[Resto] Cataclysm Raiding and Discussion - Page 3

alizat · 12/14/10, 9:52 AM

From personal experience in using Tellturic Currents, 1/2 gives a negative mana return by about 300 give or take a few hundred.

But 2/2 has been doing me just fine. Just going off base mana, LB for me is 1405 mana. I haven't been looking at the straight damage of LB, but I've been watching my mana return, and I'm getting anywhere as low as 1780 to as high as 3008 mana returned. So we'll say a positive mana gain of anywhere between 303 and 1603.

Granted though, I haven't gone into a raid yet, But for heroics it's came in handy PLENTY of times as a last minute ditch to get some mana back and heal the tank to kill a boss.

As for Mastery, I haven't seen the numbers, and I haven't really been able to find a time when mastery saved me, but that's possibly only because it's one of those things that can be easily mistaken for something else causing the bigger heal or save of sorts. I -feel- like I'd want mastery higher, I've reforged my gear into it atleast, but as everyone else is feeling the pull of the casts, I'll probably do what this says and switch over to haste and gem up some different gems.

exschwizer · 12/14/10, 11:53 AM

Originally Posted by Cyfir

Bottom line: including Improved Water Shield, HS is almost exactly the same HPM as GHW, and much, much better HPS. Factor in the added flexibility HS adds to your healing, and there's no reason GHW should ever be on your bars, except inasmuch as it is part of a NS macro. HW is *much* lower HPS than either heal, but *much* higher HPM. With high enough crit values, HW approaches mana-neutral.

GHW can fill the role of heavy tank damage phases when mana restraints will kick in. In HPS terms we have HS>GHW>HW and in HPM terms HW>GHW>HS. GHW is the middle ground. And there is also no reason to heal the target of a random single target boss effect with HS, if it is not in intermediate danger of dying. Also, GHW can be precasted to end with the impact of casted special boss attacks. To use all the healing tools we have in the right circumstances and keeping an eye out for these opportunities will seperate the excellent Resto-Shaman from acceptable, in my eyes. But we will see after the first HM kills, if such skill will be rewarded.
And could you please elaborate the calculations you did in your spreadsheet concerning Mastery, preferably in the sheet itself. I have a hard time following your train of thought there. I appreciate the effort!

Cyfir · 12/15/10, 1:17 PM

Originally Posted by Corunix

I believe that your spreadsheet contains a computational error or you do not take into account the fact that 1 mastery point equals 2,5% bonus.
13% crit, 20% mastery (baseline), 6500 non-crit HW on 100% target.
Let's assume 90% bonus for crit (50% from crit and 27% from AA) and try to add 5% crit or mastery at the expense of the same amount of rating (1% crit costs as much rating as 1 point of mastery does).

Hypothetical 1% health target:
crit-powered heal = 6500 * (1 + 0,18 * 0,9) * (1 + 0,2 * 0,99) = 9048
mastery-powered heal = 6500 * (1 + 0,13 * 0,9) * (1 + 0,325 * 0,99) = 9451

So, if your targets are lower than ~60% HP, mastery is better than crit for direct heals in terms of HPS with zero overhealing (up to 5% bonus for 5%-equivalent rating spent).
Of course, the regen and utility of IWS and AF should be taken into account, along with the fact that indirect healing (hots) now contribute much larger amount of our healing, being affected by crit but unaffected by mastery.

I did have a computational error, incidentally however it was actually over-valuing mastery. It's now updated. To get the average value of a heal, factoring in crit and mastery, it would look like this:

critHeal = baseHeal * 1.5
AAHeal = critHeal * .286
masteryEffect = (baseMastery+gearMastery*.025)*(1-targetHealthPercent)
critValue = (1-critChance)*baseHeal + critChance(critHeal+AAHeal)
totalHeal = critValue*(1+masteryEffect)

Plugging everything in, we get:

totalHeal = ((1-critChance)*baseHeal + critChance(baseHeal*1.5 + baseHeal*1.5*1.286))*(1+(.2+gearMastery*.025)*(1-targetHealthPercent))

Simplified:

totalHeal = baseHeal (1 + .929*critChance)*(1+(.2+gearMastery*.025)*(1-targetHealthPercent))

This is what you've put together, and you're correct. Your example leaves out two things, however:

First, as the value of Mastery increases, it increases the effect on your critical heals as well. The differential between adding crit and adding mastery will thus be affected by diminishing returns. In fact, at exceptionally high levels of mastery and low levels of crit, mastery actually becomes less effective at throughput. The second sheet I've added to the spreadsheet will show this, but you're better off manipulating that data on your own computer as an excel - it runs very slowly.

