I originally posted this in the Holy Priest Theorycrafting thread, however that thread has a lot of running topics (especially "The Great CoH vs. DS Debate" ) and I can see how something very math-oriented would be hard to keep going in that thread. So, that said, decided this would benifit from having its own discussion space.

Anyhow, as I've been spending more and more time raiding on my Priest, I've been trying to look at the potential min/max options for my gear. I've always been a fan of regen, and it served me very well when gearing up--however, now that I'm rather solid at just over 1900 healing/200 mp5 non-raid, I feel a bit more flexible about playing with stats to see what is "best."

Of course, this lead me to lootzor.com as it's quite handy for looking at things in a numeric sense... however, the quesiton is always "how to weight stuff?" (This also applies to Pawn, in-game.)

Realistically, a lot of this goes by feel. Math can't prove everything in regard to healing, but -trying- for a moment to look at things from a pure mathematical perspective, I came up with the following rationale:

-Healing would be the base value to compare everything against (1)
-Mp5 would be valued based on its impact of downranking manacost vs. healing lost
-Spirit would be based on various multiplier of Healing and Mp5 (Imp. DS/Meditation/etc.)
-Int would be based on a multiplier of Mp5 over an 8 minute boss fight

The numbers I ended up with are:

Healing
Healing Value = 1

Mp5
Assumptions:
5/5 Divine Fury
3/5 Empowered Healing
3/3 Improved Healing
5/5 Spiritual Healing

GH7 gain per heal: 1.14
GH2 gain per heal: 0.93
GH1 gain per heal: 0.83

(These numbers were pulled from my current values, substitute if needed)
GH7 701 mana -- 4934 avg
GH6 637 mana -- 4568 avg
-64 manacost (128 mp5), requires +323.7 healing (2.52 mp5/heal)
GH5 603 mana -- 4397 avg
-98 manacost (196 mp5), requires +471 healing (2.4 mp5/heal)
GH2 386 mana -- 3054 avg
-315 manacost (630 mp5), requires +2022 healing (3.2 mp5/heal)
GH1 314 mana -- 2584 avg
-387 manacost (774 mp5), requires +2831 healing (3.65 mp5/heal)

This is a bit variable, depending on what spell you use the most. Given that GH2 is my favorite downrank, I will use a value of 3.2. This may be slightly different for those who use different ranks.

Mp5 Value = 3.2

Spirit
Spirit is the tricky one, because it affects so many things. Impact from Spiritual Guidance/Imp. DS needs to be calculated, as well as Mp5 due to Meditation and potentially the PMC bonus if you use it. Also, one's average OO5SR time may vary heavily from person to person. This section may require heavy substituation of values, so I provide the whole formula below.

Imp. DS + 3/5 Spiritual Guidance Contribution:
Healing Value * .25 = 0.25

OOC Regeneration:
0.625 mp5/spirit (while not casting)

5SR Regeneration:
0.09375 mp5/spirit (with meditation)
0.125 mp5/spirit (with meditation+PMC)
0.1875 mp5/spirit (with 30% meditation, patch 2.3)
0.21875 mp5/spirit (with 30% meditation+PMC, patch 2.3)

Approx. Time OO5SR = 30% (This may vary)

Formula:
(0.25 * Healing Value) + (5SR Regeneration * Mp5 Value) + ((0.625 - 5SR Regeneration) * OO5SR% * Mp5 Value)

With PMC:
Spirit Value = (0.25 * 1) + (0.125 * 3.2) + ((0.625-0.125) * .3 * 3.2)
Spirit Value = 0.25 + 0.4 + 0.48
Spirit Value = 1.13

With PMC (Patch 2.3):
Spirit Value = (0.25 * 1) + (0.21875 * 3.2) + ((0.625-0.21875) * .3 * 3.2)
Spirit Value = 0.25 + 0.7 + 0.39
Spirit Value = 1.34

Without PMC:
Spirit Value = (0.25 * 1) + (0.09375 * 3.2) + ((0.625-0.09375) * .3 * 3.2)
Spirit Value = 0.25 + 0.3 + 0.51
Spirit Value = 1.06

Without PMC (Patch 2.3):
Spirit Value = (0.25 * 1) + (0.1875 * 3.2) + ((0.625-0.1875) * .3 * 3.2)
Spirit Value = 0.25 + 0.6 + 0.42
Spirit Value = 1.27

