Divine Hymn only gives one stack, the same with Binding Heal. But Penance gives four stacks, really nice as a Discipline Priest.

And regarding the possibility to wear both versions of the CC trinket, take a look at our Retribution Paladin (The World of Warcraft Armory). He wears the normal and heroic version of the CC ring, so I think the same is true for the trinket.

Those rings says unique-equipped and not unique: <item name>. I have yet to see anyone wearing two items saying the latter.

Well as the post which you pulled those number from states they are based on my priest theorycrafting for 3.2 post. If you are after the nitty gritty of how to calculate each of these numbers use the assumptions from the priest theorycrafting 3.2 post and apply them to the detailed formulas in my set of six theory crafting posts.

My numbers being different from the OPs don't necessarily make them wrong. It depends on the assumptions used. Don't get me wrong, I agree with 95% of what constantius says in his OP, but just because he is the OP doesn't make his view the only valid one.

Just because you don't agree with the conclusion doesn't necessarily make it wrong.

It doesn't, but it doesn't make it right either. My trials on Hard Modes has in general indicated that Haste increases mana consumption, that mana consumption of a active priest is very high as it is, that mana regen is valuable and that mana efficiency is more valuable than faster heals. I did actually try to stack haste/int early in ulduar progress, but ended up using more mana for no real gains over our other holy priest who ran a crit/int heavy set. After correcting my gearing, my efficiency became better with no loss in terms of effective healing.

So far I've seen no reason to change this gearing convention around for Colisseum Hard Modes. There are some fights where less regen might work out just fine, but then those don't really tend to burden the healers anyway, so there's no reason to gear around them.

Comparing the numbers from BobTurkey and the compendium post only makes sense when taking partial numbers from BobTurkey, not his final stat rankings.

Nidaba always only lists relative values of comparable stats (which I prefer).

BobTurkey creates complete stat rankings from his math. As such, Nidabas numbers use far less assumptions and will thus be correct for more healing scenarios, but they won't answer the question "which item is better" easily. If you trust BobTurkeys assumptions (read: you think that your healing reality matches his assumptions closely), then to compare items using his numbers is far less work. The downside is: if you blindly use the numbers and your healing reality differs a lot, you are far from making the right choice.

In either case: just looking at ratios and numbers without carefully reading the explanation makes for lousy decisions.

Fair enough. I gear for crit before haste as a disc priest too.

Your comment I replied to seemed to indicate that you didn't like the weightings because you didn't like Spark of Hope. My mistake.

I think the simple point here is that Bob's value for haste does not include the additional mana consumption in the haste calculation (only the healing gain), and the best solution would be to help improve the calculations.

Looking at

Priest Theorycrafting – Haste Rating (Part 4) � BobTurkey’s WoW Blog

we see that 1 haste gives 0.03% more casts (this is about 1/3279) and ignoring surge-of-light procs and just spamming flash heal that would mean you use 625*5/1.5*0.03% more mana per haste-rating and 5 seconds or about -0.64mp5 (yes, negative weight). Subtracting that from the final stat weightings give haste a negative total value (but I wouldn't put it below zero since you can always slack for the extra time allowed by the haste!).

Nor should it.

Haste does not change the mana consumption per amount healed. It just allows you to cast more spells per time. Whether you do that or not is solely your choice.

In fact, carrying haste may have tiny positive effects on mana consumption. If your scenario is heal-spamming with pauses, the more haste you use, the more time oo5sr you have. However, this effect is probably too small to bother including in models.

I agree on the second part and that it is probably marginal.

I also agree that the choice of using haste or not is your choice, but don't agree with the conclusion that haste is beneficial in a mana-regen set and think there are two different answers:

Short answer: When considering healing gains in a mana-regen set you should not directly add a weight for spell-power, but a weight for healing done per mana, since you are trying to maximize healing per mana, i.e. keeping people alive without going oom (a bit simplified).

Spell-power indirectly helps by increasing healing done per mana and thus can get the same weight as before, but haste does not influence healing done per mana (as you noted) and thus get zero weight.

