1. I agree with the CS analysis. I personally don't see how you can justify using CS unless you're already at full mana (especially since iirc the CS CD isn't even hasted - thus costing more than 1 GCD in relative time).

2. I have noticed that I haven't gotten back to back Eternal Glory procs since I've specced back into it (for a couple of weeks after Cata I avoided it due to the LoD being ridiculous). Haven't analyzed what the ICD might actually be, but 15 seconds is a believable number.

3. The main reason why 4 pc T11 is controversial is the lack of haste, but looking at Evarella's analysis of spirit vs. intellect pretty much demonstrates that spirit is by far the best secondary stat (and probably close to 1.5-2x better than any other secondary). If you can generate ~100 average spirit per each piece of T11, you could easily justify at least 4pc; this represents about 25% uptime on Holy Radiance, or once every 40 seconds. At that level of usage, 4pc t11 seems easily better (especially since you can reforge a portion of the extra secondary into haste anyway) and probably justifiable even at lower usage of Holy Radiance.

4. I personally use WoG over LoD because I can't always hit 6 people with it. You suggested the same reason later, so this question seems odd. Using your numbers, WoG is better or equivalent if you have 3 targets or less (~4 if one of those is the Beacon). On any fight that requires range to spread out, I can typically only guarantee hitting the melee, myself and maybe the tank, which is ~3-4 people for me in 10 man, and requires all of those targets to be hurt (and requires me to be within 30 yards). If you factor in 30% from Eternal Glory, LoD is only really a win with 5+ targets. There are also other situations where hitting the most hurt individual is more important than hitting a lot of people. Also, without the LoD glyph, LoD is only very slightly better when considering EG; if I accidentally miss someone or have to move to hit the 5th, WoG is generally better. Given the point about the EG ICD, however, it might generally be better to get in the habit of WoG->LoD on proc if LoD has a decent, not optimal number of targets (especially since you can use the WoG GCD to move slightly if needed).

I am very interested in figuring out if using Crusader strike is worth it. I initially didnt do it when HL could generate HP on the beacon target. After the hotfix I started using CS to make up for the loss. And having fairly good results. So trying to transcend my normativity:

- What is the exact HPS and HPM implications of Crusader strike healing behavior? That somehow needs to be detemined in a resonable manner.

- The rational intent of me using CS has been something like: Generating holy power enables increased use of LoD. LoD is the spell that puts out the most healing in one global. So increasing its use can only be good on average (then how much better/worse is it?). Is this detrimental in any way? Generating on demand holy power through other sources than beacon/holy shock can mean a lot of things. Putting out several LoDs within a very short timeframe. Having to resort less to ToR healing (that has a tendency to result in overhealing).

- LoD in itself is a smart heal and chooses targets better than I do (assuming a situation where the six lowest targets are actually the best to heal).

- Trying to figure out more good reason to use CS has actaully put me at a loss. It seems to come down to the HPS/HPM situation using CS instead of HL/DL.

Other thoughts?

EDIT: On another note I almost never encounter a situation where I use WoG over LoD. The only time is when healing is needed on a single target and it only makes sense to WoG. If healing is needed on several targets I almost always have 6 targets. So the in between scenario rarely exists for me. This is from a hardmode 25man perspective.

Am I missing something, or is the formula for Spi regen off by about 200%? What it says in the main post is:

My Int is 4047. The value I get when I plug my Int into the equation is 0.534330059.

Checking my acutal charsheet, by adding and removing a trinket with 215 Spi, I can see I'm getting 106.511628 out of combat mp5 per 100 Int, or 53.4883721 in combat mp5. The reason I mention that last number is, well, it's almost exactly 100 times the figure I got from the formula.

Seems like it's off by a factor of 100. Maybe the OP means 1 spirit provides this formula's worth in mp5, not 100 spirit.

Based on the discussion around back-to-back EG procs (I also haven't seen any since respec'ing back to EG) I thought it would be interesting to see whether EG generated HP falls off in the same way normal HP does (i.e. 1 point at a time instead of all 3 at once). It does, and after refreshing just the last point I did get another EG proc.

