it is really worth it? I don't think so . Before 3.1 when you could combo IF and IHC it was nice but now it doesn't have much use at all. However, the reason I don't run OOM might be because I don't use PoH at all, because we have many "one-second" chain-healing shamans in raid and by the time my PoH lands on a party, half of its members are already fully healed so don't see the point. Ok I admit, I don't have the best computer, running around with 15-20 FPS on boss fights which might be the reason why my PoH doesn't land in time even with the Serendipty stacked up.
04/24/09, 12:58 PM   #64
Venelar
Glass Joe

Barthilas
 Originally Posted by Ehitos Hello guys, here are some of my thoughts... Almost every holy priest has Inner Focus, but why? Imo Inner Focus is just a waste of 1 talent point because there is only few fights when you could actually run oom. Ok I get that people take it so they can have +25% crit to a spell every 3 minute but still, it is really worth it? I don't think so . Before 3.1 when you could combo IF and IHC it was nice but now it doesn't have much use at all.
I've started using IF with Divine Hymn for those "oh shit" situations and it is awesome, give it a shot. To me it's worth the point if i'm doing it once a fight and saving a ton of people with it.

 04/24/09, 1:00 PM #65 Zhaera Glass Joe   Zhaera Human Priest   Kul Tiras (EU) I use Inner Focus twice more or less on a fight. Once on Divine Hymn which is 2k mana, and once on PoH which is 1.5k mana thats saving 2.5k mana each fight. In a 5 min fight that is ~42mp5, not bad for 1 point right? Additionaly while casting Divine Iwinn, you have free OOC regen and you can probably get some with that PoH. So basically its a bit more than 42 mp5. As for PoH it depends. It is usually 15% of my healing. Is it a lot? No, does it save situations? yes!
04/24/09, 3:08 PM   #66
TheMutt
Glass Joe

Draenei Warrior

Alleria
 Originally Posted by Zhaera I use Inner Focus twice more or less on a fight. Once on Divine Hymn which is 2k mana, and once on PoH which is 1.5k mana thats saving 2.5k mana each fight. In a 5 min fight that is ~42mp5, not bad for 1 point right? Additionaly while casting Divine Iwinn, you have free OOC regen and you can probably get some with that PoH. So basically its a bit more than 42 mp5. As for PoH it depends. It is usually 15% of my healing. Is it a lot? No, does it save situations? yes!
I think this has take a nice turn from the last healing thread. Couple things I would like to add to the whole debate of regen vs output is that depending on the fight you can run same same total output with higher and lower downtime.

Take for example Kologarn to xt for example. Total boss damage would be about the same however the spells used vs mana used in a given period changes.

I find its important to note that on fights where you'll blow a couple PoHs quickly then have downtime / lower THS time. Compair this to Xt where you your throw a more contant mix of Flashes and gh's on the raid from light bombs.

2 different types of fights and I heal them in different ways as holy and personally I'll use SP flask on Kol because I want my fewer big heals to hit for more, were as xt is more of a keep mana up and pound is my only big mp5 time.

I tend to agree with the math that states higher output is more affective then int/mp5 mathematically, however from experience I find that unless your in a perfect word spike dmg and using larger heals/overhealing is more common and mana usage is higher them I think your taking into account.

Healing comp also takes a key roll in this bebate imo as different healer compositions require holy preists to change the way they heal. Just as before with a large amount of shamies you tend to do less PoHing, however on the other hand I personally raid with 3 holy paladins and end up having to cover more of the raid splash. In the later case Mp5 and int take a higher priority for me as I need to throw more large mana spells and having a larger pool to work with is more to my advantage then the extra HPS Spell power would add for me.