With PMC + The Human Spirit + Spirit of Redemption
Spirit Value = 1.13 * 1.10 * 1.05
Spirit Value = 1.3

With PMC + The Human Spirit + Spirit of Redemption (Patch 2.3)
Spirit Value = 1.34 * 1.10 * 1.05
Spirit Value = 1.5477

Intellect
Fight time: ~8 minutes (480 seconds)
1 Int = 0.15625 mp5

Formula:
(0.15625 * Mp5 Value)

Intellect Value = (0.15625 * 3.2)
Intellect Value = 0.5

Final Numbers
Healing Value = 1
Mp5 Value = 3.2
Spirit Value = 1.06 (no PMC), 1.13 (PMC), 1.3 (PMC+SoR+Human)
Spirit Value (Patch 2.3) = 1.27 (no PMC), 1.34 (PMC), 1.5477 (PMC+SoR+Human)
Intellect Value = 0.5

This generates this list (or this list for Patch 2.3)

(For those of you who use Pawn, this would come out as something like: (Pawn: v1: "Healing": Healing=1, SpellPower=1, Intellect=0.5, Mp5=3.2, Spirit=1.3, RedSocket=18, YellowSocket=18, BlueSocket=18))

Of course, these formulas were thrown together somewhat quickly and it's very highly possible I made some glaring error somewhere, or missed something important. So, wouldn't mind people's feedback and thoughts as to how this kind of thing could be calculated.

As I said above, certainly healing is one of those "by ear" things in some cases, so math alone isn't always the answer...but given that lootzor is an interesting tool--especially in how it weights gems/socket bonuses automatically--I would like to investigate this angle a bit further.

(I must admit that prior to now, I had rated Mp5 a bit higher than it has been derived by the above formulas. I have been going with an assumption of around the 4.5:1 mp5 to healing rather than 3.2. So, I'm challenging a bit of my own feelings here--which is another reason why I'd prefer some more sets of eyes on it!)

One thing I did not mention in the other thread was that if these numbers get confirmed or worked on some more, so that there is some feeling that there are some resonably conclusions at work, I will probably make a more in-depth spreadsheet that can calculate these values on the fly, especially as talents, current +healing value, and other external factors will alter the weightings somewhat. I didn't want to start work on the sheet until the numbers had a bit more work, however, so feedback would be greatly appreiciated.

Unfortunately I do not have time to post all my math, but I did similar analyses. I use a 6 minute fight assuming 35% time spend outside the 6-second rule. I also assume improved Divine Spirit, Spiritual Guidance, Mediation, and Spirit of Redemption. I don't assume Blessing of Kings or any tailoring benefits.

Here are my relative rankings:
Int 15
Spirit 21
MP5 72

These values are not far from yours, relatively speaking. Int is higher for me probably because I assume a shorter fight.

It is hard to compare +healing to mana directly, but I usually rank 1 point of healing as 10 in this scale. So 2 +heal is 1 +spirit, or 5 mp5 is as good as 36 healing. In your scale you value healing twice as much (5 mp5 equivalent to 18 healing), but I feel that this comparison depends a lot on the individual and their style. For me I hate running out of mana, so I favor regen a bit.

edit: nvm, to much of a derail really.

The numbers I came up with for Lootzor are based on a different model. I assumed that my preferred spell will be Gheal rank 3, 2/5 Tier 5 bonus and the same build as you. Next I looked at a WWS parse for me in SSC (everything bu Vashj) and noticed that 80% of my greater heals procced the 2/5 t5 bonus. Hence I subtracted 80 mana from the cost of gheals for the purpose of these calculations.

+Heal

Now for a rank 3 Gheal the coefficient for healing from gear is 0.95x1.1x1.2 = 1.254 . So that is the value I use for healing in Lootzor.

Mp5

I had a bit of an issue coming up with the exact number for this and ultimately I think it is very dependent on playstyle. If you are chain casting heals then the it would be simple to just look at 0.5 x Average_Heal / Mana Cost = 0.5 x 3595 / 383 = 4.69 . The 0.5 in the formula is a result of the fact that 1mp5 would generate 0.5 every 2.5 seconds which is the cast time of a gheal.

