He's not saying get rid of all your crit, he's saying pick up the crit/haste spell power pieces (ie those geared for mages mostly) not the ones with spirit. Most fights, especially as disc, are not very intensive on mana so there's really no reason for it, and the fight he's talking about, Sindragosa, allows for large periods of time outside the 5SR = full mana.

So it would be Spell power > Haste > Crit >>> Spirit

Also to the poster above, mana is not really an issue. Never really has been for a long time. 95% of the fights in ICC either are too short for one to spend their whole mana pool or allow for tons of regen time. If you're not ooming and your tank is dying, then you should be stacking more haste/crit. And if you are ooming then you probably need better gear or better cooldown management.

What Veldefice is suggesting is perfectly accurate though, we're not going to outshine holy paladins obviously, but if its a weird day and one hpal is missing or you're doing a ten man and you're on tank heals, this is the best way to do it, not really much of an argument to be had. If you're healing tanks you want the highest hps, which is gheal with as much haste/crit as possible.

And for the record, I've downed Sindragosa heroic many times with 1hpal and 2 disc priests on tank heals, meaning that there are times when its just 2 disc priests on the tank when the hpal gets unchained, and never is a problem. We are weak tank healers, but not that weak.

I would agree mana is not generally an issue, I run all crit+haste, but I don't spam Greater Heal either. Are you saying you can spam GH and never have any mana issues in all/mostly crit+haste?

It’s also an issue of synergy. Yes, GH is more HPS than FH, but because we can't afford most of the talents for it, its expensive and lower HPS than other classes spam spells (lowest of all last time I checked). If you make a special build for it, you can get more of the GH talents, but even then, it'll gimp you for other jobs, though I would be interested in seeing the results of that, compared to druids and shaman (both also spec’ed tank healing to be fair).

I might do it this weekend if I have some time. I guess part of the problem will be the criteria: would we just look at base heals/self buffs? That wouldn't seem useful, so do we use full 25 buffs and what sort of gear? Again leaning towards tank healing would probably be good, but I'm not sure what would be best for the other classes, or even ours, considering fight variation. How long of a fight, what percentage 00FSR? Assume the fight was sufficient to cause a mana issue, in all haste+crit gear, how would we account for that? Or do we just ignore mana all together (not sure it'd still be meaningful, but certainly much easier) spec/gear each class for full throughput, and only compare burst single target throughput. That's probably what I'd end up doing, even of the usefulness of it is debatable.

Based on previous testing, I would expect it might be close, but we’ll be in towards the bottom, not right below pallies. Once you’re going for many of the GH talents in the lower part of holy, why are you majoring in Disc? Penance is great, and it will take work to account for the full rotation in Disc, but its probably also worth exploring deep holy, with more of a Disc minor than normal, though I’m not sure how far down to go into disc for the best return. Would it be worth losing the GS CD? Probably not in reality, but it may be the only way to stay competitive.

It's not that we can't do it, it's that we do other jobs better (standard tank assist (spike damage with FH being used unless you’re under borrowed time) comes to mind as being similar, while playing to disc's strengths) and other classes have more sustained throughput, not just pallies. If you have nothing but priests, or your priests are just that much better than your other healers, sure, but if you have capable shaman, why bother?

I think one of the major limiting factors preventing disc priests from tank healing effectively is the high base healing of holy light. While greater heal has a higher spellpower coefficient at the highest ranks of each, it requires a level of spellpower that is unobtainable at the current gear level (based on the information found at Wowhead). Coupled with the fact that talents such as Empowered Healing are inaccessible when speccing for Divine Aegis relegates disc priests to the support position. This is reinforced by both [Glyph of Holy Light] and Beacon of Light which both provide more utility when focusing a build around one spell. However in practice it may not be plausible to cover your tank healing needs with holy paladins and a disc priests absorption may synergise well with the raw throughput of a holy paladin.