Long answer: I tried to look at three scenarios:

Assume your cast cycle is not influenced by haste: healing done and mana use are the same (almost). So zero weight.

Assume you spam heal as quickly haste allows: healing done and mana use both increase. To compare with stacking spell-power or crit instead of haste it makes sense to include both effects.

Assume you spend e.g. one third in heal-spamming (during the burst-phases) and two thirds in a fixed cycle uninfluenced by haste. As far as I can see the combined weights is one third of second weight and two thirds of first weight. Since the second weight is negative and the first zero the result is negative (same for other ratios).

Obviously the scenarios are simplified and the goal during a burst-phase is not to maximize healing done but to keep people alive, and haste can help with keeping people alive. That would take a much more complex model and be more dependent on the actual fight.

Let's try with a real world example:

If you have a stack of dollar notes, their buying power does not change whether you spend them at once or distributed over a day. If you have bought everything you wanted or you are out of dollar notes, it's over. At the end of the day you have spent the same amount of money and got the same goods. Now if you assign stat ratings to the buying power of the dollar notes and to the speed with which you can take the notes out of some pocket and hand them over, it's clear that the speed doesn't have an effect on the buying power of the dollar notes. It does have an effect in "first come first served" scenarios, of course, where you may not get the item at all if you are too slow (like players dying because your heals were to slow).

Sure, you can compute a mana burn rate, which is higher the more haste you carry if you are chain-casting all of the time. However, the amount healed in an encounter is always dictated by the damage received by players. The only thing that haste allows you to do is perform the heals faster. You heal them to full faster, but the healers on the whole do not heal more. Thus, you do not have a negative MP5 effect for stat ratings. If you try to do that, I think you are mixing up cause and effect.

Crit and spellpower have a positive effect on efficiency, which is included in stat ratings, and which haste does not have. There's no need to introduce an artifical negative MP5 effect into the stat rating for haste.

Nobody thinks about dropping Divine Fury because it has a negative MP5 effect - is hasn't. As hasn't bloodlust/heroism.

Hmm.. I don't see that there's any real disagreement if we consider efficiency instead of spellpower/healing in the weights:

• We have a pure mana-regen stats where intellect, mp5, spirit, and crit has some positive weights.
• We have efficiency stats in terms of healing/mana where spell-power (and spirit through conversion to spell-power), crit (and thus intellect) has some positive weights.
• We have survivability where stamina has some positive weight.

Bob wants to combine all stats and uses the conversion that 1 spellpower is 0.6 mp5. The combined weights should then also have 0 value for haste, i.e. neither positive nor negative.

The value 0.6mp5 for 1 spell-power doesn't have any real explanation, and if we discuss healing/mana and mana-regen we could get some estimate of how to combine mana-regen and efficiency:

Assume you only cast flash heal, not as quickly as quickly as possible - but only fast enough to not go oom before the fight is over. Casting as quickly as possible without going oom seems like a reasonable assumption for a mana-regen set.

To maximize healing done you can either get more mp5 to be able to sustain shorter time between heals or more spell-power to increase the efficiency of each heal. Say that you on average cast 2 Flash heals costing 625 mana per 5 seconds healing for about 4000 each and 1 spell-power increases the healing done of each flash heal by 1 (including effects from crit).

Then 1mp5 increases healing/5 seconds by 4000/625=6.4 due to increased casts, and 1 spellpower increases the healing done/5 second by 2 (due to the existing flash heals).

So 1 spellpower is worth the same as 2/6.4=0.3 mp5 under these assumptions (and a higher value if you have higher mana-regen allowing more casts).

Unfortunately the assumption about only casting flash heal is far from realistic (and the computations are a bit simplified, and breaks down completely when you have enough mana-regen to spam-cast flash-heals); so I don't know how realistic the value 0.3 is - but at least it should give a starting point. I think it also works under the burst-scenario for healing.

This is only partially true and may lead to false decisions.