However, I find myself using chain casting in combination with a stopcasting macro frequently so I looked at a fight like Morogrim, where I am using gheals almost exclusively (with a few Binding Heals thrown in), to see actually how often I am finishing a gheal cast. The number I came up with was every 6 seconds. Which means that 1mp5 generates 1.2 mana for every cast of gheal rather than 0.5 . This bumps up the value of 1mp5 to 11.27 .

Within this range I am inclined to pick a number closer to the lower end because mp5 would be a more relevant stats in a chain casting situation. The value I use for Lootzor is 5.2 .

Spirit

I assumed 80% time spent in FSR (again from WWS) and the calculation was straight forward. Each point of spirit adds 0.25 healing and then used mp5 value to calculate the contribution from mana regen.

0.25x1.254 + 0.8x0.625x0.15x5.2 + 0.2x.0.625x5.2 = 1.35 .

Intellect

My way of calculating the value for Intellect is the same as Jayde. However, I think he made a calculation error. The value I am getting is (15/480)x5.2 =0.1625 .

The final values I get are as following

Healing = 12.5

Mp5 = 52

Spirit = 13.5

Int = 2

The interesting thing is that once you scale my values relative to healing, they match up pretty well to Jayde's even though our method for calculating the value of mp5 was quite different.

Edit: In 2.3 Spirit should be valued at 17.4 .

Thanks for the feedback. It's 2:45am and my brain is a bit too slow to run through your numbers completely.. I will do so in the morning and post my thoughts.

I did want to note about Int, though, is that as I was scaling it based on Mp5, it was:
15/480*5 to multiply it accordingly. 15/480 would be mana/sec, instead of Mp5. (I think!)

I don't think Int is really an extremely good stat for Priests, although it can be OK as an additional value here and there. You want enough Int to be able to chain-heal when needed, but realistically Healing/Mp5/Spirit are generally more effective as far as budget is concerned for increasing both throughput and longevity.

As you note, I do find it interesting that although we calculated things from a different angle the ratios between the stats are very similar--especially if you multiply your int value by 5. That does give me a bit more confidence that the numbers are on the right track!

In your situation, I think the T5 2-set bonus really does make a big difference, and that wasn't something I tried to model quite yet. Shooting for hopefully getting that bonus myself, so I will run some numbers to see how it would affect my model tomorrow.

I've been playing around with numbers like this for quite a while now - here's my thinking:

Let's assume:

• 75% of your time is spent inside the FSR (call this the duty cycle)
• The fight lasts 360 seconds (6 minutes)
• You use 4 mana potions during this time
• You get 0.125 mana per spirit per second outside the FSR
• While raid buffed you have 11k max mana, 520 max regen, and 250 casting regen

The total amount of mana you have to play with during the encounter is given by:
= Max_Mana
+ (Duration/5)*Duty_Cycle*Casting_Regen
+ (Duration/5)*(1-Duty_Cycle)*Full_Regen
+ Num_Portions*2400

With the above assumptions, that gives us 43,460 mana.

How are we going to spend that? Well, that's highly variable, but for the sake of argument let's assume we're using a mix of GHeal 7, GHeal 2, and Flash Heal. If we assume that we use a 15%/35%/50% mix, and reasonable holy talents, then our "average" heal costs 475.25 mana, hits for a base effect of 1430.8 HP, and has a coefficient of 79.68%.

Now, if we assume that you've got around 1800 raid buffed +healing we can work out the average effect of each heal. Then if we divide our total mana by the cost of the heals we know how many of them we'll be casting. Do the math, and it works out that over the course of the fight you'll be pumping out 130,844 HP in healing.

There's our baseline - now we can run some what-if scenarios to calculate the effects of increased stats. To do this, however, we need to make some additional assumptions, since the actual effect of 10 spirit (for example) is highly dependent on various talents, racials, buffs, and set bonuses. Let's assume:

• You get a 10% bonus to all stats (BoK)
• You get a 15% additional bonus to spirit (human racial + SoR)
• You get 15 mana per int (no Mental Strength)
• You get 25% of spirit as +dmg/healing (5/5 Spiritual Guidance)
• You benefit from 20% of your spirit while inside the FSR (Meditation + Whitemend)

Now we can calculate the actual effects of adding an extra 10 more int, spirit, mp5, or +healing. If we do that, we find that adding 10 of a given stat gives us the following extra HP healed over the course of the fight:

Crunch the numbers to normalize them and you get weights of the following:

Stamina, obviously, doesn't help you heal directly, but it's still useful. I tend to value it as half as useful as int, but that's obviously arbitrary.