 04/24/09, 3:27 PM #67 Senres Von Kaiser   Senres Undead Priest   Dragonmaw I'm curious why [Darkmoon Card: Greatness] isn't listed in the trinkets section? After having [Soul of the Dead] only drop once for my guild, ever, and losing the roll on it I finally turned in my completed nobles deck for Ulduar. I replaced [Spirit-World Glass] and I've been very pleased at the impact it has had on my regen. I've also been surprised that the +300 int proc has been so useful -- I'm not at all disappointed to have it instead of the +300 spirit proc. The trinket will almost always proc near the beginning of a fight after a few casts. As a Blood Elf I can immediately pop Arcane Torrent to get back more mana, even going above my normal max mana, and use it up leaving me still at full mana with a good deal of healing done by the time the proc ends. It's been pretty useful on a couple of fights. I can also usually line up Shadowfiend + Hymn of Hope with a proc to pretty much refill my entire mana pool.
 04/24/09, 3:42 PM #68 binaryz Glass Joe     Lime Blood Elf Priest   Thrall On the topic of trinkets, I think [Sif's Remembrance] should also be added to the list. It's a nice upgrade to [Je'Tze's Bell] & easily replaces [Forethought Talisman].
04/24/09, 8:24 PM   #70
spathos
Glass Joe

Night Elf Priest

Bloodhoof
regen versus spellpower

 Originally Posted by tedv By the way, I don't buy the argument "more +healing just produces overhealing so regen is better". The argument relies on constructed examples and there are other examples where the +healing setup happens to win. Here's a typical argument.
I think your statement that more spellpower always equals more +throughput is what is flawed, though you are right that constructed examples are not the way to address this. To break it down, let's first look at the cases where having more spellpower is better than having less spellpower, and therefore see what extra spellpower exactly gains you.

The first case, and the most situational, is where having the extra healing power would prevent an instagib either on the tank or on the DPS. While we can all produce scenarios that this would solve for, the reality is that the majority of the time this happens is because someone made a mistake or you have a very unlucky RNG. Neither of these cases is something to gem for, any more than you should ask the DPS to gem for extra stamina to fix the problem of them standing in the fire for too long. That being said, if you consistently find your heals are too low, boosting your spellpower or boosting the tanks gear is obviously the right call.

The second case (and the least situational) is when the difference between the amounts of spellpower would allow you to cast one less heal or a cheaper heal to top the target off (e.g. when your greater heal hits enough that you can use a flash heal or a hot instead). While it seems like having a large number of extra spellpower would allow you to do that, the numbers you have given so far (for example ~243 extra hit points per target with a prayer of healing), amount to something between 1-1.5% of a target's health. This is hardly ever going to allow you to substitute a cheaper spell. The only thing that might be considered an exception is if the stacker would bring the heal target to 100% (i.e. near 0 overheal) while a non-stacker would have brought them to 99%. This is hardly something that most healers would be worried about as a smart healer would simply wait until the target takes more damage and use a heal then, or would alllow another healer to top them off with a very cheap heal (like a rejuv or flash of light).

Of course, back in BC, this would be a different argument entirely as you could have downranked if you had enough +healing and saved some mana. But the end result is that gemming for spellpower once you have enough to top your target off will only result in increased overheal (actual waste) while gemming for regen is only a potential waste as having more mana allows you to fix mistakes made by the raid and increases the margin for error.

Given the fact that having enough spellpower to increase your heal strength by 1-2% of a target's health isn't going to make much difference in the number of heals you have to cast, then haste really is the better way to increase throughput effectively as it allows you to cast those heals faster. That being said, I would agree with you that gemming for haste once you have a decent amount will only result in marginal gains in most cases.

Since gemming for spellpower does not increase the size of your heals to the point where you can cast fewer heals, it is my opinion that the only real purpose for having more spellpower on gear upgrades, therefore, is not to increase throughput, but to match the gains in hit points that accompany gear upgrades for the rest of the raid and increased damage to the raid. To put it simply, you can not gem for another tier's worth of spellpower without gimping other stats, but you naturally get the spellpower you need to heal new content as you get gear upgrades and stacking spellpower or any other single stat, is superfluous.