1. Haste makes your individual heals land faster (except front-loaded instants, of course). This alone increases the HPS of the single cast without incurring an additional mana cost, and this alone is worth something. I think we can all agree on this.

2. Haste also allows you to squeeze more casts within a period of time, by reducing GcD (within limits) and cast times. Lets look at the two possibilies:

2a. You use this to perform more casts during the entire fight

In this case, you obviously need more mana. The question though is how often this is the case. The fact remains that the amount of effective heal possible in a fight is dictated by the damage received by players and not by the amount haste carried by healers.

In the whole time since starting WotLk, I have only had one combat log where I was constantly casting, and this was our Freya+3 first kill with two healers, showing a 100% activity on my part. Even in this case, though, given more haste, how much more mana could I have consumed? Since only one player died relatively late in the fight, we're obviously talking about less than 10 casts or a few group heals over the course of almost 9 minutes. In fact, given more haste, I might have used PoH more, actually saving mana.

So, if you use option 2a deliberately, you will only achieve that by sniping other healer's heals.

2b. You do NOT use this to perform more casts during the entire fight

In this case (which I consider to be the case for most fights), haste has no negative effect whatsoever on mana consumption.


Let's have real examples of fights where we actually spend more mana by using haste.

I really think the best way to value stats is to first have a seperate metric for sustained hps and peak hps.

Peak hps is clear : that's the maximum hps you can achieve, without taking mana into account. It benefits greatly for haste and spellpower, moderately from crit.

Sustained hps is the mean hps you can produce for the fight (given a fixed fight duration). It takes into account the mana pool, mana regen and mana efficiency, and benefits from int, spirit, spellpower and crit at diverses level.


A few advantages for these metrics are the following :
1/ Most fights have healing intensive phases, and healing light phases. You need both the hps to survive in the high damage phase (that's peak hps), and enough mana to last the fight. However, you won't be spamming for the whole fight : mean hps measures wether you have enough mana to heal at the needed level for the whole fight.
2/ You can measure how much more mana would be useful for you. More mana only increases your sustained hps, with the limit of your peak hps. If both are close, it means you won't need more mana. If there is a huge gap between both, you can be oom fairly easily.
3/ If you want to merge them in a single metric, that's quite easy to do. You're adding apples and apples : it's a lot more clearer than adding mp5 and peak hps.
4/ The only "parameter" is the fight length, which is common parameter (for example to value the weight of int). It is quite a "fixed" parameter for most fight. Naax fights are not much shorter than coliseum fights, they just need more hps.
5/ You don't need to "twist" weight to correct / adapt special effects. You may just twist the peak hps / sustained hps ration, depending on what you want...

You are looking at the equation from the wrong side.
If You are setting a fixed Raid-Heal-Performance and calculate backwards from there, you obviously do not need more haste other than increasing personal HPS to render one of the attending healers obsolete or to make up for slacking performance of other people just because the encounter is trivial.

The question is to value haste for encounters that are not trivialised or yet have to many people waiting for the light at the end of the tunnel (aka rez).

Sure. Still, the point in question was whether haste has a negative effect on total mana consumption. This is what I reject. I agree that, especially during progression, there is a lot of avoidable damage that needs to be healed. But this is exactly what dictates mana consumption. Not haste. Thus haste does not have a negative effect on MP5.

Now, one can argue that, in specific progression encounters where you struggle to deliver the total healing required to keep everyone alive, haste enables you to perform more total healing if you have the mana to do so. Still, there's no negative MP5 effect, it's just that heal per mana doesn't go up by using haste.

But is this really true? Do we really fail on progression because we cannot deliver the total healing necessary for a fight due to mana constraints? Or do we fail because we cannot keep up during burst phases or because we're not quick enough to react to player mistakes (including ours)? Would mana have sufficed if we had been able to save these 3 damage dealers, just due to the shorter combat duration? Who knows.

Anyway, stat ratings need to watch out for cause and effect. Just one set of ratings is not what I would recommend or use. But if someone wants to use just one set, these still need to be built around actual benefits of a stat. If one starts arguing that haste has a negative MP5 effect, then one can also argue that MP5 has a positive effect on an amount healed. And then we're in really deep shit.