Secondly, please note that comparing a 5%/5pt increase against a 1%/1pt increase involves issues of scale, and will therefore give you different results.

Originally Posted by alcapawn

Isn't this discussion analog to the tank mitigation/survival gearing? Crit might be a more valuable stat over a longer fight given expected crit rates. But in the worst case scenario, where someone has taken a large chunk of damage and is in need of a big heal, you can always count on the boost from the mastery, but even with extremely high crit chance - a bad streak can leave you completely un-boosted by your crit.

This would leave mastery as a more predictable way to boost your healing in an environment where you know you will be healing target with low HP - progression raiding. Or have I misunderstood the Deep Healing mechanics? Is there a random element I'm not aware of?

It is similar, but to make it directly equivalent, you would need a tank who gave the healers mana when he dodged/parried/blocked (Imp. Water Shield) and procced 10% damage reduction (AH). Mastery may have a more reliable effect, but the additional effects from talents are what makes crit the more desirable stat. Without those effects, mastery would be the better choice in almost any situation that mattered.

Originally Posted by exschwizer

GHW can fill the role of heavy tank damage phases when mana restraints will kick in. In HPS terms we have HS>GHW>HW and in HPM terms HW>GHW>HS. GHW is the middle ground.

As the linked speadsheet shows, GHW is only better HPM than HS if you're not casting it under Tidal Waves. Since we can assume you will be using TW under most, if not all, circumstances, we can say that HS is better HPM and HPS than GHW in virtually all situations. This is because the extra 30% crit for HS significantly increases its mana efficiency from Imp. Water Shield procs.

Spoonicus · 12/15/10, 6:34 PM

After switching to HW over GHW i find that while I dont OoM as often, I im not quite healing to the level I should. At what point of SpellPower/Haster/Crit does HW become a reasonable to use spell during periods of medium to heavy damage?
Also I've found that Spiritwalkers Grace is actually rather good for resto shamans. When it says "cast while moving" it quite literally means that, and you can continue casting while in freefall, trapped by cyclones, or during periods of knockback.

Corunix · 12/15/10, 6:36 PM

Originally Posted by Cyfir

I did have a computational error, incidentally however it was actually over-valuing mastery. It's now updated. Your order of operations is a bit off. To get the average value of a heal, factoring in crit and mastery, it would look like this:
...

totalHeal = (baseHeal + .929*critChance*baseHeal)*(1+(.2+gearMastery*.025)*(1-targetHealthPercent))

I'm sorry, but I fail to see how this is any different from my formula (especially in terms of order of operations), except for the 0.929 instead of 0.9 crit scale:

Originally Posted by Corunix

Hypothetical 1% health target:
crit-powered heal = 6500 * (1 + 0,18 * 0,9) * (1 + 0,2 * 0,99) = 9048
mastery-powered heal = 6500 * (1 + 0,13 * 0,9) * (1 + 0,325 * 0,99) = 9451

Also, your spreadsheet demostrates that mastery is better than crit when target is below ~60%.
I agree that crit is a proper competitor to mastery, especially on the regen and "smart" healing AA nature, yet I think that mastery has it's value as a baseline boost to healing which cannot be cast aside easily (one can't rely on crits when at around 20-25% raidbuffed, also the "smart" portion of healing might get wasted when there's some sort of decimate or healing absorb ability on the encounter).

Cyfir · 12/15/10, 7:01 PM

Originally Posted by Corunix

I'm sorry, but I fail to see how this is any different from my formula (especially in terms of order of operations), except for the 0.929 instead of 0.9 crit scale:

Also, your spreadsheet demostrates that mastery is better than crit when target is below ~60%.
I agree that crit is a proper competitor to mastery, especially on the regen and "smart" healing AA nature, yet I think that mastery has it's value as a baseline boost to healing which cannot be cast aside easily (one can't rely on crits when at around 20-25% raidbuffed, also the "smart" portion of healing might get wasted when there's some sort of decimate or healing absorb ability on the encounter).

You are correct about our formulae; I have updated my post to not call you out incorrectly. However, you appear to ignore the other two issues (Mastery suffers from DR; %5/5 does not equal 1%/1.) Again, there's no question that Mastery in most circumstances is better throughput. What makes crit the better overall stat is 1) 1 CR is about equal to 1/3 SPI in terms of mana regen, and 2) Mastery does not proc AH.