As a disc priest, I was wondering if it was viable outside of encounters where you are assigned to raid shield to use borrowed time to cast instants before the stack falls off, especially below the haste soft cap. With that method, a priest may cast prayer of mending, renew, 3 channels of penance and a greater heal benefiting from the haste from borrowed time before the buff wears off. Would it typically be more throughput to shield, flash heal and shield again? Assuming a priest will cast at least 2 instant (or channelled) abilities and one with a cast time, would it be worth shielding someone who would not predictably be taking damage in order to obtain the haste buff?

Also, with spirit being a relatively low priority stat for disc priests, assuming a disc priest has obtained a full set of crit/haste gear, would it be reasonable to drop points out of meditation for situational abilities such as Unbreakable Will?

Finally, has any disc or holy priest considered using the [Scarab Brooch] trinket in current raids and if so, could they provide their reasoning for using it?

I wouldn't recommend dropping anything out of Meditation unless you're fully stacked on mp5, and have an '5sr' of around 800. Meditation is essential for tank-healing DiscPriests, and still a pretty decent throughput for all specs. I've raided ICC with Spirit Tap and Meditation separately, and Med was better by a mile (currently Disc/Holy, so my gear's not properly corrected.)

Here's the promised analysis, sorry it took a bit longer than I planned. Assumptions as follows:

1. Rawr - A tool I'm familiar with, that does a good job of modeling all three classes (better than anything else I know of). Please don't argue with the tool, though I'm happy to discuss the numbers it came up with. I'll try to answer any questions anybody has about the results, but TNSE will have to answer any in depth ones.
2. iLvl 264 - It's just easier to limit it (less choices) and it seemed an appropriate medium.
3. Choices were made favoring burst not sustained healing since mana is not generally an issue.
4. Rather than pick a random arbitrary fight length, I chose 10 minutes, to make sure the sustained value was as efficient as possible for longer fights. Only the Druid module calculated time to OOM with burst rotation, so I listed it for them.
5. Talents were chosen for tank healing as a distant first priority, if there were no more single target healing improving talents, I took the best utility I could find. Because of that, certain all but essential talents, such as the one improving CH for Shaman were skipped. Long story short, these are not builds I necessarily think are good, they serve one purpose and one purpose only: tank healing. The actually builds are linked in the Class (Style) names.
6. I picked the talents to start with and allowed Rawr to pick amongst the best of the top few (burst/single target) pieces for each slot using the Optimizer. If anybody wants the XML for the gear I can e-mail it to you or something. After the gear was added I checked the talants out and moved a few points around, then double checked Optemizations. I don't want to argue about it, but everybody got Tuskarr's since I believe it is essential for almost every serious raider, if you don't like it tough, but it doesn't effect the overall results anyways.
7. Spell Priorities were chosen by the Rawr, or at least based of the numbers it provides. Notes on the them follow the Conclusion.
8. I originally started out with DPCT as my primary measurement, but since both the Shaman and Priest model list everything in HPS, and the overall numbers are calculated that way for all the modules, I gave up converting it for the Shaman. The numbers in parenthesis are the DPCT, which I calculated manually based on the overall healing numbers for the spells by taking the overall heal and dividing by the cast time I'm not sure how TNSE has them displaying, but I don't think they take into account DA, maybe they average the crits in, but I'm unsure.
9. Paladins are not included. Since everybody knows they are best single target, I didn't see a point, and it saved me time. If anybody is actually interested, I'll do the work.
10. Overheal (anything about it) is not part of my analysis, and I don't intend to make it one. Yes, Druid Hots are likely to be high, but so what? You think GH's is going to be low?

Shaman
RT > HW (LHW for sustained)

• Mana: 30726
• Int: 1774
• Spi: NA
• Mp5 (casting): 305
• SP: 4109
• Crit: 54%
• Haste: 49%, 1159, 1 sec Base GCD

• Burst: 22036
• Sustained: 11262
• HST: 351
• ES: 1793
• R+HW: 19825
• R+LHW: 13047
• HW: 18356
• LHW: 14394
• R: ?

