Some issues of note, overgeneralizing a single parse into grandiose, unjustifiable pronouncements about the Priest class as a whole:

1. Expand Tiaran's full healing for the evening and we can see some interesting aspects of his changing critical rate.

a) Flash Heal/Penance have virtually the same critical rate, slightly higher than Holy Nova's. This indicates the heavy use of Renewed Hope (affecting both) but the +critical portion of IFH being nearly worthless (only affecting Flash Heal).

b) Prayer of Healing has a strongly elevated critical rate, almost certainly due to Inner Focus use.

c) His Glyph of PW:S healing clocked in at 21% +/- 0.7%. Given this is "gee, I won Powerball twice" level likely under pure random chance, I'm not sure what is going on here. I suspect that Holy Specialization doesn't apply which seems obvious in retrospect but never occurred to me. But that doesn't quite bridge the gap, so there might be something else going on here. Maybe Tiaran has a healing style that involves routinely dashing out of range of totems whenever he's thinking of casting a PW:S.

d) Discipline Priests should be forbidden from ever casting Renew since the lack of an ER upfront heal makes it a nuisance to analyze the data.

2. Wreath was convieniently Holy the entire time so we get better aggregate data than Nidaba. Indeed, he's so hardcore he pfahs at those puny mortals who need the crutch of a second spec.

a) Renew had 991 casts (3/3 ER), indicating 2.14 ticks on average. So Wreath's Renew was not only twice as mana efficient, but he actually overhealed (including non-healing ticks) less than he did with Flash Heal. A point in blessed resilience would have increased his total healing by about 500k. Very roughly, each point in ER is worth 2% on the ticks and then the upfront heal, so he's getting about 300k out of each of those extra ER points. Note that this is total healing, not effective healing (no way to tell what would overheal).

Comment: The ER vs. BR choice isn't just a matter of healing performed, but also in terms of how much healing ER might potentially allow you to shift from Flash Heal -> Renew. Since, at least for Wreath, casting a Renew to heal a single target is about twice as mana efficient as casting a Flash Heal, those two points in ER provide some (unknown model) worth of implicit mana 'regen'.

To clarify this, if Wreath had simply replaced his Renews with Flash Heals, he would have had to cast 1.83 more Flash Heals than Renews to achieve the same amount of healing. This is ~610k mana saved by the playstyle decision. With 1/3 ER, this drops to 1.62 more Flash Heals, dropping the difference to ~465k. So, in some sense, each point over the one 'mandatory' ER point is worth the throughput stated above plus ~72.5k mana over the course of the raid. That is, if you're Wreath healing the linked raid.

b) He had 890 procs of Surge of Light, and an estimated 505 consumptions. This is a total savings of ~350k mana. His Holy Concentration had 1h18' uptime, so presuming 800 oFSR, that would be ~190k mana regenerated. Worst-case (all multi-target heals always hit all possible targets), Healing Prayers saved ~75k and if he had taken 3/3 MA, he would have saved ~160k. If we claim that multi-target heals averaged 4.5 (+1 for glyph of CoH) except ProM with 2.5, these numbers become ~180k for MA and ~110k for HP. So per-talent-point, SoL > HC > MA > HP for these fights.

3. Nidaba. Nidaba switches specs back and forth, so aggregate figures get a little sketchy and we have to use more limited data sets. So the only real issue of interest here is Surge of Light procs.

Let's compare three unfortunately very dissimilar sets of data:
This specific fight. 42 procs, ~35 consumptions, 43% of FH free/instacast..
This entire raid. 795 procs, ~535 consumptions, 37% of FH free/instacast. (Neat trivia: ~200 of those unused procs were overwrites).
The entire raid from post #177 (Thorim). 297 procs, 265 consumptions, 22% of FH free/instacast.

Two discrepancies worth noting:

a) 43%/37% in first two parses. My suspicion is that this is due to a fair amount of Discipline data corrupting the second analysis.

b) 43%/37% vs. 22%. In the #177 parse, Nidaba has 1/2 SoL. In the other two, he had 2/2 SoL.

Ideally, Nidaba should have wiped the raid repeatedly at 1% with varying levels of Surge of Light to give us a better data set, but I suppose we'll just have to soldier on with what data we do have.

All of the various Holy parses demonstrate that Serendipity has an almost negligible impact on raid-long haste, but we already knew that. There doesn't seem to be any decent way to determine how it impacts response time vs. risk without re-parsing the entire raid with a complete different parsing mechanism. Nor can I figure out a way to reasonably test for the actual value of Test of Faith with a complete re-parse.

15 shields / 15 seconds, absorbing 7000, dealing 45% damage back, 35% crit, 14% miss.
15/15*7000*.45*(1 + 0.35*50%)*0.86 = 3183 dps. Too bad its mostly reflected back at Freya for no efficient DPS.
I feel I did something wrong with these calcs.

Obviously I did, too bad. I forgot the Reflective Shield only works for yourself. Could get interesting if they ever decide to fix that. :P

Sorry, I was just looking at all the BiS lists and saw some people had the Leash and some had the Cord. Thanks for the reply.

Re: [Leash of Heedless Magic] vs [Cord of the White Dawn], I think the best argument for the Leash (and the main reason I'm still trying to drag my guild back to Maly25) is just the sheer lack of haste options available in Ulduar. Maybe I've just been unlucky with drops, but my haste is really low compared to where it was pre-3.1. Crit is on *everything* it seems. Thankfully I was able to pick up t8.5 shoulders last night, but my haste is still not as high as I would like it to be.

