Note that I never said you're making a big change to what you're casting. Having some more +healing will have a very small change on the line where you turn from FoL to HL, HL9 to HL11 etc, but that change is there and cannot be simply ignored. Even if your 100 extra +healing caused you to cast 1 extra FoL that's already a considerable efficiency increase (in comparison to how significant other item effects are).

Remember that gear optimization is always about very small changes and upgrades, gearing better will not turn you from a terrible healer to a great one or from one that always lets the tank die to one that is able to keep him alive. It's more about trying to gain that extra 1% here and extra 0.5% there. Using the "it's too small of a difference" arguemnt can be directed at just about any comment made about gear choices (not counting totally obvious comparisons of course).

Also the last comment about tank gear is completely irrelevent. You're trying to optimize your gear, not the tank's gear, so you use whatever kind of damage your tank is taking to determine what kinds of spells you're using and use that to calculate your efficiency and how it would increase with any given stat.

None of your examples are chain casting rank11 HL. If you casted even 1 FoL during the fight then it means FoL is useable and the more HPS you'll have the more useable it'll be - slightly but it adds up just like any other itemization difference.

Even bloodboil the majority of the time you can actually get a positive mana regen since it's not too healing intensive between fel rages, and the fel rages are actually far less than 1/2 the time. So again having more HPS would allow you to increase your efficiency, although really on bloodboil it's all about the fel rages - you shouldnt' have mana problems and should probably care mostly about your HPS during those fel rages (and worst case regen mana between them when not much healing is needed).

Sorry, but you are wrong. If you are casting one FoL on raid or even tank (risking tank getting a big spike btw) it doesn't mean you (a) you should have casted that FoL instead of continue cast HLs on tank and (b) it doesn't imply you can cast more FoLs to up efficiency.

In Brutalus fight you only cast HL9/FoL and not HL11 to not go oom completely at some stage of the fight. You would always aim for higher HPS staying exactly at that lvl of efficiency to be oom at the end of the fight. If you add healing you WON'T do more FoL. Since in that kind of fights keeping a stream of MAX possible heals is the goal and FoL isn't the top HPS spell.

Once again to make it cristally clear - goal of the fight is pure raw brute etc HPS-stream. The more the better with no cap acheavable with current lvl of gear. You are allowed to cast lower healing spells only to not go oom. Extra healing WON'T let you cast more FoL and less HL. Since your goal is NOT to KEEP some lvl of HPS (thus being able to throw more FoL on some stage if you have more +healing) but to MAX OUT your HPS, +healing will only add HPS and not efficiency.

You are only right, Gal, if we assume there is a cap in HPS after reaching which you can keep it constant by swapping some fol with hl and thus upping efficiency. But there is no such cap.

Well, I hope you agree, that in general +healing -> efficiency conversion is only usable if you do reactive healing or have some soft HPS cap. Only in this case you can do more FoL and less HL (either to keep hps constant, or by requiring less reactive HL casts).

The thing is there is always a max HPS you can do since you can't heal the tank for more than he takes damage. You can overheal more for more safety, but it's not actually doing more healing. Let's say you cast HL11 during stomp, and HL9 otherwise, if you gained enough HPS you could cast higher HPS HL11 during stomp and HL8 otherwise, meaning your "sustained" HPS is the same but your burst is higher. While that doesn't directly say anything about the values of stats, it shows that just about in any situation you can come up with there will be efficiency gains when your HPS goes up. Granted when you run with every mana support possible and never cast a single FoL then it'll be calculated differently*, but unless you're spamming max rank HL there's always some kind of benefit here.

*Normally I'd check how much healing I need to give FoL to match HL's HPS (which is a lot), and how much mp5 I would need to give HL for it to match FoL's efficiency. Then divide the healing by the mp5 to get a healing->mp5 conversion value and add that to the efficiency value the spreadsheet is giving to raw healing. For example for me with a shadow priest I effectively get 1 mp5 from HPS->mp5 for each 12 +healing on top of the spreadsheet's value of 12 healing from the increased HPM.

If you instead get to cast some HL8 over HL9s, then instead of comparing FoL to HL you do the same kind of calculation with HL8 and HL9. You will need a lot less +healing for HL8 to get the same HPS as HL9 and a lot less MP5 for HL9 to get the same efficiency of HL8. In that case, though, since HL8 is not much more efficient than HL9 but is quite lower HPS and has higher downranking penalty, assuming I didn't have any formulas wrong, will take around 150+healing to gain 1 extra "effective" mp5.