The problem with these weights is that, as with any weights, they're hugely dependent on talents, gear, healing style, and other factors. What's perfect for generic trash healing won't work nearly as well for CoH spam on VR, for example. They also emphasise total throughput over burst healing, and so possibly undervalue +healing slightly.

Healing style obviously impacts stat weightings in a significant way, simply due to the fact that the nature of mp5/healing/spirit are more (or less) useful in certain situations.

I mean, for instance, I'm not sure if your "15%/35%/50%" example was meant to imply using Flash Heal 15% or 50% of the time--but I probably use it less than 5% of the time myself, and am lucky to get off more than 1-2 Flash Heals per raid. (Although I use Binding Heal a fair bit.) If I'm assigned to off-healing, I'm probably looking at something like 80% GH2 and 20% GH7--whereas if I'm main-healing I'm looking at probably 90% cast-canceled GH7s.

Your way of calculating via healing pool is quite interesting, though. However, the ratio of spells probably has a large impact on it due to the HPM rates involved. e.g. heavy use of GH1/GH2 will make mana/regen more valuable due to the increased HPM of those spells when compared to something like Flash Heal or even GH7.

Anyway, very interesting thoughts... I find it to be a bit of a cool topic simply because there are so many ways of calculating the same thing. That said, in all 3 serious examples in this thread, the ratio of stats is quite close to each other--which leads me to believe that the general ratio is somewhere at least in that range. Even though the method of calculation is quite different, the jist seems to be in a somewhat tight range of values.

1, 3.2, 1.3, 0.5 vs. 1, 2.9, 1.14, 0.64 for our numbers aren't really that disparate. The 1, 4.16, 1.08, 0.8 numbers from Irise more heavily weight Mp5, but considering that he is specifically taking into account the T5 2-set bonus (which obviously dramatically changes average HPM ratios) that deviation makes a lot of sense.

Either way, there seems to be some consensus that Mp5 is generally worth between 3-4 +healing, depending on the scneario, that Spirit is worth between 1.1-1.3 +healing depending on race, talents, and set bonuses, and int is worth somewhere in the 0.5-0.75 range, depending on the length of the fight. It is, at the least, a generally good ballpark.

Also, interestingly enough, other than changing the order of a few items that are -nearly- identical anyway, the Lootzor list for your vs. my numbers is virtually identical in every way. (Only real difference is the preference of +18 healing vs. +9 healing/2mp5 gems in some pieces of gear. For instance, the slight difference seems to make it values 3x +18 in the Leggings of Eternity better than 3x +9/2mp5 gems to get the +9 healing socket bonus.)

Oops. Embarrassingly, I just found an error in my spreadsheet. I've corrected it, but now I'm wondering if there's others I'm still missing... My weights change fairly dramatically too, and they don't "feel" quite right to me now. *sigh*

Irise: With the assumptions above, I get the following weights:

About what you'd expect I'd think - as fight length increases, int drops and regen increases.


Jayde: I got my spell mix by running a few raids, logging them, parsing them, and seeing what I was actually casting. At the time, about 50% of my heals were flash heals. High, perhaps, although in my defense I'd note that we were running with 0 resto shamans, 1 tree with poor attendance, 1 holy pally, and 6 holy priests at the time. Weird, I know.

If I return to a 6m assumed fight, use the assumptions in my original post, but plug in 80% GHeal 2 and 20% GHeal 7, I get the following weights:

Again, about what you'd expect - your healing mix has a higher weighted coefficient, which increases the value of +heal compared to other stats.

My numbers aren't too close to yours, but I see you made some different assumptions. If I plug in an 8 minute fight and a 70% duty cycle, I get:

That's closer, but still off a ways. Basically, you appear to value +healing as being about twice as good as I do. I'm not quite sure why (although I wouldn't rule out a math error on my part ). I would note, however, that my values are surprisingly close to the itemization costs. Normalizing with Spirit = 1 we see:

In other words, my model throws up numbers very close to how Bliz weights stats (not counting int, since it doesn't fit the pattern). Coincidence, I'm sure, but amusing anyhow.