Druid
RJ > RG > LB > N

• Mana: 28251
• Int: 1669
• Spi: 952
• Mp5 (casting): 216
• SP: 4358
• Crit: 44%
• Haste: 36%, 463, 1.1 sec Base GCD

• Burst: 19207 (195sec)
• Sustained: 9840
• RJ: 2145 (24755.49)
• RG: 1127 (23335)
• LB(SS+Bloom): 1985 (18526)
• N: 19185 (19185)

Priest (GH)
PW:S > P > PoM > GH

• Mana: 32818
• Int: 1949
• Spi: 410
• Mp5 (casting): 300
• SP: 4018
• Crit: 50%
• Haste: 57%, 1223, 1 sec Base GCD

• Burst: 16855
• Sustained: 8650
• GH: 8802 (7026)
• FH: 6867 (5487)
• PoM: 4673 (4673)
• R: 896 (10751)
• PW:S: 8882 (8882)
• P: 12058 (12243)

Priest (FH)
PW:S > P > PoM > FH

• Mana: 32818
• Int: 1949
• Spi: 602
• Mp5 (casting): 371
• SP: 3951
• Crit: 49%
• Haste: 50%, 1009, 1 sec Base GCD

• Burst: 15165
• Sustained: 9390
• GH: 7643 (6143)
• FH: 6872 (5431) (33% sub-50%)
• PoM: 4614 (4614) (single)
• R: 739 (11082)
• PWS: 8781 (8781)
• P: 11312 (12125)

Priest (Disc/Holy Hybrid)
Renew > PoM > GH

• Mana: 32818
• Int: 1949
• Spi: 443
• Mp5 (casting): 313
• SP: 4192
• Crit: 44%
• Haste: 55%, 111, 1 sec Base GCD

• Burst: 14895
• Sustained: 7777
• GH: 11533 (9295)
• FH: 9144 (7371)
• PoM: 5468 (5468) (single)
• R: 1100 (13202)
• PW:S: (7711)
• P: 0

Conclusion:
Shaman are best, as expected, followed by druids, as predicted, though they don't offer the physical damage reduction. I'm not surprised the freakish Holy/Disc Hybrid came out last. As people suggested, the Disc GH spec can be the best single target HPS for a Priest, but at over 11k, the resulting heals are big, and will undoubtedly result in more overheal (I know I said no discussion, I lied but its only a brief mention!), which is one of the main reasons everybody uses FH in the first place.

Notes:
Shaman - One of the main points of Riptide is it's effect after it's cast (it increases your haste for your next two spells) so it it's effective HPS depends on what you cast next. I didn't feel like calculating its HPS for both options, so live without it.
Druid - Using Swiftmend lowers the Burst HPS. Maybe it's caused by Nursh's crit being boosted by talent?
Priest - I could not figure out how to make Renew part of the Rotations without manually entering the total number of times each spell would be cast, over the entire fight I suppose the fact that it can't crit might be part of the issue as well. Maybe TNSE knows?

I apologize for the length, but I wanted to make any important factors of the testing apparent to all. If you're still reading, congrats!.

Thanks for the analysis - I just chatted with TNSe about it - if you used custom priorities it may be that Grace isn't factored into the HPS calculations. For the predefined tank healing spell priorities, it is (as is Aegis).

This basically confirms my napkin math/gut feeling. There's just no way why someone should choose a priest over a shaman for primary tank healing. Druids are closer to priests than I thought, though.

Regarding renew, I don't know it really makes sense to include it for a disc tank healing rotation/priority list. If it's really a challenging tank healing job, I wouldn't really risk a GcD for it, actually I'd rather cast-cancel a GH if there's nothing to do for a moment or shield someone else (like the other tank) to have BT up.

Lets see, DA is not included in the Spell Stat Display. It is calculated for the Throughput and Sustained however.

