You already have so much Haste that the Haste on the Circlet is offsetted by the extra regen from Cowl's Spirit.

The short answer is that it's entirely subjective which is better, if you ignore the set bonus then it largely depends on whether you lack throughput or regen. They are both decen items and perfectly usable, if I use my standard stat weight I rank the Shadow one higher, but I get around a 2% difference which is fairly insignificant on one item. I won't be seeing Yogg for a while so I took the healing one to get the set bonus quicker.

It's been some time since the list of trinkets has been edited and many new ones have been discovered. I thought it would be useful to compile a spreadsheet to 1) give an updated list of trinkets and 2) evaluate the oftentimes subjective processes of comparing procs. The blue cells in the table are meant to be updated so that one can evaluate with different stat weightings. Options are available for on-use and on-proc stats since some players may find these effects more or less valuable than those always present on the item. For example, setting the "Proc Weighting" to .5 means that a trinket that procs 40 MP5 with 25% uptime, it would be considered equivalent to a trinket that had a static 5 MP5 (40 * .25 * .5 = 5). A "Combined Ratio" of .6 indicates a 60/40 split between throughput and longevity. I have listed a few of the best trinkets (using the weighting from the beginning of this thread) below with their equivalent SP or MP5.

Throughput:

• [Illustration of the Dragon Soul] -- Sartharion 25 (200)
• [Scale of Fates] -- Thorim 25 (176)
• [Eye of the Broodmother] -- Razorscale 10 (167)
• [Pandora's Plea] -- Mimiron 25 (159)

Regeneration:

• [Meteorite Crystal] -- Algalon 10 (126)
• [Spark of Hope] -- Kologarn 10 (106)
• [Pandora's Plea] -- Mimiron 25 (80)
• [Soul of the Dead] -- Sapphiron 25 (25)

Combined: (60/40 weighting)

• [Pandora's Plea] -- Mimiron 25 (159/80)
• [Illustration of the Dragon Soul] -- Sartharion 25 (200/0)
• [Scale of Fates] -- Thorim 25 (176/0)
• [Eye of the Broodmother] -- Razorscale 10 (167/0)
• [Show of Faith] -- Yogg-Saron 25 hard (140/16)

I don't know if you purposefully left out [Forethought Talisman], but its value in 100% throughput and 60:40 throughput:longevity is 171. Unfortunately, it has no longevity stats. That makes it 3rd BiS for 100% throughput and 5th in 60:40.

Throughput/Longevity

• [Forethought Talisman] -- Multiple Bosses Naxx 25 (171/0)

I appreciate what you're doing with the spreadsheet, which I think is useful. However, I really do not like the 60/40 split, which is totally and completely arbitrary. It was used as an example but without any justification for those numbers.

I'd prefer that everyone come up with their own weightings, but if some "neutral" comparison between mp5 and spell power is needed, use blizzard's own stat budgets unless/until someone provides some justification for something better.

If people want to use a 60/40 split, then fine, but they should realize it puts very low value on mp5 relative to spell power compared to the item budget, so using that split puts very strong emphasis on throughput -- to the point that the forethought talisman blows away spark of hope and meteorite crystal. That fails a sanity check as far as I am concerned.

I was considering whether to include Grace 2/2 in my spec or to stick with 1/2, so I wrote a little macro to parse some of my logs and give me an idea of the relative value of the two points. Basically the macro just goes through the log, keeps track of who grace is on and the size of the stack and records the number of casts of each spell on people with each stack size.

Context: These logs are taken from Ulduar 10 raids, excluding Vezax (healing Hodir may be awfully different than healing Freya, but I feel like I'm playing a completely different game on Vezax). Hard modes for Thorim, Hodir, XT. Freya +2, Yogg -1 and easy mimiron. I cast a lot of AoE heals, from what I've read, way more G. Heals than a typical priest (though that's probably because most people are raiding 25s).

Results are written as:
Spell - Average Grace stack when cast (1/2 Grace) - Average Grace stack when cast (2/2 Grace)

Flash Heal - 0.31 - 0.33
Penance - 1.35 - 1.69
Prayer of Healing - 0.19 - 0.20
Greater Heal - 1.11 - 1.29
Prayer of Mending - 0.92 - 0.97
Binding Heal - 0.63 - 0.64
Power Word: Shield - 0.35 - 0.38
Holy Nova - 0.04 - 0.13

Basically that second point of Grace would be netting me +1% effect on my Penance +0.5% effect on my G. Heal and a negligible effect on everything else, whereas the first point is +1% to +3% healing on most heals and +4% on penance. Doing this really convinced me not to bother with the second point in Grace.