Edit: If anyone is interested, you can find a copy of the spreadsheet here. Comments, suggestions, and corrections welcome.

Hmm, well your new set of numbers is considerably different (nearly double the regen value! whoops!) so I will have to consider that some more...

I think the primary difference in our two method is that your is based on healing pool, which obviously favors regen... however, my regen calculation was primarily based on downranking parity, which is favoring maintaining GH7-level throughput.

There is certainly merit in your system, healing pool as a raw value is certainly a nice absolute value one can compare to. However, I would be concerned that it may value regen a bit too highly compared to +healing.

The other factor that I did differently is that I look at mp5 vs. +healing on a per-heal basis, meaning the length of the fight has no impact to the calculation of mp5 vs. healing. This might be the wrong way of looking at it, but again...basically the idea that my throughput/consumption would be a constant, with downranking used to compensate the difference.

Like I said in my OP, I'm not 100% convinced that my way is particularly "correct", but I can see arguments made for both. I suppose ideally a hybrid system would be ideal, but one would have to know a "target HPS" throughput to maintain, then figure out what you would need to meet that goal. It gets a bit complex in that case, however, because downranking requires X +healing to save Y mana, however X or Y may not always attainable numbers. You would have to consider a certain assumed statistical baseline, I suppose. (I can see this being some really fun spreadsheet logic here. )

The flaw with only looking at healing pool is that in many situations (especially MT healing) one must occasionally (or often?) maintain a throughput that is higher than what one could maintain for 6 minutes using the highest HPM available, which hurts one's ideal healing pool yet is often required. Another thing to consider in raids if looking at healing pool is external mana regen sources--e.g. Shadow Priests, mana pots, Blessing of Wisdom, Flask of Mighty Restoration etc.--as those would potentially allow less gear-based mp5 while maintaining a rank with a higher coefficient, thus potentially increasing the value of +healing.

Another issue is, of course, that regen and healing "work together." For instance, the higher one's +healing, the more the value of regen will be if one maintains that level of +healing without sacrifice. This is something really important to consider if you're going to look at healing pool.

I will ponder this when I get home and get access to some spreadsheets.

I personally use the following weights for Pawn and Lootzor:

22.23 Red
19.76 Blue
16.06 Yellow
32.12 Meta
4.82 Mp5
1.23 Healing
1.99 Spirit
1.23 Intellect
0.07 Mana
0.45 Stamina
0.04 Health

(Weighted out of a total 100, for consistency's sake)

They are roughly equivalent to your's, Jayde, except that mine values intellect a little bit higher.

If I normalize for 1 healing, I get:
1 healing = 1 int = 0.62 spi = 0.26 mp5 = 2.73 sta

Hmm.

Yeah, I'm not sure how useful a "total healing throughput" model is, since there are times when burst healing becomes important. I'm not sure how to model that though - it's easy to construct a function to measure total healing throughput, but I'm sure how to measure burst healing ability at all...

To some extent, we can model burst healing just by tweaking the assumed heal usage. You mentioned than when off healing you use 80% GH2 and 20% GH7, while when MT healing you use almost all GH7. It's easy to calculate the average cost, effect, coefficient, HPM, etc of the heals you use - so can any healing style be modeled based on these numbers plus an estimate of time spent in the FSR? And if not, what else do we need to know?

The point on additional sources of mana is an extremely good point - the more mana you have to use, the more heals you can cast, and the greater the benefit of each point of +healing. I've been counting potions and buffs, but ignoring shadow priests, set bonuses, trinket procs, etc.

Well, what is a good number for total effective raid buffed mp5? I was assuming 420 mp5 outside FSR, and 250 inside. If we add, oh...200 mp5 for a shadow priest, and 60 mp5 for 2p T5 (both of these are just random guesses), then that brings us to 680 mp5 outside FSR, and 510 inside FSR. For an 8 minute fight, mostly casting GH2, I get:

Those numbers look better - and closer to those others are posting.

Random thought: Perhaps instead of modeling, we could look at simulation...

Hmm?

It's how I value stats/gear, for every 100 healing you (I) should also gain 10mp5. The goal for most would probably be around 2,000 healing and 200mp5. As for Spirit, Int, and Stam, well.. have fun with lootzor :P

A simulator is an interesting idea, and I'm almost tempted to work on making one if I can get some spare time over the weekend.