It also seems that HP falls off at a rate slower than 1 every 15 secs. Would it be worthwhile casting regular heals for 15 seconds before firing off another WoG? Naturally delaying the use of HS so it is available to start generating HP again if we don't get another EG proc.

Are you 100% certain of this? Since the Cataclysm release the Holy power I've gained from Eternal Glory has NEVER procced again after using another word of glory. I'm reading several reports of people saying they have had 5 or 6 back to back procs.. I don't want to call them liars but I guess I am I guess I'm one of those. "I need to see it to believe it people."

The problem with that is the only thing we use Spirit for is to boost regen. According to Lightrender's figures, 4T11 returns around 1900 mana but it costs 9400 mana to trigger it. Spending 9% of our mana to get 2% back is not a regen boost. As such, Lightrender is correct in saying the true value of 4T11 lies in its effective cost reduction on Holy Radiance. This goes double in 4.0.6, when we will no longer be able to sync Holy Radiance with Mana Tide for massively boosted regen.

Re: Crusader Strike. When I was championing this as a way to gain HoPo, it wasn't meant to be as a primary source. It's simply the cheapest source of HoPo we have after Holy Shock. If we cannot afford to generate HoPo by casting Flash or DL on the Beacon - or we aren't specced for Tower - then it's the only reliable way we have to use a big heal for free more often than every 18 seconds. That's a mana saving from both ends; less expenditure to generate our HoPo, and more frequent use of 3-point Word/LoD.

Lightrender, your rotation figures are out - you've only included one of the Holy Lights in the total healing. Regardless, the better CS rotation with Beacon would be HS - DL to Beacon - CS - Word - Judgement. Plugging in your numbers and assuming DL is 22k and EJs is 10k, you get ~81k healing in 8 seconds at a cost of 11k mana, 3.5k of which is returned. Even including your second HL that's equivalent HPS and HPM, needs no glyphs where your rotation needs two, and mana return is guaranteed rather than proc-dependent.

Well, I'm certainly not advocating using more HR with 4T11 simply in an attempt to regain mana. But if the encounter benefits from that level of HR usage, 4T11 looks pretty attractive. Viewing 4T11 as a mana cost reduction on HR is a valid interpretation, but it's not the only or necessarily the most useful one. The primary issue at hand is whether or not the gear set up is better than the alternative; in order to do that, we need to be able to formulate some comparison with the alternatives. The method I am proposing is to calculate the equivalent value of the set bonus in terms of spirit on gear, for a given uptime on HR, without suggesting that it's necessary to change the usage of HR.

Again, my basic argument is:

1) If the alternate secondary stat on the set pieces were switched out for reforge-level haste and rest spirit, the pieces would be worthwhile without considering the set bonus;

2) The uptime required on HR in order to get the effect of 1) is not completely unreasonable (~400 spirit equivalent should be necessary, which seems to correspond to about 25% uptime on HR, or every 40 seconds);

3) You get the benefit of the actual stat left after reforging plus the spirit-equivalent of the set bonus plus the haste.

@Evarella: Are the values for Intellect and Spirit in your graph buffed or unbuffed?
Also, what has been said about the beacon could be said about PotI (it's getting nerfed though), am I right? Given it appears to have a very low spellpower coefficient...
Or did you already include that into your calculations?

Thanks to the feedback I received, I was able to update my calculations. I'll try to expose briefly what I changed and how I modeled IoL and Daybreak. This is still from a pure theoretical standpoint, focusing on single-target healing, and doesn't include spirit into the picture, nor the mana-related advantages of intellect. Though Evarella has already provided with an excellent contribution to the int vs spirit debate.

Given the amount of computations needed, I used a MATLAB script to automate the process and try different stat values. Give it a look if you're interested, even if you don't know the language, it's very intuitive and heavily documented. A free interpreter of MATLAB scripts is Octave.

My first step in updating my calculations was to include PotI and account better for beacon of light.
We already know the base heal for a 3/3 PotI is 4002 (pre-nerf), so I gathered a small sample of PotI heals using different spellpower levels and specs.
From this, I got a spellpower coefficient of 0.0564, and I verified PotI is affected by the 15% passive healing increase of holy spec. So even though I didn't verify it explicitly, I assumed all other healing modifiers apply to PotI as well.