I've gotten a lot of questions about my trinket rankings post, both on this thread and via private message. I will try to answer them as best I can. In my original post I placed a link to the spreadsheet, which was the real point of the post. I then linked a few of the best trinkets using the TheDoctor's weightings as a convenience for people who didn't want to play around with it themselves and just wanted a quick answer. I did not, by any means, expect them to be the final list of best-in-slot trinkets for all players. They obviously are not.

While most of the trinkets procs can be modeled as a static stat increase down weighted by uptime, a few cannot. The [Forethought Talisman] is one of the worst offenders and required quite a bit of a hack to come up with a value for it. Basically the way I modeled this one was assume it procs for a heal of 3572 once per minute, then calculate the average heals per second (hps) from the proc and to ask how much spellpower would result in an equal hps increase assuming you cast all flash heals at one every gcd. The other "hard" trinkets were [Meteorite Crystal], [Spark of Hope], and [Living Ice Crystals]. I show my work for these on the second page of the spreadsheet titled Proc Worksheet.

I can't reproduce your 171 number. It depends on what you use for the stat weighting, but if you get a score of 171 for 100% throughput, you should get a score of 171*.6 =103 for the combined rankings. Using my original weightings it came up fifth in throughput with a score of 143, and I included the first 4 trinkets for each list.

I agree 100% with this. 60/40 is an arbitrary number, and instead of coming up with my own arbitrary number, I just used someeone else's. Not only should people come up with their own throughput / longevity weightings, they should consider the individual stat weightings for Int, Spirit, Spell power, etc. also. Additionally you can play with the way "on equip," "on proc," and "on use" effects are weighted. For some people, 400 haste on demand for 15 seconds ever minute is worth more than 100 haste continuously. For other players, the opposite is true.

I expected other readers to play with the weightings themselves (I even colored the background of the key cells to edit blue). As stated above, the lists were simply for the convenience of people who did not want to do this themselves.

Not exactly. [Darkmoon Card: Illusion] is item level 200 and I entered only those trinkets that are 1) not clearly useless (e.g., [Blood of the Old God] and 2) item level 213 or higher. I would like to extend the list down through item level 187 and also add a column for hit so that this table can be used by dps casters as well. As of yet, I have not had time to do so. In the mean time, if there is a particular trinket you want to check, it should not be hard to fit it into the framework. The permissions should be set such that you can save a private copy.

I have been working through some possible measures for longevity, and I think it is possible to come up with a rough estimate for healing longevity (the time before you go out of mana), as long as we assume HPM (heal per mana) is approximately constant. Doing so allows one to estimate the rate at which you burn mana as HPS/HPM so that:



whenever the denominator is positive. Longevity is then defined as infinite at or below a certain key rate of healing, where the denominator is zero:



As noted in previous posts, this infinitely sustainable HPS and peak HPS provide a useful functioning range for HPS, since you can always sustain your infinitely sustainable HPS but you cannot exceed your peak HPS. So your HPS will generally be running between those two extremes. Importantly, we can simplify our longevity formula to:



which is useful because the denominator is then just a measure of how far your HPS is above your infinitely sustainable HPS. A simple and potentially useful longevity is one midway between those two extremes, namely:



This "median longevity" would give you a rough idea how long you could heal at moderate intensity. Conveniently, your minimum longevity (which is how long you can heal at your peak HPS) is exactly half your median longevity. So the practical utility of knowing your median longevity is that it is easy to do the quick math in your head and get a rough idea how long you have.

For instance, if your median longevity is 6 minutes, and you know a certain encounter (or a portion of it) requires near peak HPS the whole time, then you might estimate that you have a little more than 3 minutes. For light healing it might be slightly more difficult to estimate your longevity, but it hardly matters since longevity really takes off near your infinitely sustainable hps and becomes a non-issue. So you have a quick and easy way to estimate how long you can heal moderately to intensely, which is generally what matters most.

To figure out your median longevity, you will need to know your peak HPS (or throughput), your available mana, and 2 of the following 3 things:

• Mana regen
• Heal per mana
• Infinitely sustainable HPS

Once you know those, you can use the above formulas to figure it all out. If you are using a sustained HPS metric (a measure of how much dps you can maintain over a fixed time interval), then you can figure out HPM from:

The equip for Spark of Hope seems like it really isn't too hard to model, if you know what fight you're doing. For example, here's a fight were I was raid healing. My cast weighting was something like:

4 PW:S
3 Penance
1 Flash Heal
1 PoH

Given my Haste % and talents, I cast these in about 14.77 seconds. Round it up to 15 for simplicity (since I'm not casting 100% of the time anyway).