There are some "tricky" things such as Clearcasting -> OO5SR issues to consider, in addition to Clearcasting -> Inner Focus -> cancel-cast OO5SR tactics... those things are hard to average, but could be easier in a simulation environment.

But, yes, I think a major point to consider with the healing pool model is simply that there is often a minimum HPS threshold needed in most situations. For instance, if you are raid healing an the average damage intake is 3-4k in short intervals (think trash healing, Karathress off-healing, Morogrim off-healing w/o CoH, etc.) you will probably be chain-casting a lower-rank heal that you want to land for at least 3k, or around 1.2k-1.5k HPS.

While perhaps adding primarily regen instead of +healing could increase your heal pool over a 6 minute fight, you may not actually be able to cope with maintaining 1.5k HPS for a block of time without having to uprank, and thus lose mana quicker.

For example, plugging in some "random" numbers into your sheet... Using 100% GH2, 100% duty cycle, 480 second duration: 100 mp5 and 3000 healing would result in 288158 healing. 300 mp5 and 1000 healing would result in 263867. 200 mp5 and 2000 healing would result in 299567. (Well, technically your sheet was puting out some too high numbers for the high regen scenario, but I modifed the calculation to cap the number of heals if it exceeded what could be cast in the allowable time. Total Heals = IF(Duration * Duty_Cycle / Heal_Speed < Total_Mana / Heal_Cost, Duration * Duty_Cycle / Heal_Speed, Total_Mana / Heal_Cost))

Now, of course, getting stats in that ratio might be a bit silly, but it's mostly an example... the difference between the healing pool in the examples is actually quite small--yet one cannot argue that the "healing experience" would be nearly so close as the healing pool might indicate. After all, at the extreme ends, the GH2 being cast by the guy with 3000 healing will heal for 86% more HP than the guy with 1000 healing. (Meaning, of course, he could potentially downrank to save mana if he wanted to match the same HPS as the other guy. You have to consider downranking at some point as a mana-saving device.) So, the practical results can probably vary wildly even if the end healing pool is similar.

Just something to consider...

Some things (clearcasting, some trinkets, some set bonuses) would be easier with a simulation. Others would be quite a bit harder. *shrug* Even if innacurate, however, it'd be inaccurate in different ways than existing models, and thus interesting.

Also, your point about the "healing experience" is a good one. With a given set of assumptions (20% GH7, 80% GH2, 6 minute fight, 100% inside FSR, normal-ish talents and stats), someone with 700 mp5 and 381 +healing will be able to do 300,030 HP of healing in 6 minutes. With the same assumptions but 50 mp5 and 4086 +healing they will be able to do 300,001 HP of healing in 6 minutes.

Now, just imagine trying to heal a boss fight with those two sets of stats - obviously the experience would be radically different. In the former case you'd be chain casting like crazy, and even though you'd do 300k of healing, it'd be spread out very evenly, and might not be enough to handle spikes. In the latter case, however, you'd be casting far fewer heals, and even though you'd do 300k of healing, it'd be very spiky, and might well all be wasted on overheal.

So yes, with extremely skewed gear you will be unable to effectively heal, and your potential healing output will be wasted. Note, however, that the calculations encourage you to pick an even mix stats. For example, here are the weights it produces with the above two scenarios:

What that's telling you is that in both of these cases, your gear is so skewed to one extreme that you're crippling your ability to heal. That is, I think, quite correct. Still...what is the proper metric, if any? Trying to maximize burst HPS is useless as a metric for evaluating gear. Potential healing output breaks at extremes - so can we just say "maximise potential healing output while maintaining burst HPS potential of X"? Or what?

Ultimately the issue is that healing tends to be pass/fail - either you kept your assigned target up or you didn't. It's not at all like DPS where we can compare two rogues based on their damage output; any metric we choose almost has to be pretty artificial.


Alright, so what about instead of using a weighted average heal, we construct a "heal cycle". For example, something like (picking stuff completely randomly):

GH2 -> GH2 -> 2s pause -> GH2 -> GH 7 -> Renew -> 5s pause -> Flash -> GH 2 -> GH2 -> 4s pause

That should work out to about 30s counting lag, and if you picked a good cycle, it would closely reflect how priests actually heal - spikes and all. So, if we assume this cycle is repeated for the entire fight duration, then we can work out the mana needed to sustain it, and calculate how much it would benefit from +healing, and so forth. Then if you wanted to be really complex you could pick a target "burst HPS" and calculate the amount of +healing needed to obtain that figure.