This way, on average, and accounting for critical strikes and beacon, each direct healing spell not cast on self will heal for an additional:

This is obviously in the best possible case, with the caster not at full health, otherwise the beacon multiplier will go down, reaching 0.5 in case of no useful healing on the caster.

Including the PotI value calculated just before and Mastery, a direct heal will generate, in the best possible case:

(Assuming "Heal" is the average value of the healing spell, taking into account critical heals and the appropriate healing modifiers.)

That's basically the formula I used to evaluate the contribution of stats. As we'll see, the apparently very low coefficient of PotI puts intellect at a disadvantage when speaking about pure HPS.

I also updated the calculations for Word of Glory to account for the attack power scaling component.
First our buffed AP, as I'm not 100% sure it's correct:

And what currently appears to be the fundamental formula for WoG (@ 3 holy power) is:

To account for Daybreak and Infusion of Light I used a very simple theoretical model, and evaluated an upper bound for the contribution of stats increases to Daybreak proc chance, the average number of HL/DL that can be cast while HS is on cooldown, and the holy power generated per second. By estimating the effect of stat increases on those quantites,
I could account for the change in the throughput of HL, DL, HS, and WoG.

More in detail, for a given Haste%, the cast time of a HL/DL will be:

The model assumes the first spell cast is an HS, and the second is a WoG/LoD. At this stage, the first 1.5 seconds are used by HS and just after that 0.5 seconds are used (on average) by WoG/LoD (GCD * Holy Power / 3).

In case the HS was a crit, the next HL/DL will require 1.5 seconds less. So, on average, the next HL/DL cast will last:

(as IoL gives +10% crit to HS).

So there are still (6 - 1.5 - 0.5 - IoL_cast) seconds before HS goes off cooldown. Assuming nothing other than HL/DL are cast will give an upper bound to Daybreak proc chance, so the average number of HL/DL that can be cast on this time span are:

(it is not rounded or truncated)

This gives an initial Daybreak proc chance of (the +1 is the HL/DL accounting for the IoL proc):

The algorithm is then iterated to account for Daybreak procs, calculating the average "cast time" of HS and WoG using:

and calculating again the number of average HL/DL casts possible before HS goes off cooldown.
Only a few iterations are needed for the Daybreak proc chance and the cast time affected by IoL to stabilize.

The increase in daybreak proc chance is used to scale HS healing increases, while the change in holy power per second is used to scale WoG healing increases. The change in the average number of HL/DL cast in between HS is finally used to give a better estimate of overall haste increases to HL and DL.

While this model shows that haste gives an indirect increase to the output of HS and WoG/LoD, it seems to be pretty much insensitive to crit changes, suggesting that a lot of crit is needed for IoL to make a difference, or perhaps that my calculations are wrong!

I should point out I was somewhat tricked into reasoning that crit is a peculiar mana saver, but the truth is it isn't any more than intellect or mastery. Since crit saves mana by making spells heal more on average, the same can be said about intellect and mastery (albeit in different ways). It remains to see if haste can give comparable benefits in mana preservation...

On the other hand, better accounting for mastery gives a more realistic value of crit with respect to the ones I previously calculated.
The amount of mastery on gear needed to make crit overtake haste is probably very high though, and this is still assuming the value of haste isn't influenced at all by mastery, which is a bit unrealistic.

Ok, sorry if I bored you, here are some values I estimated for various stat values. The "naive" estimate for haste is calculated similarly to how I did in the other post and there are values for crit under different situations.