EDIT: The bottom line is this - crit and mastery are not directly analogous stats. Mastery provides only throughput; crit provides slightly less throughput in exchange for efficiency and AH procs. If we're going to compare for min/maxing your gear, and you accept 1 CR as ~ 1/3 SPI for mana regen, then you would need enough additional throughput from mastery such that 3 CR < 2 MR + 1 SPI, and that doesn't even consider AH procs. MR is nowhere near that.

ellotheth · 12/15/10, 7:36 PM

Originally Posted by Keiyakusha

Been keeping this to myself for a while, as I wanted to be sure, but I think now is a good time to share. The UL bonus is affecting Riptide's HoT and will linger long enough for you to cast a heal into a queued Riptide and gain that 30% boost to the heal, Riptide's DH, and Riptide's HoT. I generally have UL on constant cooldown as a result of this. This works 100% of the time (it's dependable).

Same here, obligatory combat log snip:

"Lotiel","0x511",61295,"Riptide","0x8",5066,5066,0,"nil"
"Lotiel","0x511",61295,"Riptide","0x8","BUFF",,,
"Lotiel","0x511",61295,"Riptide","0x8",,,,
"Lotiel","0x511",61295,"Riptide","0x8",1597,1597,0,"nil"
"Lotiel","0x511",61295,"Riptide","0x8",1597,1597,0,"nil"
"Lotiel","0x511",61295,"Riptide","0x8",2395,2395,0,1
"Lotiel","0x511",61295,"Riptide","0x8",1597,1597,0,"nil"
"Lotiel","0x511",61295,"Riptide","0x8",1597,1597,0,"nil"
"Lotiel","0x511",61295,"Riptide","0x8",1596,1596,0,"nil"
"Lotiel","0x511",61295,"Riptide","0x8",1597,1597,0,"nil"
"Lotiel","0x511",61295,"Riptide","0x8",2395,2395,0,1
"Lotiel","0x511",61295,"Riptide","0x8","BUFF",,,
"Lotiel","0x511",73685,"Unleash Life","0x8",4322,4322,0,"nil"
"Lotiel","0x511",73685,"Unleash Life","0x8","BUFF",,,
"Lotiel","0x511",73680,"Unleash Elements","0x1",,,,
"Lotiel","0x511",61295,"Riptide","0x8",6448,6448,0,"nil"
"Lotiel","0x511",61295,"Riptide","0x8","BUFF",,,
"Lotiel","0x511",61295,"Riptide","0x8",,,,
"Lotiel","0x511",73685,"Unleash Life","0x8","BUFF",,,
"Lotiel","0x511",61295,"Riptide","0x8",1887,1887,0,"nil"
"Lotiel","0x511",61295,"Riptide","0x8",1887,1887,0,"nil"
"Lotiel","0x511",61295,"Riptide","0x8",1887,1887,0,"nil"
"Lotiel","0x511",61295,"Riptide","0x8",1887,1887,0,"nil"
"Lotiel","0x511",61295,"Riptide","0x8",1887,1887,0,"nil"
"Lotiel","0x511",61295,"Riptide","0x8",1887,1887,0,"nil"
"Lotiel","0x511",61295,"Riptide","0x8",1887,1887,0,"nil"
"Lotiel","0x511",61295,"Riptide","0x8",2830,2830,0,1
"Lotiel","0x511",61295,"Riptide","0x8","BUFF",,,

That's tasty.

Draewind · 12/16/10, 9:04 AM

I am about to do my first raid tonight. I have found, that in 5man heroics, Telluric Currents is almost a necessity, but I am going to drop it for raiding. I am oom just as the boss dies in good groups, so I am going to try the routine listed by prim0c. Chain Heal appears to be useless. I don't even have it on a top priority bind anymore. It is a ctrl-left or right ... I forget. Greater Healing Wave is all I have to really heal. I like Healing Rain, and will get much more use of it in raids. Healing surge is quick, but not very effective and I am oom as we are wiping because of it.

The Spi, Mastery, and Crit are difficult for me to determine. Haste seems to be a pie in the sky at this gear level.

What has the math told us about HEP? I am using:
Haste 2.5
Int 2.4
Crit 1.6
Spi 1.3
SP 1
Matery 0.6

EDIT:
For now I have decided not to reforge Spirit. I will reforge Mastery to Haste or Crit if Haste is already on the item. I put [Brilliant Inferno Ruby] in red sockets, [Quick Amberjewel] in yellow sockets, and [Purified Demonseye] in blue sockets (Meta = [Ember Shadowspirit Diamond]). Once I get into better gear and know more about the environmental demands, I will better be able to judge.