The equip says that each time I cast a spell, I save 42 mana. All of those spells cost at least that much, so 9 spells * 42 = 378 mana saved over 15 seconds.

Saved mana == regen'ed mana, so you can call it 378 MP15 or 126 MP5. To be more exact, I suppose you could add up ALL the spells cast over the entire duration of the fight and you'd get a more accurate estimate, since it would take into account different casting rates throughout the fight.

I believe the Spark of Hope mana discount is affected by talents (and perhaps glyphs, tho I'm not certain). While it ends up being small numbers concerned (5-15 mana per spell), this could contribute to the difficulty Kamaia had in modeling it.

I've been reading EJ for some time now, and really appreciate the discussions and tools you have all created in an effort to maximize efficiency and plumb the depths of possibility. Thanks a lot, everyone.

Judging from the logs I've been taking since 3.1, I'm not so sure that disc is "low on the meters" anymore.

First, yes, I'm aware that log parsers guess how effective absorbs are since it doesn't show up reliably in the combat logs. But in the ones I'm posting, I'm the only disc priest. The only other absorbs come from Sacred shield and the few shields the holy priest puts up.

But has any other disc priest looked into how effective their shields really are? If WOL is to be believed, PWS is actually powerful enough to get a disc priest on top of the healing meters (on normal modes anyway). Logs for 5 bosses done a few days ago:
World of Logs - Real Time Raid Analysis

But it really seems to depend on how many healers there are and the boss. For example on Auriaya, where I can easily pre-shield, I'm doing more "healing" by a large margin.


But after trying hard modes, DA actually seems to be my top healing spell on those, and really, there's not much time to shield the raid because of the less amount of healers and more damage on the tank. So it really seems that pre-shielding the raid, although incredibly powerful, doesn't have much place in hard modes which calls for dps stacking.
Heartbreaker: World of Logs - Real Time Raid Analysis
Steelbreaker: World of Logs - Real Time Raid Analysis


Oh and I find it absolutely ironic that the Ret paladin is actually 1st on heartbreaker and almost 1st for Steelbreaker only due to JoL -.-


Don't. There are better glyphs out there. The POH glyph usually just gets healed over and is pretty crap for both disc and holy.

The only time pre-shielding is really effective is when you can:
a) Predict a lot of incoming damage.
b) Prevent gibs by effectively raising people's HP by 7k+.

For these reasons, I use mass PWS spam only on two hard-mode fights: Freya+3, and Vezax. Freya, because the Tremor + any of the other sources of damage is sufficient to kill a mage or similarly geared cloth-wearing caster, and Vezax because you can absolutely predict the incoming damage, and pre-shielding as you enter Animus phase means you "eat" most of the first ticks. It's also the only spell that works while standing in Crash (at full efficiency).

It still makes sense to throw some PWS around on Heartbreaker, since the combination of Tantrum and Gravity Bomb will kill anyone in the raid, as will (if not carefully healed) Tantrum and Searing Light. It's just not feasible to mass PWS in those situations, since there's so much incidental damage that will eat the shields before the actual spike. Besides, Heartbreaker is one of the few fights where Body and Soul can really shine, and if you hit the whole raid with WS, they're not going to be helped while moving out of the raid.

As far as Steelbreaker goes, it makes far more sense to be tank healing for P1-3 than to be a somewhat effective raid healer. Disc is the only tank-healing capable class that can actually solo heal the Runemaster tank for all of P1 without going OOM in the process (possibly rDruid, although they lack Inspiration and absorption). I heal the Runemaster tank, then assist on the Steelbreaker tank (dispels, etc.) and help set up for P3, then spam heals on the current Steelbreaker tank for the ~ 85 seconds of P3.

PWS spam can be decently effective for Thorim and Hodir as well. It depends a bit on your guild and their setups. On our first kill I was PWS spamming the raid in between throwing Penance on the tanks in order to raise the "effective" hp of some members. Priority was mage/shamans/priests.

PWS can be decently used to pre-empty Frozen Blows as well.

A quick reaction to what you said is that PW:S is incredible to "steal" healing when there is too much healers.
The issue was known with COH. At 3.0, instant, intelligent-targetting raid wide COH was insane, especially when healing was easy. You could heal faster as any other healer could, and you basically snipped heals, and shined on meters.
It seems to be the same here: when there is not enough damage, you pre-shield the raid, and heal even before the damage, meaning that you can still heal at "full" time when others have to overheal or take a break.
The real question would be on bosses where you have not much healing margin. Or wether your PW:S where needed or not.