There are also weighted averages for more diverse arrangements. They are:

Single target, heavy damage:

FoL: 30%
HL: 20%
DL: 10%
HS: 20%
WoG: 20%

Single target, moderate damage:

FoL: 5%
HL: 20%
DL: 30%
HS: 20%
WoG: 25%

Multiple target, low WoG usage:

FoL: 20%
HL: 35%
DL: 20%
HS: 20%
WoG: 5%


Setup: Erthadan (Entry-level raider)

4201 intellect
1729 spellpower on items
674 crit rating
1623 haste rating
104 mastery rating
(for all setups, stats are as read on armory/character sheet)

Increase to heal values, in HPS:

Stat	+FoL	+HL	+DL	+HS	+WoG
+100 Intellect	229	73	181	76	179
+100 Haste rating (naive)	247	100	202	0	0
+100 Haste rating (including Daybreak)	209	84	171	32	49
+100 Crit rating	141	53	119	56	131
+100 Crit rating (factoring IoL)	141	53	119	56	131
+100 Mastery rating	99	33	88	34	95

Weighted averages:

Scenario	Intellect	Haste rating(naive)	Haste rating (+Daybreak)	Crit rating	Crit rating (+IoL)	Mastery rating
Single: Heavy	152	114	113	102	102	71
Single: Moderate	140	93	97	97	97	68
Multi	132	125	114	88	88	60

Setup: Full ivl 359 (focus on haste)

5045 intellect
1954 spellpower on items
558 crit rating
1904 haste rating
90 mastery rating

Increase to heal values, in HPS:

Stat	+FoL	+HL	+DL	+HS	+WoG
+100 Intellect	234	74	185	76	179
+100 Haste rating (naive)	264	105	215	0	0
+100 Haste rating (including Daybreak)	223	88	181	34	51
+100 Crit rating	153	57	129	59	139
+100 Crit rating (factoring IoL)	153	57	129	59	139
+100 Mastery rating	109	36	96	36	102

Weighted averages:

Scenario	Intellect	Haste rating(naive)	Haste rating (+Daybreak)	Crit rating	Crit rating (+IoL)	Mastery rating
Single: Heavy	155	122	120	110	110	77
Single: Moderate	142	99	103	104	104	74
Multi	134	133	121	95	95	66

Setup: Full ivl 359 (focus on mastery, not a real gear set)

5045 intellect
1954 spellpower on items
558 crit rating
904 haste rating
1090 mastery rating

Increase to heal values, in HPS:

Stat	+FoL	+HL	+DL	+HS	+WoG
+100 Intellect	225	71	178	79	185
+100 Haste rating (naive)	264	105	215	0	0
+100 Haste rating (including Daybreak)	225	89	183	35	53
+100 Crit rating	146	54	122	60	141
+100 Crit rating (factoring IoL)	146	54	122	60	141
+100 Mastery rating	102	34	90	36	102

Weighted averages:

Scenario	Intellect	Haste rating(naive)	Haste rating (+Daybreak)	Crit rating	Crit rating (+IoL)	Mastery rating
Single: Heavy	152	122	121	107	107	74
Single: Moderate	141	99	104	102	102	71
Multi	131	133	123	92	92	62

Setup: Full ivl 359 (focus on crit, not a real gear set)

5045 intellect
1954 spellpower on items
1558 crit rating
904 haste rating
90 mastery rating

Increase to heal values, in HPS:

Stat	+FoL	+HL	+DL	+HS	+WoG
+100 Intellect	224	71	176	78	184
+100 Haste rating (naive)	271	108	220	0	0
+100 Haste rating (including Daybreak)	232	92	189	36	55
+100 Crit rating	143	53	120	59	139
+100 Crit rating (factoring IoL)	143	53	120	59	139
+100 Mastery rating	104	34	92	37	104

Weighted averages:

Scenario	Intellect	Haste rating(naive)	Haste rating (+Daybreak)	Crit rating	Crit rating (+IoL)	Mastery rating
Single: Heavy	151	125	125	105	105	76
Single: Moderate	140	101	108	100	100	73
Multi	130	136	127	90	90	64

Setup: Diamondtear (High-end)

5462 intellect
2207 spellpower on items
482 crit rating
1796 haste rating
259 mastery rating

Increase to heal values, in HPS:

Stat	+FoL	+HL	+DL	+HS	+WoG
+100 Intellect	234	74	185	77	180
+100 Haste rating (naive)	274	108	223	0	0
+100 Haste rating (including Daybreak)	232	91	188	35	53
+100 Crit rating	159	59	133	61	144
+100 Crit rating (factoring IoL)	159	59	133	61	144
+100 Mastery rating	113	37	100	38	106

Weighted averages:

Scenario	Intellect	Haste rating(naive)	Haste rating (+Daybreak)	Crit rating	Crit rating (+IoL)	Mastery rating
Single: Heavy	155	126	124	114	114	80
Single: Moderate	142	102	107	108	108	77
Multi	134	137	126	98	98	68

EDIT: The haste estimate is incorrect, I'll try to fix that soon.

