Elitist Jerks Discipline priest raiding compendium

02/13/09, 10:54 AM   #61
Caltiom
Von Kaiser

Human Priest

Eredar (EU)
Great to have a discipline-thread established, even if there are many similarities, it's good not always reading about coh-nerf

To add more mathematical formulas to this guide, i tried to determinate the exact coefficient of rapture returns with some testing.

Assuming the formula: Rapture_return = c * Max_Mana * Effective_Healing_done

Testing in Patch 3.0.9
At 20233 Mana:
 Healing done Rapture_return 3791 168 3712 165 3675 162 3903 173 5610 249 3934 174 7825 347 7829 346 7869 349 7839 349

And a 12'204 Crit Gheal returned 506 Mana, which ist 20233*0.025 rounded up. So the 2.5% Cap holds.

c=2.1915 * 10^(-6)

At 14880 Mana (half-naked)
 Healing done Rapture_return 2908 90 2734 88 2758 85 2886 89 4327 133 2836 87 5975 184 5915 182 8653 266

c=2.071 * 10^(-6)

So there is quite a difference for the 2 Mana_Levels. Maybe i'll investigate it further, using a full naked set if I have the time and the space in my inventory. Or maybe the assumption of the formula is incorrect

Interesting is the cap on rapture. rapture_return= c * max_mana * effective_healing_done <= 0.025 * max_mana

Solving for the healing_point where the maximum rapture_return is reached, we get heal= 0.025 / c . Using the above coefficient, we see that at 11'500-12'000 healing done the maximum rapture_return of 2.5% of max_mana is reached. So Crit Gheals may become more and more Mana-Inefficient as we gear up.

Last edited by Caltiom : 02/13/09 at 11:16 AM.

02/13/09, 12:40 PM   #62
The Not So Evil
Piston Honda

Night Elf Priest

Trollbane (EU)
 Originally Posted by Caltiom So Crit Gheals may become more and more Mana-Inefficient as we gear up.
This would mean that the cost of a GHeal increases as we gear up wouldn't it? But to the contrary, GHeal cost is constant, thus, as long as our Intellect, Spell Power and Crit Increases, GHeal becomes more and more efficient.

Rawr - Coder of HolyPriest (Healer) and ShadowPriest (DPS) Modules.

02/13/09, 1:41 PM   #63
Turrin
Glass Joe

Night Elf Priest

Medivh
 Originally Posted by Gor Didnt know where to write - here or in Healin Compendium - but as i am DISC priest - ill ask here. Renew vs FH - it was long discussed over this board. i was reading how FH is superior as compared to Renew - so wanted to test it myself. My char is not best equipped - we are just clearing 10Naxx, not 25instas yet. FH Glyphed, Renew glyphed, changed GH to Renew in talants. number based on last patchwerk stat FH cost is 690 mana and crit gives 6091 HP max per crit Renew costs 590 mana and gives 1594 HP per tic. one cast = 4 tics = 6376 HP Please tell me - what is wrong with my calculation - or WHY renew is worse than FH? Yes, i know that renew cannot crit - but it is not supposed to replace FH - but add smoothness to healin. It is may be better use GH in fights as Patchwerk - but in that case haste 10%+ unbuffed is required - otherwise its too much "jumping" in tank's HP. If it is wrong topic to post to - i am asking admin to move it to right one.
Nothing wrong with the caculations as far as I can tell. But even glyphed, you need to ask yourself why are you wasting a GCD on renew in a fight like patchwerk? When you are learning that encounter on 25s, it can hardly smooth out 10k+ hits per second on your hatefull tanks (probably more as they wont be geared). Ask yourself what else can you be casting durign that GCD that will heal more than 1.5k and I think you can come up with a better answer than renew. Until the tanks and healers are well geared, you should only be casting your best then your biggest heals.

Others have covered the reason why, but the chance to gain mana via rapture should by itself make you move renew off your most used key binds.