No, if using just one set of ratings, we can only say: okay, MP5 and haste are not comparable at all, still we want decide what to purchase, so we attribute a positive value, because both are fully beneficial. We say: MP5 is good, and haste is good. We try to find a value at which comparative amounts we would choose MP5 over haste. But we should no try to do actual math proving how haste affects MP5. It doesn't (except for tiny positive aspect I mentioned).

(I agree with everything else said outside of the quote.)

Mp5 has a soft cap. That cap is reached when you hit 0 mana at 0 Boss health (assuming optimal SF, Manapot, Inervate etc. usage). Every point of Mp5 above that is theoreticly wasted. It's a total number that is effected by Encounter length, Raid Performance etc. One will theoreticaly want to gear for the worst case scenario (aka Firefighter, Algalon for example)
This cap can not be evaluated to stat rankings.

Not always true. For some fights your "crunch time" might come mid fight, where you need more mana to sustain peak HPS longer, even though you will eventually regain enough mana to finish the fight with a surplus. I believe this is especially true for holy priests who are capable of burning mana faster than any other healer I've seen. A good example might be going into Beasts undergeared, you're likely to run oom during p2, even though you'd have no problem finishing off p3 with mana remaining. So in effect the Mp5 softcap should be considered "ability to maintain sufficient sustained HPS throughout the fight whilst being able to sustain max HPS for the required periods of the fight without running oom at either critical point". Technicality, but it sort of emphasizes gearing regen past the fightlength.

So I have to agree with this:
Trying to compare the throughput and regenerative aspects of any stat can not be done with confidence. In all cases what you actually have is a change to the healing output "throughput" or to the mana regeneration "regen" profile of your character. These two items are linked only in the ability to perform you role as a healer through potential mana consumption. To truly compare stats you can only compare the relative throughput value of a stat to other stats relative throughput value and likewise with regen. When you begin making comparisons at the macro level you have to make certain assumptions that will not be true in general.


Mana efficiency & Mana regen
For those stats which improve mana efficiency this is often times correlated directly to a mana regen component. Which I find to be incorrect and entirely misleading.

- Mana efficiency is a potential to reduce mana consumption during the course of a given encounter.

This potential is not realized at all until you reach a change in healing output that would guarantee that you can drop at least one full cast. Example: If you can not cast 1 FH where you at one point required 2 FH then you really have saved 0 mana and just provided a higher throughput resulting in generally higher survivability. If you still cast 2 FH even a more mana efficient FH has saved you 0.

So until you can fully quantify that you will cast X-1 casts vs. X casts you have netted no actual mana consumption gain. Furthermore even at the point that you have reached a consumption gain you have not improved regen you have only reduced the regen requirement of the encounter. The regen requirement is a variable that can not be strictly defined as it relies on the damage pattern, playstyle, strategy, and general raid performance (more dmg income -> higher healing requirement -> higher regen requirement, impacted by spell selection and assignment).

It is important to understand that one of the goals in stat & gear balancing, is to balance Mana Regen with Mana Consumption for a given encounter. This is an indication that at no point during a given encounter does the mana consumption requirement exceed the mana regen capability to a point where a negative mana pool state is reached. This is similar but not equal to the concept that when boss health equals 0, mana pool can equal 0... Due to the nature of encounters some will end with a large mana pool based on the damage pattern. Examples: NRB phase 2 is more mana intense than phase 3, Mimiron phase 2 is far more mana intense then phase 3 or phase 4...


Throughput Stats

Haste: Increases Peak HpS capability. No impact on heal efficiency.
SP: Increases Healing output per cast. Per spell has a varying impact on heal efficiency.


Regen Stats

Mp5: This is the only pure regen stat. Having more mana does not increase throughput, it only provides the means to balance mana consumption.

Mixed Stats
These stats impact both the mana regen & the healing throughput by varying amounts.