With the HPS->mp5 value being much lower for HL downranking than swtiching to FoL, mp5 becaomes actually an amazing stat for efficiency even when a shadow priest is present on fights where you only spam HL (1 mp5 = ~6.7 healing with my gear with 150 healing/mp5 conversion).

This is more about deciding wether you want efficiency or burst HPS rather than how to actually calculate either one of those:
The way I understand the fight from discussions around here, on Brutallus, you don't only care about efficiency, but also care (a lot) about how much HPS you're putting up during stomp, which is not dependant on how much mana you have. It's up to you to decide wether you want to stack healing/haste to maximize ability to heal during stomp or stack mp5 to max your ability to heal during the rest of the time (where you just burn leftover mana with as much healing as possible). So even though for the healing style you're describing mp5 is by far the best stat for efficiency, it definitely doesn't mean that it's what you should be stacking for that fight, but it could be depending on what your raid healers look like... In my experience though, healing in "weaker" periods is much easier for most raids to the point where on some fights I almost feel like I can go AFK while my mana regen, but when bursts happen healing seems to fail very often no matter how much mana they got. So unless your tank keeps dying without even getting a stomp, healing/haste for max reliable burst is probably what you should stack regardless (to help with safety during stomps), and not mp5 for efficiency that only helps with safety *not* during stomp.

Edit: actually changing my cast ratio to 40% mana to HL11 and 60% mana to HL9 using the (rounded value of) 150 healing->mp5 I get 1 mp5 being worth ~8.6 healing which is quite higher than I stated above. This is because HL doesn't scale its efficiency as well as FoL when you gain more +healing. 1 crit rating in this setup is worth ~2.3 healing using 120% effective crit value. Using 150% makes 1 crit rating worth almost 3 +healing. This is only in terms of efficiency, of course - in terms of saving tank during stomp when spamming HL11 crit would still horrible, so if you're thinking crit for HL spamming you should probably just gem mp5 anyway for even better efficiency than crit.

The reason spell crit is preferred for HL healing in Sunwell has nothing to do with its efficiency in terms of regen, though that doesn't hurt. I have a 45% HL critrate raid buffed (lets round to 50% for ease). Thats a 75% average case increase in my throughput. Paladins are the only healer who can consistently heal 40% or more of a tank's HP with one cast.

Armory him, clearly he has not.

It's not 12 healing equal to 1 mp5, it's 12 healing giving you the extra HPM you'd normall account for it plus 1 mp5 on top of it. Obviously this works both ways, gaining 1 mp5 will allow you to increase your average HPS by the same amount +12 healing would if you ignore the extra HPM of that 12 healing (although looking at it this way has a lot less meaning considering the whole reason you don't spam FoL in the firstplace is needing to put up a certain level of HPS). I know it sounds complicated but it gets much easier if you seperate the gains of HPM (which the spreadsheet takes into account just fine) and HPS (which the spreadsheet ignores and you have ot add yourself).

Regarding the crit% I already said I'm still not sure how to estimate it properly. I'm trying to think wether what recount shows is actually a good number or not - hadn't thought of what could be bad about it but getting the full 150% made me suspicious enough to not automatically accept it without thinking about it some more. I would at least like to see its results on other fights as well as using other addons etc. Remember 150% average crit value means that your targets' health deficit on average is higher when your crit heals land, and since target's health deficit has nothing to do with your crit dice it doesn't make sense.

Regarding crit->HPS, if anything 50% crit is the most UN-reliable place you can be at crit-wise, it has highest possible variance you can get for your heal sizes meaning you never really know how much you heal for. I mean, taking it to the extreme, we can all agree that healing for 5k always is better than healing for 4k 1/2 the time and 6k the other 1/2, meaning for HPS purposes, going from 0 to 10% crit is just not as good as increasing HPS by a fixed 5%.


Not to mention crit is quite expensive on stats (which it should be due to its mana returns) making it even less of an HPS stat - with 150% crits at 40% crit you'd need 2.4% crit to increase your average HPS/HPM by 1% and effective mana supply by a bit under 2%, costing 53 crit rating.

Using same budget on healing would provide 116.6 +healing which is (with 2k +healing) ~1.8% to HL11 size and 3% to FoL size, reliably - for HL providing more efficiency but almost 1/2 the HPS and for FoL about 1/3 the HPS and about the same efficiency.