You should remember that it should be significantly easier to get off-set pieces than tier pieces, especially when it come to ilvl 372. I wouldn't take tier items (or pay a lot of VP, which can be turned into gold) for a set that might be better on some encounters.

Eternal Glory
The back to back procs were changed by blizzard shortly after release IIRC. I can go dig up the blue post later if you want. The wording implied to me that you can never use the HP from a proc to get another proc, which implies that a 2pt proc, followed by another gain of some kind (DL with ToR delay) would allow a 2pt -> 3pt -> 3pt chain, but would need testing.

Correct, this would be my mistake... I went back on forth on whether or not to even put those in the post like they are, and I guess I was not very thorough. I am thinking of changing the way the post is setup when the ptr patch goes live, since it will need some changing then anyways. Thanks for verifying, it was also listed as OOC regen when it was the formula for IC regen, also the one under Int was not taking plate spec and Kings into account.


@elvez

Thank you for putting so much work into all that, it really does a good job of showing a few important things. Still there appears to be a few minor things that need changing.

Should be:
Also don't forget that PotI is cast on self, and thus gets double benefit from Divinity, which is not transferred through the beacon, so really it should be:
Unless you want to do your formula another way.

The SP one is a 3pt, and AP is a 1pt, needs to be 0.209/0.217 or 0.627/0.651, using rough numbers, also the Glyph should effect both Coefficients as well as the base value.

Is your IoL generating the proper 1.5/(1+Haste) GCD when it gets used? It doesn't appear to be currently.

This is not really correct, but the method you are using cannot really avoid it. Restricting yourself to the HS cooldown is a loss in all situations that I found in my spreadsheet, and will cause issues evaluating haste. Putting in a 1/3 GCD each time for WoG is troublesome too, since it will delay 3 HS instead of 1. HS is about 75% of a HL, so if you can fit Half a HL in before the HS cooldown, its worth it (ignoring other factors).

My sheet came up with a series of 'rotations', which were arbitrary and could probably be improved with a better ordering. HL=DL for simplicity.

HS-> 3HL -> HS -> 3HL -> HS -> 2HL -> HPdump = 3HS:8HL At higher haste turns into 3HS:9HL

While this pushes back the Holy Shocks under most haste values, it is also better HPS then 3HS:6HL, which has slightly higher HPM due to holy power. As haste increases, you add more HL/DL in unless HS becomes a null factor and is skipped, which only occurs under DL or higher levels of output at >40% haste.

I am also not sure if I agree with ranking the stats on heal value added, when using a collage of spells with differing cast times. It should be Heal / Time in this case, since Mana is not a factor.

Using the 359 Focus on Haste:
Total Haste: ~35.41%

+Heal

Stat	+FoL	+HL	+DL	+HS	+WoG
+100 Intellect	258	137	340	114	268
+100 Haste rating (naive)	392	259	534	0	0
+100 Haste rating (including Daybreak)	287	189	391	52	78
+100 Crit rating	169	105	237	87	196
+100 Crit rating (factoring IoL)	169	105	237	87	196
+100 Mastery rating	121	67	178	54	143

+HPS

Stat	+FoL	+HL	+DL	+HS	+WoG
+100 Intellect	233	74	184	103	242
+100 Haste rating (naive)	354	160	289	0	0
+100 Haste rating (including Daybreak)	259	102	212	47	70
+100 Crit rating	152	57	128	79	177
+100 Crit rating (factoring IoL)	152	57	128	79	177
+100 Mastery rating	109	36	96	54	129

This would make the weighted values have more meaning, which I would also change. It would also allow haste to effect the HPS value like it should.