I'll look at it again in the raid tonight, but my main question is this:

Is Chain Heal dead?

EDIT:
After running a raid, I would say not. Chain Heal still has a place in raid healing. For the heroics, I found that the routine given by prim0c has saved the day. I no longer spec into Telluric Currents. Many thanks.

Levva · 12/16/10, 10:23 AM

Originally Posted by ellotheth

Same here, obligatory combat log snip:

"Lotiel","0x511",73685,"Unleash Life","0x8",4322,4322,0,"nil"

What was your spellpower with that? The heal was 4322, the base is average 1995 so it looks like you are getting 2327 heal which assuming you have around 5.5k SP gives co-efficient in the range of 42% are you able to do more testing with different SP and get some more logs to narrow down the SP vs heal range so we can establish the co-efficent.

I did some numbers on Telluric Currents which suggests that due to lack of hit it is NOT reliable and on average will cost you mana. Edit: Note that due to the far lower miss rate you WILL gain mana in heroics but this calc is for Raiding as Heroics aren't really important.

LB cost 6% base mana. Base mana at lvl 85 is 23430. Therefore LB costs 1406 mana. With Teluric Currents the return is 20% of the damage dealt per point I'll assume 2 pts spent.

Base av LB damage is 770 and the co-efficient is 0.713999987 * sp. Miss rate assuming no +hit is 17% (hit rate is therefore 83%).

Total damage done is therefore hitrate * ((1 - crit) * (770 + 0.713999987 * sp) + crit * 1.5 * (770 + 0.713999987 * sp))

The *1.5 is because spells do 50% more damage when crit. Now lets assume 20% crit and 5k spellpower obviously you can fill in your own numbers.

Total damage is 0.83 * ((1-.2) * (770 + 0.713999987 * 5000) + 0.2 * 1.5 * (770 + 0.713999987 * 5000)) = 0.83 * ((0.8 * 3570) + (0.3 * 3570)) = 3141.6 damage. Now if you are only getting 40% back you get back 40% of 3141.6 which is 1256.64 mana returned vs 1406 mana spent a net COST of 149 mana.

Now of course as you get more crit and more SP these figures will change. However the average is that due to the 17% miss rate TC is on average costing you 149 mana. At better gear or with a very small amount of hit (pointless for Resto though) you will have a net gain but at current gear levels its a net cost.

Since these numbers fly in the face of what was suggested by others I can't help but think I've missed something out. Is there something that lowers LB cost? You don't take convection and I can't see an obvious damage multiplier I'm missing.

On UE I did some numbers on MMO-Champion thread but forgot about the intial heal anyway in case it helps the calcs were...

Unleash Elements costs 7% base mana and boosts heal by 30% - 20% base * 50% boost from Elemental Weapons.

Its usual use as far as I can see is to boost GHW or CH. For a GHW the cost is 30% base mana. Since base mana at 85 is 23430 mana this is a 7029 mana cost. UE is a 1640 cost. This then means you can spend 1640/7029 = 23% more mana for a 30% healing boost if you prefix a GHW with a UE.

So UE+GHW = 8669 mana for 1.3 * GHW vs 14058 mana for 2 * GHW. However UE+GHW = 4.5 sec cast (reduced by haste) vs GHW+GHW 6 sec cast (reduced by haste). ie: UE is only boosting your heal by 5.4% over what casting two GHW back to back cost.

However since UE uses a GCD the difference is simply cast time. UE+GHW = 1.5sec+3sec hasted down to your level of haste. So it takes 4.5/6 = 75% of the time and is 5.4% better. So overall the boost seems to be 40.54% if you cast UE+GHW vs GHW+GHW

For Chain Heal its less as mana cost is 20% so the costs are 4686 mana vs 6326 ie: 35% more mana for 30% healing boost. The casts are UE+CH = 3.5 sec (reduced by haste) vs 4 sec (reduced by haste).

So the costs are 6326 mana for UE+CH vs 9372 for CH+CH for a 3.5sec vs 4 sec cast thus by the same calcs above we get only a 10.05% boost by using UE+CH. Still that's decent.

All this assumes 2 points in Elemental Weapons.