The problem with this though is that it doesn't let us construct weights, which was kind of the entire point of this discussion! What we want to know is "how many +healing must I gain to make trading away 1 mp5 worth it", but if we keep the healing cycle static we can't answer this - to actually answer the questions you'd have to calculate optimal healing cycles dynamically, which would be....hard.

Blah. That doesn't seem like a better approach.

Ah, the joys of theorycrafting.

Anyhow, I think your model is generally fine other than the fact that it ignores the potential mana benifit of downranking as an additional bonus to +healing. Due to downranking always being an option, +healing not only raises your healing pool but can also have the option of increasing mana efficiency by being able to use a more HPM-efficient lower rank while maintaining the same or similar HPS. (This is due to the fact that HPM is not consistant across all ranks, and that GH1 is significantly more efficient than GH7.)

For instance, in your example of the very far extremes, I'll use GH7 MT healing mode instead. 700 mp5/381 healing results in 92 GH7s, which would be 3051 per heal. The guy with 50 mp5/4086 healing would result in 30 GH7s, averaging 7359 per heal.

However, the guy with 4086 healing has another option. Intsead of using GH7, he could use GH2 for instance... even GH2 would have an average value of 5118 per heal, netting a significant increase in healing pool, allowing 54 casts.

This brings the healing pool up to 277902 instead of 220044. Given that the high regen version only has a healing pool of 281381, could one argue that having only 1% higher healing pool is worth having 60% lower HPS? Given that with high +healing, GH1 is slightly higher efficiency than GH2, downranking to that would probably provide a higher healing pool than the regen scenario, even though when comparing identical heal usage the regen scenario has a higher healing pool.

So, without considering the benifits of downranking as a positive to +healing, I don't think healing pool alone is totally fair on its own. It's missing a lot of the potential value of +healing in high enough quantites to make shifting down a rank to a higher HPM spell viable. (This is what my model is based around, which is probably why we get such different results--given that we are really looking at two very different aspects of mana conversation.)

Upranking is another option available, but as upranking generally decreases HPM it is rarely a positive thing. I think it's probably best to look from a basis of GH7 then look at the benifits of downranking from that point, as it is fairly straightforward to do so. Of course, mp5 has a higher impact on the sustainabilty of GH2 due to its lower relative mana cost, but without the +healing to back it up compared to the relative healing value of GH7 it is not really worth looking at. GH2ing for 1.5k isn't really a practical endeavor--think at that point you're looking at being replaced by a FoL-spamming Paladin rather quickly.

One potential way of looking at it would be to add to the value of healing the percentage of mana one could gain based on the percentage gap it brings between your current GH7 and GH1 (or GH2). While it's not perfect--as adding +10 healing wouldn't specifically mean you can now downrank--it would be perhaps a bit more fair.

For example, nice round numbers to make this easy...2000 healing, 200 mp5.

GH7 = 4934, 701 mana
GH2 = 3142, 386 mana
Difference between GH7 and GH2 is 1792 healing, 315 mana

+10 heal would make this:

GH7 = 4946, 701 mana
GH2 = 3152, 386 mana
Difference between old GH7 and new GH2 is 1782 healing, 315 mana

In essence, you have gained 0.56% of the difference in power, with the same change in mana reduction, or a gain of 1.76 mana per cast.

As you have the option of downranking at any time, you can always take advantage of efficiency gains with higher +healing while also being able to use the higher top-end potential of a max rank spell, however simply stacking additional mp5 only increases the longevity of existing spells and does not increase one's HPM via downranking at all. Therefore, +healing is a bit more complex of a factor.

This would not be as much of an issue if lower rank GH spells did not have notably higher HPM than GH7... however, that is sadly the case. If it was only a matter of picking the right size for the missing HP, downranking would still have its place...but the impact would be less dramatic from a mana-conservation point of view. Realistically, though, GH1/GH2 are significantly more efficient than GH7. (Looking like 7.05 HPM vs. 8.17 HPM in the above example, for instance...which is a good 15% difference.)

Hopefully I'm making some manner of sense here, although it's always possible I'm insane.