Given that the largest reason for getting the set is the 4 piece, and that the competition stat, mainly haste, achieves the same effect on Holy Radiance, but also increases the HPS:

The ~5.88% haste is not enough to outweigh the combination of the Crit, Int, and 4Set Bonus, in terms of efficiency. In terms of maximum output, the haste set wins by far. Given that Holy Radiance is already very efficient, and that the primary 'secret bonus' of the set was Mana Tide, which is being altered, I am not in much hurry to get the set either way. I don't believe these numbers have reforging counted, so that will benefit the Tier set. The mana cost of HR can be considered trivial. 40%/30s compared to 9%/2.5s for Holy Light.

Today I was looking around different sites and found a post talking about Conviction being bugged. I decided to test this.

I did not take screenshots or a video, however, it would not be hard to test for yourself. As of now, Conviction is only giving 3% increased healing to yourself per stack (at 3/3), not the 3% increased healing by you per stack that the tooltip says. I'd imagine it will be hot-fixed but as of now it seems like a waste of 3 talent points.

For those interested, testing was done by speccing out of Divinity, casting Holy Radiance with a helper right under me, and observing how much it healed both of us with 3 stacks of Conviction and without any stacks of conviction. The healing done to him was constant as stacks of Conviction increased, however, the Holy Radiance did heal myself for more as the stacks increased.

Edit: What I said may have been misleading, I'll quote myself. "Conviction is only giving 3% increased healing to yourself per stack (at 3/3)". I did not test this with another person healing me as I stacked Conviction. It may also increase healing done by other people to you, too, but to me that doesn't seem likely.

Can't test it right now, but according to Wowhead it does:

According to Wowhead EG really seems to have an 15s ICD.

The PotI "variable" on that formula was the value calculated just before, which already factored in the beacon and critical strike chance. I'll edit the post to make it more clear, though. About Divinity, yeah I forgot about that. From your formula it looks like it stacks additively so when healing on self it gives +12%, am I right?

That was another mistake, thanks for pointing that out.

What effect does IoL have on GCDs? I must've missed something again in your OP, because the talent tooltip only refers to casting time.

I thought about modeling a more realistic rotation, but it was going to become complicated. Anyway I'm not delaying anything, I'm just "splitting" casts while forcing HS use on cooldown, which I guess is inefficient in term of HPS. But I think the fact that I never waste time by allowing partial casts may compensates for that.

Yeah, that's true, it's pretty much pointless now that I think better about that. It isn't hard to update my formulas to use HPS instead of raw increases, and that's what I'm doing right now.

EDIT: Post updated with HPS values. They don't seem to agree with yours though, as I don't get 35.41% buffed haste from the haste focused setup, only 17.53%.

The int and spirit values are buffed, after modifiers have been applied (e.g. +5% from BoK). I.e. the values you would see in your char sheet at the start of the fight. Although to be precise, the resulting lines (where 1 int and 1 spirit provide the same healing per fight benefit) are for 1 unbuffed int or spirit - i.e. the numbers you see on the gear, not how much they changed your buffed stats by). Which might seem peverse to some, but was more intuitive to me since what I'm looking at in game are my current buffed stats and the stats on a new piece of gear, typically.

Plus using the buffed stats makes it more general - people can factor in whatever buffs they think they are likely to have and work out their typical raid stats from that, rather than hoping that I've done calculations using the same set of buffs that they actually use.

I didn't include PotI since I figured it'd be getting too much into situational assumptions about how often that healing was actually useful (mostly by producing overhealing on yourself). It would produce some change though, unlike beacon of light. I've clarified my thinking on the subject, and it isn't actually about the base vs spellpower portion of the heal. Basically, if an 'average' spell heals for (Base + Spellpower x Coefficient), then there is a term in the int vs spirit equation that is proportional to:

Coefficient / (Base + Spellpower x Coefficient)

In the case of beacon of light, it effectively multiplies Base and Coefficient by 1.5 each (when healing someone other than the beacon), and in this term those 1.5s cancel out exactly producing no change whatsoever. Coincidentally, it also explains why spell choice was having next to no effect: Holy Shock, Holy Light and Divine Light (and Word of Glory, for what it's worth) all have virtually identical Base/Coefficient ratios, so varying the mix of those spells has no effect on the int vs spirit balance. So it only really depends on the fraction of casts that are FoL (or HR, but that's on 30 second cooldown and so doesn't factor in so much). And I personally very rarely cast FoL.