 02/13/09, 4:17 PM #64 Kraylessa Glass Joe     Krayleyna Draenei Shaman   Silver Hand Just to sum up the Renew vs. Flash Heal argument: - Renew cannot Crit (no DA, no increased throughput) - Renew does not stack Grace - Renew does not return mana via Rapture - Renew rarely ticks for it's full amount (does not tick when target is at full health, thus is essentially overheal) With Haste bringing your FH cast times down to barely over one second, Glyph reducing mana cost, crits proccing DA and therefore increasing the amount "healed", Grace stacking, and mana returned via Rapture from the spell itself AND the DA's it procs, it is just a superior spell in every way. The only time I ever use Renew is when I need instant cast spells (Maly's vortex, Heigan dance...). Pretty much any other time, FH is a far better use of my GCD. Last edited by Kraylessa : 02/13/09 at 4:33 PM.
 02/13/09, 6:29 PM #65 stoops417 Glass Joe     Knobody Undead Priest   Thunderlord First of all let me mention how glad i am to finally have a place to discuss Disc priest raiding. I was hoping to extend the discussion on general theory and practice of disc healing. Specifically, the useage of Flash Heal vs Greater Heal. I find myself when tank healing using Greater Heal between Pennance cd's based on the old holy spec habit of 1 GH > 2 FH. I'm wondering if most of you are following the same practice of using GH over FH when there is more than a 5-6k heal needed or if you just chain spam FH?
 02/13/09, 6:43 PM #66 The Not So Evil Piston Honda     Brundtland Night Elf Priest   Trollbane (EU) Sarth3D I spam Greater Heals, on Patchwerk I've found that rotating PW: Shield on all 3 tanks + flash healing between Penance to be more effective. Sapphiron I mix it up, as I know other healers can see what heals are incoming on people all the time, and use a mix of PWS, Penance, Flash Heal and Renew, all depending on situation. On Kel'Thuzad I actually end up Greater Healing a lot after the AoE Frostbolts. Call me oldschool, or maybe I just have nerves of steel, but I have quite a lot of trust in the slow hard hitting Greater Heal still, and I'm waiting for a decent 3rd Glyph, may it be Glyph of Greater Heal. Rawr - Coder of HolyPriest (Healer) and ShadowPriest (DPS) Modules. Get Your Rawr 2.3.x!
02/13/09, 7:39 PM   #67
atrinitydream
Von Kaiser

Pandaren Shaman

Alexstrasza
Some Math on Crit vs. Haste

I think a bit much of the discussion around crit vs. haste has boiled down to 'it's personal preference' without understanding the numbers that should drive that preference.

I'll state up front that my preference is for balanced gearing and believe that stacking either crit or haste too heavily decreases effectiveness. (As a general comment, I believe haste tends to be undervalued.) However, I've found some of the following numbers helpful in making decisions when picking gear upgrades:

(This is a work in progress; there's some stuff I'd like to include but don't have written out at the moment. Please feel free to correct my math if I've made any errors.)

Haste

Haste soft caps
Below numbers indicate haste rating required to cap the GCD to 1 second. Note that PI/Heroism + BT always cap haste. Haste Modifiers stack multiplicatively. Enlightenment is included.
 Haste Modifiers Talents + Buffs Only w/ PI (20%) w/ BT (25%) w/ Heroism (30%) 1405.29 624.57 468.43 324.30 3% [Moonkin/Ret Paladin] 1268.85 510.88 359.28 219.35 5% [Wrath of Air] 1182.22 438.69 289.98 152.71 3% + 5% 1052.29 330.40 186.03 52.76

I would consider the BT cap associated with your current level of raid buffs to be a 'minimum' level of haste, although that's certainly open to interpretation.

Efficiency

Haste for obvious reasons has a negative effect on rate of mana usage, but what is not widely recognized is that if you are able to add x amount of haste rating and consequently do less overhealing, the additional mana returned from Rapture can easily compensate for the loss in efficiency due to casting more.

At approximately mid-Naxx25 gear levels (2200 spell power, 30% crit, 200 haste) and average amounts of overheal (50% overheal on non-crits, 65% on crits, 25% of DAs lost due to overwriting or non-absorption), for FH adding 10 points of haste requires you do to .5% less overhealing.

This number is variable based on the amount of overhealing you do as well as your current gear level, etc., but the point is that haste does not have as great a negative impact on efficiency as is believed.