I think you're using epic gems in those numbers that are not available in game. If you're gemming spower and spower/haste as it seems the numbers for the Cord [not including belt buckle for obvious reasons] would be -

116 spower
8 haste

for a total gain of- 7 stam, 2 int, 13 spower, 48 crit at a loss of 54 haste. They're very close and I will probably end up making the Cord once I get more haste on my gear. Right now it's lower than I'd like so I don't want to drop it further for minimal gains from other stats.

It's not that I value haste over crit; rather, it's that I lack a lot of haste with my current setup due to having a bunch of crit pieces (Not that you have to choose, necessarily, but I'm taking mostly Spirit items, therefore the choice is between crit and haste). I'm not so interested in the value of the piece in terms of its item budget as much as I am with reaching a better balance of stats. As Nidaba mentions in the OP, there are "sweet spots" for crit and haste. My argument is more "this feels better" than "this is more mathematically sound." And that's only because it seems like all the cloth dropping in Ulduar has crit instead of haste.

It doesn't help that the most accessible Priest T8 items are the crit pieces.

Wreath and I have been looking at the gear from Ulduar, and basically until you get your hands on some of the hard-mode haste items (Algalon cloak, Mimiron ring, XT bracers), you struggle to keep haste above 400. It's really really easy to get crit, and not so easy to get haste.

For this reason alone, I'm sticking with Leash until I have some of those pieces, so I can keep my stats balanced. You all know my opinion on "optimal" levels of stats -- I prefer to not have haste too high (~ 425), and then crit and spellpower get stacked as high as possible.

[e] This is, of course, assuming you don't just gem for haste. If you do that, then it's trivial to keep any sort of cap you'd like. Annnnd, speaking of haste, XT dropped the bracers tonight! So keeping haste high just became trivial.

I am mainly discipline and pretty much only spec holy when I solo heal Hodir hard mode and 2 man heal Freya hard mode.
I have noticed that when I do spec holy, I go out of mana fairly quickly. Obviously these two specific encounters are extremely HpS dependent and require a lot more spamming than most others. Still I can't help but wonder if I could really sustain my mana pool as holy on other encounters, specifically heroic hard modes on which I have no experience.

Anyway, that's something I have been thinking about for a while. Now I see you linking quite a few pieces as BiS with no spirit at all. Would you really be able to wear all of those pieces and not have any mana problems?

I'm asking this mainly because I'm trying to find gear that would suit both discipline and holy specs, and the gear you linked with stamina/intellect/haste/crit/spell power would most definitely suit both specs, if I could sustain my mana pool as holy with it.

OK, I'm dumb: sorry above, I missed the <= 226 gear. In that case, I'd swap Signet for the hard-mode Yogg.10 ring (socket, crit+spirit), and leave the others as is. My apologies.

Replace [Signet of Manifested Pain] with http://static.mmo-champion.com/mmoc/...26_9806_11.jpg.

Replace [Unsullied Cuffs] with [Grasps of Reason].

Replace [Pennant Cloak] with http://static.mmo-champion.com/mmoc/...alon25-man.jpg.

Replace [Spellslinger's Slippers] with [Boots of Fiery Resolution].

Mostly, my rule of thumb is that I can afford 1, maybe 2 slots where I have no regen (i.e. non-spirit-based). Items like the [Nebula Band] are good examples of where you might want to do so. Realistically, though, we have access to some *good* ilvl 239 haste-based items, and there's always the argument of supply vs. demand. At least in Fusion, we run ~ 13 dps casters on the roster, plus 4 spirit-based healer types, and another 3-4 crit+haste-based healer types. Those of us who can use spirit-based items tend to do so. The only exception that we've allowed so far is [Cosmic Lights], which was only because all the dps were using the ilvl 226 hit version from the quest reward.

The [Nebula Band] is a great item that we *might* get access to, since 10-mans are much easier to justify (or much easier to stack with people who already have Signet+Pyrelight, for example).

And in the end, ask yourself how much gain you get from an item like that, vs. a dps caster for whom it is probably BiS forever. We *still* don't have enough [Signet of Manifested Pain], and we farmed KT for ~ 5 months. We've seen *one* [Pyrelight Circle]. And without an Algalon 10-man kill yet (maybe this week!), we haven't seen any [Nebula Band]. Our casters are salivating at the chance to replace their Kirin Tor rings or (even worse) other random stuff they have. I'd much rather they get the huge upgrade than me take the sidegrade that has no regen on it.

@ Prrancer: holy can be incredibly mana intensive, especially when in a pure raid heal role, and most especially when on a fight with heavy raid damage. Wreath and I make extensive use of innervates, and gem/gear for more regen than may be the norm in part to compensate this. When it comes down to it, I run a moderately different set of gear for Discipline. If you're actually interested, I can send you the lists of my Holy gear vs Disc gear, just so you can get a feel for how I set it up. I replace a ton of the haste-based items with pure spirit+crit items in order to dump unnecessary haste, and get my crit back up where it should be for Disc.

He said for <=226 gear. But yeah, I definitely agree with your points about generally not taking BiS dps loot when there's a regen option that's just as good (or better) for you but not as good for them.

I've run into a similar problem by min-maxing to get as much crit as possible since its pseudo-power is higher. Haste becomes extremely important (more so than crit or even spellpower) on fights with high amounts of dispelling such as Yogg phase 2 or Thorim hard mode. In both these fights the priest's ability to keep people alive relies quite heavily on the ability to get debuffs off people, and the limiting factor for this is often global cooldown.

Perhaps we need smarter item stat weightings? I'm not sure how to implement this. Throughput with a certain number of dispells per minute might work, but even then you run the risk of having so many assumptions in your model that you are effectively over-fitting your parameters. Surely there is a better way than "feels better."