I'm not saying that Disc is bad. It's the spec I'm playing (with holy dual now) since WotLK. But measuring meters when there is too much heals just measures the ability of the class to be quick.

This is very true. On Ignis, for example, I can easily top the healing metres by a good 30% or so by spamming PWS the entire fight. This is simply because the Holy Priests and Resto Druids simply run out of damage to heal, while my PWS spam continues the entire fight. It's very amusing to look at the healing metres to "prove" how OP Disc is, but it really doesn't mean much.

We followed that by trying XT-002 hard-mode with 5 healers (and failed miserably, incidentally), and in that fight PWS spam can't compete with real AOE healing. There's so much damage flying around that no one runs out of damage to heal, and the superior HPS of Holy and Resto really shines.

That's not to say that Disc would be bad for that fight; if you went with 6 healers (3 AOE healers), a Disc Priest helping by spamming PWS would be a great addition. With only 2 AOE healers however, I really didn't feel comfortable with the Disc healing output.

I would very much like to explain how we do XT hm, since a lot of people seem to struggle healing it - but I don't know if it's still off limits to discuss, since its a 25man hardmode (got nerfed anyway...).

edit: I just looked it up - it has been widley beaten since the nerf so I'll just go along and edit this post.

On normal mode yes, but on hard mode, if you're the only tank healer, there really isnt much time to shield the raid. But on Thorim, it's really nice that I can shield the 2nd tank for Thorim before the switch, while other healers have to wait for the switch to land anything -.-

And being quick isn't a good thing? Personally I think that being able to add health to raid members in between damage is far superior to actually waiting for a health drop and then healing it. It may be "heal stealing", but imo, it's the better way to go. When you're waiting for hp to drop, there's also a higher chance of people dying.

And if you'd have read the rest of my post, I did say that even though PWS is incredibly powerful on normal modes, on hard modes, it really isn't a match for real aoe healers who do have much higher hps. For example, a holy priest can within 1 GCD heal the same amount of people it takes me 6 GCDs to shield (a reason I think PWS spamming can hardly be compared to 3.0 coh), without even having to find the people who need healing like I do >>. With the heavy damage a tank takes, there isn't much time to cast shields anyway. But it does balance out, since on hard modes, aoe healers are no match for a disc priest when it comes to tank healing -.-

Personally, I think if you're capable of pre-shielding for incoming damage without having to worry about the tank dying, then, yes, it's needed. That's how Disc was designed to work, if you don't "heal steal" with pre-shielding, what are you supposed to do? Wait for the tantrums to land and then cast PWS during it o_O?

Maybe we were too light on healers or something, but on heartbreaker, I really didn't have much time to shield the raid, at least not as much as I could on normal mode.

There is something extremely fishy about the healing numbers. The log says that 48.3% of your heals were from PW:S, and 16% were from Divine Aegis. Since PW:S doesn't trigger DA, that means that means that Divine Aegis was 16% of your heals while spells that can trigger DA were 100% - 48.3% - 16% - 5% (Glyph of PW:S and Renew) = 30.7% of your heals. This makes your total shields from Divine Aegis equal to 52.2% of your healing from spells that can trigger divine aegis. It took me a while to figure out how this was even possible (since this would ordinarily require a crit rate of 174%) and then I saw the overheal numbers.

If shields are counted as 0 overheal always then you can really rack up number on the meter by just casting it on cooldown, but if your other heals are coming in as high as 80%, is it really reasonable to assume that your shields were all 100% absorbed? Even if they are largely fully absorbed, the only reason you are winning healing meters is because shields are the ultimate in heal sniping: you get to drop your heal before the person even takes the damage.

In a raid where healers actually have a lot of damage to heal and where you assume some shields are wasted (not a high amount, but 10% is more realistic than 0% except) I wouldn't expect to be winning meters, but as we've been discussing in this thread and the general 3.1 healing thread, maybe healing meters aren't really a very good tool for measuring performance (this actually serves as a great example of why that is).

Er what? I can't even fathom why a Disc priest would be the _only_ tank healer for Thorim hardmode. Unless you're talking 10-man, but I was not. And people don't need "Sif gib" protection in normal mode either.

In my experience on Thorim, raid healing hasn't been an issue at all, dispelling and gib-protection seems more important. Our first kill we ran 7 healers, 3 tank healers, 3 raid healers and me as Disc to PWS low-hp raid members.