Instead you could've taken 21.2 mp5, or 1538 mana in 6 minutes. With my 133 gear mp5, 300 mp5 shadow priest and chain 3 pots with alch stone and BoW in a 6 min fight I would have 53457 mana, so 1538 mana is 2.9% more mana. About the same efficiency but you can choose when to use that extra mana.

To sum it up if you assume 150% crit value, which just doesn't seem reasonable, then crit is about on par for efficiency with other stats for efficiency. Give it anything less and it drops under. And when you look at burst HPS it's somewhere between not really worthwhile to something similar to a tank stacking avoidance over stamina when he only has 30-50% base - which does increase your chance to be saved but not your ability to handle a worst case scenario when dice roll bad. Not to mention even if you assume you always do your average HPS, crit still falls far behind +healing on HPS when using same itemization points, not to mention how far it falls behind haste which adds even more HPS per itemization point. So while having crit on gear is obviously not useless (I never said it was), it's not something I would purposely stack at the cost of +healing.


On a side note when I say "stack X" vs "stack Y" I obviously don't really mean you take as much as you can of one stat and completely ignore the others. I mean that when you use something like lootzor, the value you will put for the stat that is "stack-worthy" will be generally higher than the itemization value blizzard gave it, while other stats will be lower. How much obviously depends on your personal calculations.

Regarding Brutallus, I never said anything was trivial or not, as I obviously hadn't done the fight. What I did say is that your choice of itemization isn't directly based on efficiency due to you needing to choose between stats that increase your ability to heal through stomps VS stats that increase your ability to heal the rest of the time. If you just maxed your efficiency you would've just stacked mp5 (or alternatively crit if you take crit value for 150%), but we all know even without doing the fight that efficiency isn't all that matters, because depending on the exact numbers in the fight and the exact numbers your healers are putting up, it may be a bit more important to increase HPS during stomp, or instead it may be better to increase total healing done on the rest of the fight.
All I meant to say here is that based on how people say the boss works and the healing required from them, your choice of stats here isn't just max efficiency.

Test it out, don't just guess it is wrong because it makes you wrong. Reset data, heal something that will overheal, use Divine Favor and heal something that will overheal. Then look at the data.

You should do the fight first before saying how we should gear for it.

We really should consider someone with better gear then you. You also don't take into account any consumables, IED, Divine Illumination, or Spiritual Attunement. Consumables are around 40mp5, 50mp5 from IED, 70mp5 from Spiritual Attunement, 80mp5 from Divine Illumination (spamming HL11 during it). Crit gets better the more mana you use. Without a shadow priest I can get 80,000 mana pool on a fight like Brutallus (WWS) with a Shadow Priest I can get to 100,000 (WWS).

You also have to consider the what you will do the mana you get from crits. Say if you have 10,000 mana and get 50% mana back from your heals because of crits (not realistic, but the point holds regardless). After you have casted enough spells to use up the initial 10k, you should have 5k left. Then you spend that mana and have 2.5k left and so on. In this case 50% mana restored will double your mana pool. So if the mana cost of your heals is reduced by n (from 0 to .6) from crits you will actually increase your mana pool by 1 / (1 - n) since it is a geometric series.

So with the situation of 2.4% increase crit rate up from 30% (average of 25% FoL, 35% HL, and 50:50 usage of them). It changes the mana pool increased from 42.8% to 45.8% or an increase of 2.1%. So with 80,000k (mana of 100k parse without Illumination) and 6min fight it gives you 23mp5. It gets a little more complicated with higher HL crit rates, but it benefits crits favor. So on Brutallus for me crit is strictly better then mp5, without even considering the increased healing done.

It is much easier to model crit as reducing the cost of heals, but it is hard to compare that against mp5 though. Since WWS (or anything else) doesn't record any information about ranks used.

I don't get what you mean with mp5 you can choose when you use it, you are never wasting the mana you get from crits.

Now for Gal:

same effective healing % for crit on Brutalus sounds fair. Since to get same % in overhealing for crit and non crit you need tank HP be a function with about the same distribution between max HP and max HP - max HL crit value. As soon as all hits are more than HL crit (and according to WWS most of them are) you gonna get about the same overhealing %. Yes, in absolute value crit will have more overhealing than non crit heals. But in % they both overheal at about the same rate (say 50%). But if both overheal in the same way, it means both effectively heal at the same way. You can say that half of your casts land on full HP tank and half are FULLY used. both crits and non crits.