Int: Provides a throughput component via the Int -> Crit conversion. Provides a mana regen component and increased mana pool that work towards balancing consumption.
Crit: Provides increased throughput on average and is a pure throughput stat for Discipline Priests. For Holy Priests this has a mana regen component when Holy Concentration is taken.
Spirit: This is primarily a regen stat and is a pure regen stat for Discipline Priests. Impacts throughput for Holy Priests due to a SP conversion.

Summary

To fully optimize evaluation of gear selection throughput and regen should be evaluated completely separate. With the resulting goal to balance not stats as a whole via one direct stat comparison but by the following two comparisons.

HpS Required vs. HpS Capability

Mana Consumption Required vs. Mana Available (Mana Pool + Mana Regen)

These two comparisons will be different for each player and for each encounter. Any generalizations of stat comparisons need to take into consideration that these two comparisons are where any conversion between throughput and regen stats can be made.

Mana Regen, only increases HpS Capability through the support of mana pool and ties into the longevity of said output. At small mana pools and low regen it is a topside bound and a longevity limit @ specific output rates.

Healing Throughput, only increases mana regen through the relationship to mana efficiency gains in balancing consumption verses regen.


SIDE NOTE: Haste can by the means of others combined calculations be deemed a benefit to mana regen at the point where you would replace X FH with 1 GH... Because of the HpM gain of GH it results in a reduced mana consumption and hence a virtual mana regen. This is primarily true for Holy Priests as the HpM difference for Discipline is not as great. At no point does haste reduce mana efficiency nor reduce mana regeneration.



The traditional sense of soft-cap in my opinion is "A point at which adding more of a stat provides are reduced gain in its effectiveness". For a stat such as Haste there is a point of soft-cap that occurs when the GCD has reached 1.0 second. Because further stacking of the stat no longer impacts a portion of its value, reducing cast time and the gcd. To that point there is no soft-cap for Mp5 because it will not ever change the amount of mana it provides unless you are already at 100% and further regen becomes absolute waste (which rarely is the case during an encounter). You could further restrict it to say that once regen exceeds the absolute maximum consumption or average consumption you reach a soft-cap, though this is well in excess of any realistically achieved regen rate and would likely result in a state of permanent 100% mana.

I do understand that what you are trying to state is there is a crux point where additional regen becomes unnecessary and hence "wasted". Though extra resource availability is usually a blessing and not a curse. At the point that you could reassign stats from regen -> throughput, you may still be impacting regen required to some degree due to the relationship of mana efficiency to mana consumption. Generally you are by your definition at the soft-cap for throughput though and it would be equally wasteful to some degree to put stat points there either, unless you are trying to run lighter on healers. Because once your throughput effectively exceeds the healing required per second for your portion of the raid healing you have reached a "soft-cap", until you can successfully run a full healer shorter and thus increase the healing required per second.

Well, the only problem with your grouping of guilds, is fairly simple.
You didn't exclude the guild's who did this fight pre-nerf to Mimiron (which Mim HM was the main reason for healing priest's to take this ability). I think only 2 guild's in your list downed Mim before the nerf to all of his magic dmg based abilities.
So that leaves:
Juggernaut:

Phenominal: Spell Warding
Tsigo: Both (This spec lacks inspiration)
Xandrine: Spell Warding
Horky: Spell Warding
Togarth: Spell Warding

Vigil:
Sholtez: Spell Warding
Sear: Divine Fury
Alvism: Both (This spec lacks inspiration)
Puresinette: Spell Warding

Which is 91% i think. But IM, without Inspiration....

"For those stats which improve mana efficiency this is often times correlated directly to a mana regen component. Which I find to be incorrect and entirely misleading.

- Mana efficiency is a potential to reduce mana consumption during the course of a given encounter.

This potential is not realized at all until you reach a change in healing output that would guarantee that you can drop at least one full cast. Example: If you can not cast 1 FH where you at one point required 2 FH then you really have saved 0 mana and just provided a higher throughput resulting in generally higher survivability. If you still cast 2 FH even a more mana efficient FH has saved you 0."