The reason Renew is not part of Disc-Tank Role is that when I tried to incorporate it in my rotation during raiding (at Algalon), I found that putting up Renew might drop tank so low that he could die. So this made me decide to not include Renew as a part of Disc-Tank Role. You can argue that it will increase your HPS, but as Hegen pointed out (and I experienced myself on Algalon), using that GCD to put up Renew can (and did) get your tank dead.

Disc does not have the neat rotation that a Resto Shaman can have with Riptide-HealingWave-HealingWave-(Lesser Healing Wave-)Chain Heal-Repeat. And in the case I had with Algalon, I could not use Penance as part of the Rotation, I had to keep it in backhand for dealing extreme burst damage (or when other healer had to move).

My original intent was to have ratios. Problem with that was, what if you only cast 1 Divine Hymn and 200 PW:S? I did not want people to start inputting decimals or start calculating things themselves, so it ended up being more natural to ask them to input data directly from logs.

It is not hard to add Renew to the rotation, but if you are focusing on tank, the added FH could add a DA, refresh Inspiration, with 2pt10 add a HoT, while Renew waits 3 seconds before it ticks even the first time. And there is no guarantee that it ticks when you need to, and it does not tick for enough to make it be felt. (900 ticks from your numbers, while a FH is over 6500.). This is just my personal opinion however.

RE: MADMark post - I have a few questions on this:

1) Why did you not lower the druid GCD to 1s like the rest? (it is stood at 1.1, 400 ish haste seemed low to me, thought it would usually be 856/735 depending whether you took CF).

2) 11k heals comes from the disc/holy hybrid, not the GH build.

3) - TNSE states that DA is included in throughput and sustained - so the numbers are 8000/9000 hps without or with DA?

4) And for the shaman, there's no mention of Ancestral Awakening?

Just trying to get some understanding through discussion.

I didn't really intend this to become an argument about HoTs, but since we're here, we're here. Obviously Renew is low HPS, but its HPCT is excellent. If you're target will die in the GCD, it's obviously not worth it, but if they are likely to live, why not? All of this is a judgment call, how much you like the efficiency of HoTs, your heal comp, etc.

I was more wondering if the custom role option is as versatile as it can be. Maybe nobody will put renew into their rotation without fully speccing into it, but why not make it easier to find out if it's actually worth it to you? Obviously, we aren't Druids, but we aren't Paladin's either. We're the most versatile healers, with more tools the anybody else, don't we deserve a model to match? You do a great job with it (thank you for that), and I think it’s a great tool. I’m just making a suggestion, do with it what you will.

An option for the disc healing roles would be one way of solving it, and probably the easiest one at that, but personally, I think allowing for different ways of entering the Custom Rotation is the way to go. It would solve not only this minor issue, but allow people more flexibility in general. I realize it's more work overall, but it seems most beneficial to me at least.

Even just having a Check Box or something that multiplies the number of spells you enter out based on the length of the fight and the % you choose to be I5SR, so it have any unintended dead time would be great, while still allowing the people who want to list the specific cast counts for the entire fight to use it that way. I realize you have a lot of stuff going on and will prioritize however you choose, I just wanted to throw it out there.

@Horlicks Sorry I didn't notice your post at first.

1. My theory is that the one second GCD is not as desirable for druids because of Celestial Focus (or whichever talent increases haste after crits, sorry at work).
2. Sorry my comment on the size of the heal was based on the actual heal amount, not the HPS or HPCT value, which I did not post, as TNSE mentioned, it does not take into account DA at the spell level of granularity, but it does account for crit I think.
3. As I mentioned, DA is accounted for in the totals (Burst and Sustained, not the individual spells), so yes.
4. As I understand it, based on the fact that the talents such as DA, LS, and AA increase the overall healing, my assumption is that Rawr assumes 100% of the secondary effects will be used. At least I saw no place to choose the percentage of them that would actually get used and I can't imagine anybody picking a random number such as 50% and making everybody live with that, but I didn't spend much time looking for it either, you'd have to talk to the developers (such as TNSE) of the modules for confirmation.