Furthermore, and perhaps more significantly, the amount of overheal needed to zero out the mana loss from haste decreases as crit, spell power, and particularly max mana increase (in other words, as gear improves).

Data is below:

```Overheal (Throughout): 50.0%, Crit Overheal 65.0%, Aegis Overheal: 25.0%

Spell Power: 2200.0, Crit: 30.0%, Max Mana: 25000.0
At 200.0 haste:
Flash MPS Cost: 464.184964928
Flash Rapture MPS: 97.8292567561
Difference (net loss): 366.355708172
At 210.0 haste:
Flash MPS Cost: 465.519213175
Flash Rapture MPS: 98.1104561144
Difference (net loss): 367.40875706
MPS Lost: 1.05304888811
Added Effective Healing to net zero: 0.568164864133%

Spell Power: 2500.0, Crit: 35.0%, Max Mana: 30000.0
Overheal: 50.0%, Crit Overheal 65.0%, Aegis Overheal: 25.0%
At 200.0 haste:
Flash MPS Cost: 464.184964928
Flash Rapture MPS: 128.560334937
Difference (net loss): 335.624629991
At 210.0 haste:
Flash MPS Cost: 465.519213175
Flash Rapture MPS: 128.929867374
Difference (net loss): 336.5893458
MPS Lost: 0.964715809115
Added Effective Healing to net zero: 0.399198539158%```
The relevant portions of the Python script (I hate Excel) used to generate this data are below (so those more knowledgeable than I can check my math). I left out basic calculations like those for cast time and base flash heal amount. Grace is NOT assumed to be up. NOTE: For clarity, '_add' refers to calculations when adding x amount of haste (10, in my sample runs above), which as they're identical for everything except the last four lines I've omitted to try to save space. I can provide the full script upon request.

```flash_total_healing = (noncrit_percent * flash_base_heal * noncrit_effective) + (crit_percent * flash_base_heal * 1.5 * crit_effective) + (crit_percent * flash_base_heal * 0.45 * da_effective)
flash_rapture_return = max_mana * (flash_total_healing / 11460.0) * 0.025

flash_cast_time = 1.5/(1.05 * (1.0 + (haste_rating / 3279.0)))
flash_mps_cost = flash_cost / flash_cast_time

flash_rapture_mps = flash_rapture_return / flash_cast_time

net_loss = flash_mps_cost - flash_rapture_mps

needed_flash_rapture_return = flash_cost - (flash_add_cast_time * net_loss)
needed_flash_total_healing = (needed_flash_rapture_return / (max_mana * 0.025)) * 11460.0
added_effective_percent = (((needed_flash_total_healing / flash_base_heal) - (crit_percent * 0.45 * da_effective)) - (noncrit_percent * noncrit_effective) - (crit_percent * 1.5 * crit_effective))/(noncrit_percent + (crit_percent * 1.5))```
Non-tangible Benefits

It also bears noting that haste increases ability (this is more noticeable in large quantities than small) to concentrate on more than one target at once as well as further supports Discipline's already strong ability to reactively, rather than exclusively proactively, heal.

Crit

Inspiration

Inspiration generally gets ignored when discussing crit vs. haste, but ability to keep Inspiration uptime on multiple targets is a significant factor in some fights and is one argument for crit 'scaling.' In a vacuum (pure single-target scenarios), crit percent is irrelevant due to the fact that it is more or less guaranteed to be up.

However, in scenarios where you can only spare between one and three casts every 15-20 seconds for a secondary target, having a high enough crit percent to make an Inspiration proc extremely likely in that time is valuable. (Penance can be very helpful for this type of healing if it is not immediately needed on the main target, as not only does it have 3x chance to proc Inspiration, but it applies a full stack of Grace.)

Basic probability rules can illustrate this pretty easily - going with the three-cast scenario:
- At 20% crit (80% * 80% * 80%) there is a 48.8% chance to apply Inspiration in three casts.
- At 30% crit (70% * 70% * 70%) there is a 65.7% chance to apply Inspiration in three casts.
- At 40% crit (60% * 60% * 60%) there is a 72.9% chance to apply Inspiration in three casts.