PotI has a very different Base/Coefficient ratio however, so would actually produce a change if plugged in, so I'll look in to how much of a change it would make (depending on amount of wasted healing produced by it).

On the above topic: it seems pretty situational, but on some fights i got pretty surprised how well it worked. Depending on your raid composition, your healing role etc LoD might come up to 30% of your healing, considering this is the smartest way to use the HoPo generated via CS. As I am not really in maths I will provide a log from our Nefarian wipes last night. I was on Ony tank duties.

WorldOfLogs-Nefarian-Report

Sorry, it was a bit confusing there. To my knowledge the Divinity is multiplicative, so [PotI * 1.06] * 1.06 for the heal on yourself, with the Beacon getting Poti * 1.06. It is just the aspect of beacon that transfers the healing before target based modifiers.

IoL will reduce the cast time of HL/DL below that of the global cooldown. While this has no effect on instantaneous HPS, it does effect the sustained casting picture. The time taken by an IoL cast is under 1 second, but you cannot start a new spell until 1.5/(1+Haste) later, which is always greater then an IoL cast, which results in all IoL casts being restricted by your haste value. If you don't have this GCD in the factor, you will get more casts then you should, and haste will have a lessened effect compared to what actually happens.

For your Haste set with 1904 haste rating:
1904/128.057 = 14.87%
1.1487 * 1.09 * 1.03 * 1.05 = 1.3541 or 35.41%

I am not really sure how you got 17%, but I tend to always assume all haste buffs since they are common enough. Hope this helps.

@Evarella
If I understand you, those very nice graphs will shift in favor of Int when you consider Kings and Plate spec? I might be misunderstanding what you meant by unbuffed though. Thanks again for those charts, I might have to end up consolidating some of this great stuff into the OP, with everyone's permission of course.

I was more interested in the PotI component that heals the beacon, as it's basically a free heal on the beacon-ed target for each healing spell that is cast. And given the low spellpower coefficient of PotI, it may be another factor in making spirit more valuable than intellect in some situations.
I agree with you in using buffed values for evaluating the values of stat, but using raw values for comparison, as we're mostly interested in comparing the stats on the available gear.

Ok, thanks for the clarification.
Now I see what you was trying to tell me. Indeed, the effective cast time of a spell can never go below (1.5 / (1 + Haste)) seconds, even if IoL procs. To be honest I wasn't aware of that, but now that I've read about that, even if hasted, the GCD can never go below 1 second, right?

My bad again, for some reason I was using:

1 + (0.1487 * 1.09 * 1.03 * 1.05) = 1.1753

I fixed my post again, but introducing hasted GCDs made the HPS increases from haste skyrocket. There must be something wrong with my calculations, so for now the values on my post are with a fixed 1.5s GCD.
Please bear with me while I try to find what's wrong...

While I guess I'm just another person giving an offhand report, I have had Eternal Glory proc multiple times in a row. The most I have had it proc is twice since Cata came out, meaning I did a 3-stack WoG then EG proc'd and I did another WoG, then EG proc'd again and I did another WoG that consumed my Holy Power. Of course there is no way to prove that since you can't really screenshot something like that unless you're running a combat log, but I swear it happened. As for the ICD of EG, I have done this back to back (i.e. as soon as the GCD finished).

Since it (anecdotally) is possible to have 3 HP, get a EG proc, holy shock / CS hit for a 'wasted' HP and have EG proc again, I wonder if it is also possible to have gained an invisible holy power from blessed life via being hit, thus allowing you to proc EG again.

A combination of mixing in CS, HS and BL procs might invisibly be gaining holy power and allowing what appear to be multiple EG procs in a row.

Just a theory.