True, but I’m not sure I see your point. I can’t think of a way to raise efficiency without also raising throughput. Even if the efficiency doesn’t pan out, you still benefit from the increased throughput. Also, in a multi healer environment, it might not take much increase in throughput to save someone that FH cast. It might not be a savings to you, but saving your fellow healers mana benefits the whole raid.

Good points. And I largely agree with Hegen.

Haste has two characteristic worth considering (IMHO) for trying to determine its value:

1) How much additional throughput can be generated by additional casts. Partial casts need to included because we are looking over the duration of a boss fight (5+ minutes) and because the aim is to quantify haste rating across multiple gear pieces in a generalised manner I doubt we can come up with a decent model using only complete casts. We are effectively trying to determine the incremental value of 1 haste rating.

2) Just-in-time casting - this is the function of haste to allow you to cast a spell faster to make a save. I.e. haste allowing you to land that heal fast enough to prevent death.

I don't think number 2 can be quantified although i'm happy to be proved wrong. If anything its probably more effected by player skill and lag than haste rating, but thats another discussion.

Number 1 is worth discussing and I think it can be quantified. If you discount number 1 because you disagree with the assumptions (chain casting over a long time frame of just one 'model' healing spell) then you are basically saying either haste has no value or that it has no value that we can determine. Neither are satisfactory to me. I totally agree the assumptions are not ideal, but unfortunantly any model except the most basic, requires assumptions.

I agree with The Doctor that throughput and regen should be evaluated seperately and then combined usign some sort of balance. So calculating the values of each and part then saying I want 30% regen and 70% throughput and calculating a final weighting for each stat. I know this is the way The Doctor has done his weights before.

I'm not sure however if it is understood that this is effectively what I have done by setting the SP:MP5 ratio. So for 3.1 when I had this set to 0.6 (0.6:1) I was effectively gearing with the balance more to regen and in 3.2 with this set to 2.0 (2:1) I am effectively gearing more towards throughput.

So given this and the discussion above, can anyone suggest a more effective way of determining a value for haste that can be applied generally than this below (quoted from my earlier post)?

The real value of haste is what we can call the cascade save. This is where the small increments you save on each cast add up over 2-3 casts to being able to put out another heal on yet another different player who would otherwise have died. Proving that haste made the save mathematically is only doable is a single healer scenario but the effect is still there in multihealer scenarios. The problem is that in multi healer scenarios the it may be your haste allowing another healer to get in the critical heal. It isn't that haste has increased your HPS (which it has) but rather that it has improved your ability to distribute your HPS over multiple targets.

A perfect example of this is the Temper Tantrum where players take sustained damage over several seconds. If you have enough haste to lop 0.3 seconds off every GCD by the time you have cast 4 unhasted spells you could in fact cast 5 hasted spells on 5 seperate targets.

This is an increase in mana usage but as has been pointed out this is healing that will have to be done by somone so it hasn't increased your mana drain one bit, however Johnny rogue who stood to close to the spark didn't die so in fact the fight ends faster and you need less overall regen. Can you math this out and quantify how valuable haste is in this scenario? Not really. Up until your soft cap more haste is always a good thing in a fairly linear way.

Trying to compare regen stats to througput stats is like comparing Horsepower to Gas tank size. Gas tank size is vitally important right up to the point where you make the trip without running out but has no direct correlation to the Horsepower of your engine. (which uses the gas at the same rate regardless of the size of the tank.) I think that most experienced healers know how much regen they need for thier raid and no ammount of dickering with stat weights changes that. As has been said before you gear for regen until you don't need any more for standard situations. Then you focus on throughput. This throughput could be crit or haste or Spell power and which one is best is going to depend on your spec and what your raid asks your to do.

Despite nurfs and talent changes my raid buffed regen has hardly moved since l got fully epiced out, what has changed is that I have a lot more haste/SP/Crit and I still end most fights with roughly the same reletavie ammount of mana.