Divine Aegis

Though DA does not improve the throughput value of crit above (or even close to) spellpower or haste, it provides another shield (due to DA shields proccing based on total, not effective, healing done) and thus extends effective HP pool for topped-off targets. This effect is difficult to model mathematically, so it should just be noted that reducing crit too low makes DA unlikely to be up on the target, therefore devaluing one of the primary unique benefits of Discipline as a spec.

Last edited by atrinitydream : 02/13/09 at 7:48 PM. Reason: So it more closely resembles being intelligent.

 02/14/09, 1:56 AM #68 Sharajat Piston Honda   Sharajat Night Elf Priest   Scarlet Crusade Looking at it from a more meta perspective though, why would you assign disc to anything other than single target healing? In general, disc is a single target healing spec. Yes, for 10 mans, crit is plenty nice, but in 25 mans my gear is certainly past the point where additional crit is doing me much good, and I'm shifting my gearing away from it. Divine Aegis just has shitty synergy with... divine aegis. It's a really sad thing, but the more you get of it, the less you want it. I think that with any boss who is not hitting remarkably fast, Aegis beyond 25-30% crit (with weakend soul) is semi-worthless if we're single target healing. If they ever change things with Aegis, I'll revisit it, but at the moment I think I'm going to be regearing to not top 30% fully raid buffed with weakened soul, and I might scale even that level down. It really depends on what Ulduar has, lots of raid healing encounters might raise the value of crit a lot. Haste, the more heals you get, typically the more healing you do. It's pretty much a linear thing - more haste = more cast (until the various caps). Oddly, I think holy may actually scale better with crit than Disc. Last edited by Sharajat : 02/14/09 at 2:15 AM.
 02/14/09, 6:04 AM #69 deamosreapos Glass Joe   deamosreapos Night Elf Priest   Silvermoon (EU) Gear related I want to submit a list of gear which currently has my priority over the lsit alrdy shown (thus leaving parts already mentioned out) Neck: Life-Binder's Locket (more int more regen) Chest: Either: Robes of Mutation or: The Sanctum's Flowing Vestments Ring: Renewal of Life Sorry for the post it's not ment to undercut anyone, but I think these options should be mentioned, currently I am wearing arround 80% of the gear on my wishlist (See list of gear mentioned at the beginning of the post and add these last 3 items) But what I noticed in game, is even with the low ammount of haste on my gear, arround 6-7% (thats not raidbuffed just gear) I end up with 15-20% haste raidbuffed. And I already get the 1 second cap on my FH. (this is with either PI or heroism, PI I got arround 1.05 second cast on FH) About the int I got fully buffed, is arround 28k. Spellpower +- on 2200, with shaman totem arround 2400 So what I also like to see added are the elixirs that are great for disc: Elixir of Mighty Thoughts Spellpower Elixir
02/14/09, 9:53 AM   #70
Elimbras
Don Flamenco

Dwarf Priest

Eitrigg (EU)
 Originally Posted by Sharajat Looking at it from a more meta perspective though, why would you assign disc to anything other than single target healing? In general, disc is a single target healing spec. Yes, for 10 mans, crit is plenty nice, but in 25 mans my gear is certainly past the point where additional crit is doing me much good, and I'm shifting my gearing away from it.
There is some fights were you have several offtanks. In Razzavious for example, I'm responsible for the healing of the 2 OT, and I spend my free time on the raid. We have several other healers on the tanking add.
In such fights, I spend less than 1 third of my time on each tank I'm healing.

Same can be on Patchwerk. We don't do it, but I can imagine to have one disc priest swapping from one tank to another, shielding every one, and then keeping grace and inspiration on each.

Last point : again, we run as 2 disc. priest for the moment. That's not optimal, but we have 4 healing priest, and only 1 paladin. The other disc. priest is often on the MT. I tend to be on the OT (when there is one). For lots of fights, OT healing is not full-time. When I have free time, I help the raid healing with flash. High crit rate means that I have more chance to keep inspiration on the MT.