Pewpewfurie · 12/16/10, 5:15 PM

Originally Posted by Draewind

I am about to do my first raid tonight. I have found, that in 5man heroics, Telluric Currents is almost a necessity, but i am going to drop it for raiding. I am oom just as the boss dies in good groups, so I am going to try the routine listed by prim0c (below). Chain Heal appears to be useless. I don't even have it on a top priority bind anymore. It is a ctrl-left or right ... I forget. Greater Healing Wave is all I have to really heal, and since I have over used it, I will try to use the routine by prim0c below. I like Healing Rain, and will get much more use of it in raids. Healing surge is quick, but not very effective and I am oom as we are wiping because of it. All I can think of doing at this point is piling on the haste, but there doesn't seem to be room with current level of gear. I don't see spirit as priority, even though it is our only regen tool, which won't help if we're dead because I was too slow. I will not reforge out of it, but I will use tools like food buffs and flasks and trinkets to boost it.

The Spi, Mastery, and Crit are difficult for me to determine. Haste seems to be a pie in the sky at this gear level.

What has the math told us about HEP? I am using:
Haste 2.5
Int 2.4
Crit 1.6
SP 1
Spi 1.3
Matery 0.6

I'll look at it again in the raid tonight, but my main question is this:

Is Chain Heal dead?

My guild is currently 5/12(10man) 4/12(25man) and we are pushing for 7/12(10man) tonight. The only fight I have used Chain Heal on so far is Magmaw, and I have used it heavily.

However, I do not think that you should drop TC at all, for me it has been an integral part of almost every raid encounter thus far.

Omnitron Defense System - I use it during Arcanotron's pool, deals extra damage while regening mana.
Mag'maw - Exposed head phase, 10%-100% mana every phase.
Conclave of Winds - Both on the Anshal and Rohash platform, you have windows of 10+ seconds where you can LB spam.
Halfus Wyrmrest - 150% damage bonus, a lot of time to regen mana prior to Furious Roar.
Double Dragon - Used much less here, but I have used this occationally during HR CD when raid damage isn't high.
Argaloth - UE+HR every Meteor Slash, LB spam in between.

Having TC allowed me to be very liberal with my mana and I have to say so far from my raiding experience in Cataclysm, it's the best 2 point investment in the resto tree.

Originally posted by Levva.
I did some numbers on Telluric Currents which suggests that due to lack of hit it is NOT reliable and on average will cost you mana. Edit: Note that due to the far lower miss rate you WILL gain mana in heroics but this calc is for Raiding as Heroics aren't really important.

Having around 6400 SP in raid, my LB is averaging 5500 a hit, that's returning 800mana assuming the LB hits. Even discounting encounters where you have damage bonus, I have always experienced mana gain spamming LB. For example on Argaloth, if you only use UE+HR for Meteor Slash and exclusively LB in between the CDs, you can almost stay full mana.

Hamano · 12/17/10, 11:59 AM

Originally Posted by Pewpewfurie

Having TC allowed me to be very liberal with my mana and I have to say so far from my raiding experience in Cataclysm, it's the best 2 point investment in the resto tree.
.

Isn't the reason that you can be very liberal with your mana also that (it seems like) you are actually spending more time dps'ing than healing. It's not even close to the best two points either, because you have to give up some other really good talents for it (Either at least 2% crit, or the 10% damage reduction)

This is purely theoretical, but I also highly doubt there will be alot of time to spent on dps'ing during the heroic versions of the bosses

dgkiller · 12/17/10, 2:09 PM

Originally Posted by Pewpewfurie

if you only use UE+HR for Meteor Slash and exclusively LB in between the CDs, you can almost stay full mana.

UE currently does not increase HR value, if you use HR after UE, it will not consume the UE effect waiting for the next healing spell, probably intended.

initium · 12/17/10, 3:25 PM

Originally Posted by Pewpewfurie

Having around 6400 SP in raid, my LB is averaging 5500 a hit, that's returning 800mana assuming the LB hits. Even discounting encounters where you have damage bonus, I have always experienced mana gain spamming LB. For example on Argaloth, if you only use UE+HR for Meteor Slash and exclusively LB in between the CDs, you can almost stay full mana.

I did some math using these numbers, and here's what it shows. Please correct any math on this, it's my first time doing anything like this for wow.

800 mana per hit would work out to 1900 mana per crit. (5500 * 1.5 = 8250 * .4 = 3300 - 1400 cost of LB)

Assuming a 17% miss rate on raid bosses (I don't think anyone is wearing hit gear), and a 17.11% crit chance (based on my armory) I come up with a net gain of 820.2143 mana for 2/2 TC

here was my formula in excel:

(800 * .83) + (1100 * .1711 * .83 )

Edit: had wrong math on mana from crits
This look right to everyone?