 04/24/09, 9:11 PM #71 Zhaera Glass Joe   Zhaera Human Priest   Kul Tiras (EU) Surge of Light - Necessary? I want to raise the issue of SoL a bit. Personally I have found that I have room for just 1 point. Is that point really worth taking over lets say lightwell, or probably desperate prayer? We're talking about an unreliable proc, because personally with 26% crit raidbuffed and 1/2 SoL I cant really say that it procs each CoH, versus reliable abilities. I could understand its use pre 3.1, where we played with OOFSR with IHC, Inner Focus and SoL procs. Now we cant really do that, so isnt it kind of undervalued at the moment? So much trouble for a free flash heal now and then? Last edited by Zhaera : 04/24/09 at 9:16 PM.
04/24/09, 9:39 PM   #72
Promethia
Piston Honda

Blood Elf Priest

Kilrogg
 Originally Posted by MavSteele I like the angle that you're approaching this from, but I think some of your examples actually argue the opposite of what you intend because the line that I've quoted seems flawed. You're arguing that because healers will have to cast more they will see a non-linear decrease in Mp5 thus regen is more important. This seems to miss the fundamental point of dropping healers for DPS: the fight should take significantly less time with more DPS which is why Mp5 becomes less important. ... since you've already dismissed "bring more DPS" as a bad strategy you use your max RHPS theory to prove that extra throughput is bad. I contend that "bring more DPS" is still a sound theory and in that context max RHPS is exactly why I gear for throughput. I think a lot of people run with too many healers; this thread is littered with comments of "why gem for bigger heals when someone else is just going to come through and heal that up". When I'm assigned to heal a group (or two or three) or an area, I'm the *only* person healing that group. With the prevalence of burst or RSTS effects, I need people to be at full in the fewest number of GCDs possible. Haste, crit and SP help me with this; regen does not.
You are mistaken if you think I dismiss "bring more DPS" as a bad strategy. I'm not sure where I said that, but if I implied it in any way, I didn't mean to. Of course bringing more DPS is great if you can, and you can until healing "breaks". All I'm saying is that for difficult and long fights, mana and mana regen is going to limit your ability to reduce your number of healers more than your healer throughput.

I am very aware that changes in healer healer longevity are very non-linear. I specifically mentioned "short" fights as an exception because those will be dominated by mana pool and mp5 becomes much less important. You offer a case as a counterexample where you are downing a boss in < 6 min even on your first kill. That isn't "long", so we're perhaps thinking about different situations.

Also, you are implying that you went from downing the boss in 5m27s to 2m30s because you added 3 DPS? I don't believe that at all. Do you really believe 3 more DPS completely explains that difference? Times always get much lower after your first kill, even if you do nothing to your number of DPS versus healers. How do you control/account for that? I don't think it is fair to credit the strategy of employing more DPS.

Estimating the time benefit of adding DPS isn't too hard. For example, in going from 17 to 18 DPS, you're decreasing the time to down the boss by no more than ~5.56%. This can be shown if you assume the time to down the boss with 17 DPS is T, then the time with 18 DPS is roughly 17/18 T or 0.944 T. Of course, that assumes tanks and healers do 0 dps, which is certainly untrue, but that just makes the estimate a ceiling estimate -- i.e. the benefit of adding a DPS should be less than that. Since you often have a lot of DPS players already, adding more should not increase your aggregate DPS nor decrease the time to kill the boss all that much.

Admittedly, in weighing that versus healer longevity, it is harder to estimate how going from 5 to 4 healers (for example) drops your healing longevity. There is ~20% less mana pool and mp5 while casting to deal with incoming damage, and each healer will be healing more and that will decrease oofsr regen by some vague and unspecified amount. In a 2.5 minute fight, that might not matter much, but in a 10 minute fight, it probably will.

But regardless, why would you suppose throughput matters more? If you look at the raw healing output capacity of a well-geared healer, it can exceed the total incoming DTPS for the entire raid. I know I've seen pallies with over 17K raw HPS. Of course, I'm not saying that one healer will work simply because he/she has enough HPS. All kinds of technical issues may prevent that. But that is kind of my point: throughput isn't generally the problem -- it is other things.