Also, to all the people talking about Heartbreaker, are you guys talking about post-Hotfixed Heartbreaker or pre-Hotfixed? Because with 6 healers I've found quite a bit of "downtime" post-nerf. Pre-nerf was more like spam spam spam spam.

In my lengthy experience using recountguessedabsorbs and now World of Logs the shield tracking seems to have around a 95% success rate. The main reason for this is it gathers the log data of "(xxx absorbed)" rather than simply calculating the strength of the shield when cast/proced and assuming it will be used. The only problem this method seems to have is when multiple absorb sources are on the same player at one time (the most common is on a tank - shield blocks, savage defense, anti-magic shield etc). For these on recount I will see a maximum shield absorb higher than the true maximum, so yes there is a little over-estimation there. On the flip side, when raiding with a holy priest often a small amount of your DA will be attributed to them, which is just odd, possibly resulting from simultaneous heals.

While not perfect these methods are great for judging where you stand (finally) in relation to the pure-heal classes/specs.

On the issue of sniping - yes a 7.5-8k shield + 1.5-2k heal on a GCD is the ultimate sniper weapon, but it is also the ultimate savior weapon. If you have more healing than needed on an encounter, someone is going to get sniped regardless, so it basically boils down to this: are you happy preventing a ton of dmg for a smooth kill (and maybe conqueror achievements?) by "padding" your numbers with extra shields? This is my experience on normal Ulduar 25 clears, and I must say it is rather nice racking up those clean 1-shots while on the road to the encounters you are currently working on. Our guild isn't on hard modes yet, but from what most people here say you have to fine tune your raid to the point where every healer is giving his/her best, so I'd reserve judgment on the true pecking-order of the hps/healing-meter for those in the top guilds. But as a Disc Priest in a family guild that also does progression, I'm very happy with Disc's numbers and performance post-3.1, and pleased that I can at long last see a Disc Priest listed towards the top of a report

Considering I only started shielding the raid within 25 seconds of her screech cooldown, and it hits each person for more than the shield can absorb, then yes, there is little to no overheal on PWS. And I rarely use renew, I probably cast it once or twice at most each week, it's incredibly useless for disc and you can use that gcd for something else.

Honestly though, I don't see how shielding the raid is so bad to some here when extending someone's health pool is far far better than waiting for a health drop and then healing that. It's a great tool, just sucky there's not much time to do it on hard modes. Heal sniping to me is something that has no positive impact on the raid, while pre-shielding has loads of pros.

And my post wasn't a "look at me, I'm pro because I can win meters", it was meant to compare with other disc priests and perhaps to show that disc isn't really low on output as most seem to think or want to believe. Is disc overpowered on some bosses? Of course, but then again, there are bosses which other healers are overpowered on too. And looking at the meters I posted, you'd see that even though I'm high on healing on some bosses, I'm low on others as well, but on the overall boss healing list, every healer is roughly having the same effective output- I'm only something like .2% ahead of the holy priest, which I take as a sign that despte the whine on the official forums, healers are a lot more balanced than they used to be.

Personally I think if a disc priest isn't shielding the raid when there's time, then that's slacking to the point of not doing his job, and the healers who whine about not getting damage to heal are just worried about their own place on the meters rather than what's good for the raid really. Pre-shielding to extend health has way more pros than just letting people take damage and potentially die.

Like I said earlier, I was the only disc priest in the raid. The only other absorbs come from sacred shield, and whatever little PWS that the other holy priest casts, which don't add up to much. I find WOL pretty accurate in general.

If you looked at the rest of the logs I posted, and actually read my post, you'd see that on fights with consistent raid damage, and large tank damage, JoL is winning the meters, right before aoe healers. On Iron Council, I was last.

And nowhere have I said that just looking at meters is a good way to judge healers. Judging from some of the replies here, it really seems like people just read the first line and then think "oh shit, my job is going to be taken". Relax and actually bother to read the rest -.-





And I didn't say I'd be the only healer on Thorim, was talking about Hodir.

I didn't mean to accuse you of bragging. But it does look to me from that log like the reason you are winning meters is simply because there is so much overheal going on. Pre-shielding the raid to the extent possible when AoE is incoming is a great use of your time and is a great reason to bring a disc priest (especially with the rapture bug). But looking at the log, it looks like the reason a discipline priest is coming out on top is because there is just a lot of healing in general and everyone is overhealing a lot.

Ultimately, though, if the point is that Discipline priests can pull their weight in terms of raw output, then I'm not sure whether being #1 proves that point, but this data and data I've seen from other raids all agree that discipline priests are not low output healers. I probably got myself too involved in minutia here.