Crit has less than 50% effectiveness for fights where tank can often get damaged for less than crit HL value. As soon as boss hits very hard as well as his abilities, crit is about 50% more effective then non crit. Saying this I also mean that for stream healing with FoL: as soon as tank has all hits more than about 2-2.2k FoL crit crit is about 50% more effective then non crit. Having HoTs and other healers casted lowers it a bit, but still, I can imagine it to be as good as extra 40% easily.

That is wrong. To be 150% effective you need same overhealing for crit and non crit. Only if you assume that all your non crit heals not overheal your tank, you need HP deficit be higher than crit for having 150% average.

Basically, I can hardly imagine the fight, where tank has less than usual cast value unproportionally more than less that usual crit value. So in many fights, if you heal tank, crit is close to being 150% effective. Again, it doesn't means all your crit heals land and used fully. It just means that amounf of overhealing % for crit and non crit is the same.

Regarding the non-linear formula, I already know this and use it to calculate. It's also stated in the other thread. The more crit you have the better crit becomes, but not by a lot when you're talking about crit chances you actually control. Your choices of gear will not change your crit chance by more than 10%, probably a lot less, and at this level it's close enough to linear (the ^2 (and higher order) factors get very small). By choices I mean things you can change with your gear without actually changing the gear level, meaning gem choices and possibly badge rewards, not going from kara/t4 to t6.

I actually like the idea of looking at the HP distribution. However we don't have enough data to use it to determine crit overheal even if we assume even distribution for all possible HP values or for all possible HP values between 0 and HL11 crit.

For example if it's an even distribution between 0 to HL11 crit, HL11 crits would effectively heal for 50%, while the non-crits would effectively heal for 100% 1/3 of the time and 50% the other 2/3, totalling 2/3 effective healing, which is 4/3 of what the crits are doing. Crits heal for 1.5X though before overhealing is factored in, so 1.5/(4/3)=1.125, or 112.5% crit value...

If it's an even distribution between 0 and 20k where HL11 heals for 5400 and (crits for 8100), using same math:
Crits: 1.5 X ( (20000-8100)/20000 + (8100/20000)/2 ) = 1.19625
Non-crits: ( (20000-5400)/20000 + (5400/20000)/2 ) = 0.865
1.19625/0.865 =~ 1.383 or ~138.3% crit value.

Of course if we try non-even distributions we'll get different values as well, but I can't think of a logically possible distribution that would actually get you very close to 150%. Getting 150% is only possible if the deficit is always bigger than your max crit size, and more than 150% is not possible. If you can come up with a way to get the distribution of tank HP before HL11s land during a fight (and then repeat for other ranks) then maybe we can get something better than just a wild guess.

Except, you did not use it to calculate in your previous post. It can have a meaningful effect that you need to consider. It is completely possible someone to have 10-15% more crit then I do. There is a paladin in my guild who has 10% more crit then I do but 500 less +healing, roughly equivalent in item points. Including buffs on top of that.

Crit Pally: 1.381 to 1.393, 1.2% increase. 46% to 47%
Me: 1.276 to 1.285, 0.9% increase 36% to 37%.

For him crit is worth 33% more then it is for me.

This is totally useless. We have no clue what the distribution of a tanks health will be when your heal hits. Even if we did it won't be linear. That distribution is totally arbitrary and doesn't prove anything.

If I expect a tank to only be down by the amount of my Holy Light by the time it lands, I am probably not going to use Holy Light, unless I am the only healer on him.

Getting 150% is definitely possible, it is not saying crit heals never over heal. It says they overheal by the same percent that non crits do. Say if all my normal heals for 2000, 500 of it overheals. All my crit heals for 3000, 750 of it overheals. In this situation crits are infact healing for 50% while still overhealing.

It is possible to even get 200% it just largely unlikely. One heal heals for 2000, 500 of it is overheal. The next heal is a crit and doesn't overheal.

If you were using 40% baseline how did you get less then 2%? I used 30% and got 2.1%.

So I have actual data showing results and you don't believe it still and cite no reason for why it will be wrong. If you think recount is displaying wrong data, prove it. Go in game and show circumstances where it is inaccurate.

I said it was around 50% give or take 5%, most were mid to high 40s.

I think you are really caught up on even distribution from 0 - 20k. That is an incredibly unrealistic. If I am healing someone there are normally 3 occurrences:

1) My heal got there before any other healers and healed for its full amount (little to no overheal).
2) My heal didn't get there first and was almost all overheal.
3) My heal didn't get there first and my heal partially overhealed.

The part that would explain a much lower then 50% is if #3 the range of the deficit mostly fell in how much a non crit heals for. But that isn't normally the case. HoTs and other small heals will change it some, but not too much.