02/14/09, 1:44 PM   #71
atrinitydream
Von Kaiser

Pandaren Shaman

Alexstrasza
 Originally Posted by Sharajat Looking at it from a more meta perspective though, why would you assign disc to anything other than single target healing? In general, disc is a single target healing spec. Yes, for 10 mans, crit is plenty nice, but in 25 mans my gear is certainly past the point where additional crit is doing me much good, and I'm shifting my gearing away from it.
Tank healing (which disc does well) does not necessarily equate to single-target healing. In 25-man raiding, true single-tank encounters are the exception rather than the norm. The ability to cover more than one target with is one of Disc's strengths. Off the top of my head, I can think of any number of fights where being able to keep healing and/or mitigation on multiple targets is advantageous:

- Malygos: Healing raid spike damage in between the tank healing.
- Patchwerk: Tanks & soakers.
- Sarth+3: Main tank & drake/add tank.

And so on. Now, none of these encounters generally require careful multi-management of tank healing, but that's a function of the (non-)difficulty of the encounters, not the nature of Disc healing. I also believe that Blizzard intended Discipline healing to work in that way. Reactive heals are usually wasted if you're on a single target. Just faceroll GHeal; it's all you'll need. Plus, Disc has a number of innate advantages when used this way:

- Passive haste (rather than straight spellpower, for instance) via Enlightenment: More haste means multi-tasking is more feasible, plus makes reactive healing much more plausible.
- Penance: A fast, big, reactive heal on a CD is perfectly suited to this type of healing. It applies all three stacks of Grace at once and has an extremely high chance of applying Inspiration as well given the amount of time casting it.
- Borrowed Time: Makes it extremely easy to drop a shield on a secondary target given a break and lose little throughput on a primary target. (Plus it makes re-applying Inspiration and DA easier.)
- Divine Aegis: As pointed out, it doesn't stack well 'with itself,' that is, when being used on an isolated target. It overwrites too much in that scenario. But if you are able to spread out healing between two targets, for instance, the chance of both DAs being used increases dramatically.

EDIT:
 Originally Posted by Elimbras Same can be on Patchwerk. We don't do it, but I can imagine to have one disc priest swapping from one tank to another, shielding every one, and then keeping grace and inspiration on each.
This is exactly what I do on Patchwerk. Shields every CD on the offtanks (or MT if you have the time) eat a LOT of the damage - and keeping BT constantly up means that Penance can fill in a good amount of what the shields don't cover.

Last edited by atrinitydream : 02/14/09 at 1:50 PM.

02/14/09, 5:33 PM   #72
Sharajat
Piston Honda

Night Elf Priest

 Originally Posted by atrinitydream Tank healing (which disc does well) does not necessarily equate to single-target healing. In 25-man raiding, true single-tank encounters are the exception rather than the norm. The ability to cover more than one target with is one of Disc's strengths. Off the top of my head, I can think of any number of fights where being able to keep healing and/or mitigation on multiple targets is advantageous: - Malygos: Healing raid spike damage in between the tank healing. - Kel'thuzad: Add tanks & main tank. - Patchwerk: Tanks & soakers. - Sarth+3: Main tank & drake/add tank. And so on. Now, none of these encounters generally require careful multi-management of tank healing, but that's a function of the (non-)difficulty of the encounters, not the nature of Disc healing. I also believe that Blizzard intended Discipline healing to work in that way. Reactive heals are usually wasted if you're on a single target. Just faceroll GHeal; it's all you'll need. Plus, Disc has a number of innate advantages when used this way: - Passive haste (rather than straight spellpower, for instance) via Enlightenment: More haste means multi-tasking is more feasible, plus makes reactive healing much more plausible. - Penance: A fast, big, reactive heal on a CD is perfectly suited to this type of healing. It applies all three stacks of Grace at once and has an extremely high chance of applying Inspiration as well given the amount of time casting it. - Borrowed Time: Makes it extremely easy to drop a shield on a secondary target given a break and lose little throughput on a primary target. (Plus it makes re-applying Inspiration and DA easier.) - Divine Aegis: As pointed out, it doesn't stack well 'with itself,' that is, when being used on an isolated target. It overwrites too much in that scenario. But if you are able to spread out healing between two targets, for instance, the chance of both DAs being used increases dramatically. EDIT: This is exactly what I do on Patchwerk. Shields every CD on the offtanks (or MT if you have the time) eat a LOT of the damage - and keeping BT constantly up means that Penance can fill in a good amount of what the shields don't cover.
I'm currently rolling disc (as you can see), I'm just questioning the value of crit as a stat for us. Take the fights you talked about:

Malygos - Phase 1: assuming we're not under 12 seconds from a vortex, Aegis on the random damage targets seems... random. Yes, sometimes he hits the same guy twice. It happens. I'd say 80% of the time Aegis falls off without doing anything.