Seeing as we won't be adding hit as gear improves, but we will be adding spell power and crit, i figured i would graph out how TC would scale.

here is what it looks like as crit chance improves, very linear. Every point of crit (at this amount of spell power) is worth 9 mana per LB, more as spell power increases

Sardaukar · 12/17/10, 4:24 PM

I think a lot of people are looking at regen stats vs haste and thinking about it in an outdated WotLK way. Haste increases your HPS a lot, just like in WotLK, but spirit and mana regen focused gear doesn't just increase your regen at the expense of throughput - it also increases your throughput greatly. Why is that? Because unlike in WotLK when we always used our highest HPS heals, regardless of mana, having high regen now lets a healer more freely use their more expensive, faster heals like surge and greater healing wave. Spirit doesn't increase your throughput in any easily measurable way, but in an intangible 'Its ok to chain a few healing surges when I need high HPS because I can regen' kind of way.

As far as TC goes, I've done 5 bosses so far and I just have not really felt comfortable nuking all that much other than the Magmaw fight.

mersenne · 12/17/10, 6:12 PM

Originally Posted by exschwizer

GHW can fill the role of heavy tank damage phases when mana restraints will kick in. In HPS terms we have HS>GHW>HW and in HPM terms HW>GHW>HS. GHW is the middle ground. And there is also no reason to heal the target of a random single target boss effect with HS, if it is not in intermediate danger of dying. Also, GHW can be precasted to end with the impact of casted special boss attacks. To use all the healing tools we have in the right circumstances and keeping an eye out for these opportunities will seperate the excellent Resto-Shaman from acceptable, in my eyes. But we will see after the first HM kills, if such skill will be rewarded.

Calculations I've done indicate that HS actually has better HPM than GHW when you have TW. This is because of the increased HS crit chance from TW granting more mana from IWS, and is the case despite the lower IWS proc chance from HS. Of course, HS also has better HPS than GHW with this crit chance increase, due to AA, and also increases AH uptime (albeit perhaps by a negligible amount in a raid). HS is also just plain faster, which means more time for Lightning Bolts, repositioning, etc. So unless I've erred in my calculations, it may be time to move GHW to an out-of-the-way keybind.

The base healing amounts I used in my calculations are estimations of the average - I wasn't able to find the actual values. Corrected values would be great but these should be close enough to tell us which spell is better.

Assume Tidal Waves, with 30% additional chance to crit Healing Surge.
SP = your spellpower
CR = your crit chance
HST = your haste

HS_basehealing = 7697
HS_coeff = 0.604
HS_casttime = 1.5 / (HST + 1) sec

HS_truehealing = (1-CR-.3)*(SP*.604+7697) + (CR+.3)*(SP*.604+7697)*1.8
                     ^ non-crit heals          ^ crit heals w/ AA

HS_basecost = 5946 mana
HS_truecost = (1-CR-.3)*5946 + (CR+.3)*[ (.6*4201) + (.4*5946) ]
               ^non-crits        ^crits w/ Improved Water Shield's 60% proc


GHW_basehealing = 10255
GHW_coeff = .806
GHW_casttime = 2.5 / 1.3 / (HST + 1) = 1.923 / (HST + 1)
                      ^ Tidal Waves' 30% cast time reduction for GHW
GHW_truehealing = (1-CR)*(SP*.806+10255) + CR*(SP*.806+10255)*1.8
                     ^ non-crit heals        ^ crit heals w/ AA

GHW_basecost = 6607
GHW_truecost = (1-CR)*6607 + CR*4862
                ^non-crits    ^crits w/ Improved Water Shield's 100% proc

I've posted a very rudimentary spreadsheet with these formulas plus HPS formulas on Google Docs, linked below:
https://spreadsheets.google.com/ccc?...Xc&hl=en#gid=0
The inputs given are my current self-buffed stats. (My Shaman is just barely in heroics...)
Playing around with the inputs reveals that HS will always have better HPM and HPS than GHW with all feasible values for SP and CR.
And in practice, it really does feel better on my mana to never use GHW under TW. (Choosing a heal when you don't have TW is another story...) I heard through the grapevine that top-level raiding Shaman are forgoing GHW altogether - while I haven't done any research into this trend, I wouldn't be surprised if they came to the same conclusion I did.

Of course, if I screwed this up somewhere I'd be more than happy to edit/redact as necessary. Thanks for reading.