What I'd like to see is someone make a solid case for how throughput is really what is stopping raids from being able to bring more DPS and fewer healers. Exclude mana as an issue. Exclude range and LOS issues. Exclude everything except throughput. I have seen plenty of claims (mostly implicit, btw) that throughput matters, but no explicit and convincing evidence that it is "the" limiting factor,

04/24/09, 10:25 PM   #73
Elimbras
Don Flamenco

Dwarf Priest

Eitrigg (EU)
 Originally Posted by Promethia Also, you are implying that you went from downing the boss in 5m27s to 2m30s because you added 3 DPS? I don't believe that at all. Do you really believe 3 more DPS completely explains that difference? Times always get much lower after your first kill, even if you do nothing to your number of DPS versus healers. How do you control/account for that? I don't think it is fair to credit the strategy of employing more DPS. Estimating the time benefit of adding DPS isn't too hard. For example, in going from 17 to 18 DPS, you're decreasing the time to down the boss by no more than ~5.56%. This can be shown if you assume the time to down the boss with 17 DPS is T, then the time with 18 DPS is roughly 17/18 T or 0.944 T. Of course, that assumes tanks and healers do 0 dps, which is certainly untrue, but that just makes the estimate a ceiling estimate -- i.e. the benefit of adding a DPS should be less than that. Since you often have a lot of DPS players already, adding more should not increase your aggregate DPS nor decrease the time to kill the boss all that much.
I agree with you that the most part of time gain is because of more control over the fight, and not because of 3 more dps.
But being scientific minded, I can't let you with your false maximum bound on dps gain.

What you can state for sure is that the dps increased at most of 100/17 %. However, this doesn't mean that time was divided by a factor lower than 18/17. The obvious counter-example is when part of your dps must take care of adds / run / etc. In short, having 17 dps doesn't mean that they all dps, and if only a small fraction of your dps is doing effective dps, and your added dps can be effective, then your gain might be much higher.
Typical example would be any fight were some dps must do some kitting, or must kill adds that spawns constantly.

I must say that going from 5min30 to 2min30 by only adding 3 dps would still mean that you have only 3 "effective" dps, and that's really really low !!! So your conclusions still holds.

04/25/09, 4:04 AM   #74
Promethia
Piston Honda

Blood Elf Priest

Kilrogg
 Originally Posted by Elimbras But being scientific minded, I can't let you with your false maximum bound on dps gain. What you can state for sure is that the dps increased at most of 100/17 %. However, this doesn't mean that time was divided by a factor lower than 18/17. The obvious counter-example is when part of your dps must take care of adds / run / etc. In short, having 17 dps doesn't mean that they all dps, and if only a small fraction of your dps is doing effective dps, and your added dps can be effective, then your gain might be much higher. Typical example would be any fight were some dps must do some kitting, or must kill adds that spawns constantly.
Alright, fair enough: there could be DPS players doing some important things besides DPS, and you do need to consider only those actually doing DPS. Better yet, if you want to be more precise, you can estimate how much faster you can do a fixed amount of damage with:

$\frac{RaidDPS_{before}}{RaidDPS_{after}}$

So if you are able to increase your raid's aggregate DPS from 100K to 110K, you can do a fixed amount of damage in about 100/110 or ~91% of the time. Is that better?

 04/25/09, 7:25 AM #75 Liriel Piston Honda   Liriel Troll Priest   Alexstrasza (EU) I know, most of you will not agree with it, but sometimes if you want to increase RaidDPS it would be better to take another healer. In theorie you hurt your Raid-DPS with another healer. But in most cases where you realy need that little bit more Raid-DPS you encounter casualities. People WILL die. Due to lag, due to not running out of the fire fast enough, due to not beeing at full health at the wrong moment, due to bad luck. That would not happen in a perfect world but it does happen in reality if you are not heavily overgeared. (Or at least will happen to the majority of raids.) Even if you have enough healers adding another one may be bad on paper but could end those casualitis. So you would end up with more living DPS who can concentrate much more on DPS and have to worry about their own life less. So while you don't waite (or wipe) until you get the perfect run with less healer, in reality you may reach your goal with less trys. Sure, that may not be the option for world firsts. But it may be an option in some other cases.

 Elitist Jerks WotLK Healing Compendium v3.2: Same Old Thing