Phase 2: Enough random damage to enough targets that I could be really generous and say that only 50% of the time it falls off without doing anything.

KT - DEFINITELY agree. I've healed both tanks many times, Aegis shines here. That being said, its worthless on ice blocks.

Patchwerk - we tend to assign healers to tanks rather than have people jump about. I GUESS we could do it so that the disc priest is keeping a grace stack etc, but honestly... too much work for no reward. I toss shields about too - but shields don't crit.

Sarth+3 - not enough experience, may shine here.

I just can't see enough reason to gear for crit over haste from these examples. I'd rather have more heals than Aegis most of the time. It's just much better as a passive mitigation skill that is a bonus, rather than something that is proccing and overwriting itself constantly.

02/14/09, 8:14 PM   #73
atrinitydream
Von Kaiser

Pandaren Shaman

Alexstrasza
Ironically, I think you actually hit the crit nail on the head yourself:

 Originally Posted by Sharajat It's just much better as a passive mitigation skill that is a bonus, rather than something that is proccing and overwriting itself constantly.
I think this is exactly right. Crit(/DA) is not a good throughput stat, as such. It is, however, a good mitigation ability that can provide extra effective health for affected targets.

On Malygos: I think you've got the emphasis backwards. I would argue Aegis' value isn't on the random damage targets. It's value is on the tank (on which I'm assuming your primary focus is), giving you some extra mitigation time which you can then use to spot-heal the raid. Outside of fights like Sapphiron, DA is a terrible raid-healing talent.

My point in bringing up the examples I did isn't that crit (or haste) provides particular benefits for those fights, but that if you isolate the tank+tank or tank+raid healing scenarios in them, the talented value of both crit and haste (as well as where Discipline can bring something unique and valuable to the table) become much more apparent.

 02/14/09, 8:34 PM #74 Shaejin Glass Joe     Shaej Human Priest   Doomhammer The discussion of crit vs haste on different boss fights suggests to me that we should be preparing multiple gear sets for Ulduar. (Depending on the fight and your current assignment, for example, you might swap out 3 pieces of haste gear for 3 pieces of crit gear. This might include swapping out the exact same piece with one that has different gems/enchants). Question: How many different gear sets should we be preparing? I can see the following being situationally useful: 1. Crit set (with ~5-10% more crit than your normal gear) 2. Haste set (with ~5-10% more haste than your normal gear) 3. Endurance set (stacking intelligence, and perhaps secondary regen stats) 4. Holy set (for dual spec as holy) 5. +Hit set (for dual spec as shadow, or MC pulls) 6. Stamina set (I am not including resist sets in this list, as I expect Blizzard to provide level 80 patterns for them when the need arises).
 02/15/09, 2:28 AM #75 Imua Von Kaiser   Malama Dwarf Priest   Kalecgos Raid healing: It's not BAD to raid heal as disc, but you need to understand a few things first. If there's a localized AOE, CoH and CH are just plain better. *BUT* there are certain situations where disc can do decently enough as raid heals. If it's a fight like Sapphiron, where there is steady damage going out there, you should feel free to pop out a PW:S on pretty much anyone. Any caster will benefit from the uninterruption, and you get mana back from the shield being absorbed. So putting a PW:S on anyone is beneficial. Also, disc is TERRIFIC at saving the random person who's in danger of getting 2-shot (Malygos phase 2). A shield+Penance+FHeal can easily save some random in the raid. Again, though: those fights are isolated. There are situations where disc can do an adequate job raid healing, although it still doesn't have the tools like CoH. A targetable PoH in 3.1 may prove to be worthwhile, though.

 Elitist Jerks Discipline priest raiding compendium