You also have to consider what you gain from dropping Divine Fury. Spell Warding, Improved Renew, Desperate Prayer (to name some possible candidates) are not very compelling talents either, so I'd rather keep Divine Fury for the occasional Greater Heal and Smite.

I tend to save penance for spikes (or unlucky avoidance series), and use mainly flash / PW:S / POM in my rotation.
When there is a need of more hps (tank has lost 10k HP or more), I use penance, and gh if needed.
Most of the case, I'm either alone or with a holy paladin on tanks.

So in short there are three types of disc assignments :
- Tank-centric : shield tank, PoM tank, Penance on CD + GHeal/FHeal precast, this is more of a solo healer on tank raid assignment.
- Tank-reactive / raid-preventive : keep tank shielded (maybe renewed ymmv), PoM raid/tank appropriately, bubble spam on raid to keep BT up and smooth incoming raid dmg, keep Penance off cooldown to react to spike with hasted Penance + GHeal. This is ideal when paired on the tank with a druid or a FH-spamming paladin.
- Raid heal : while not as good as druids and holy priests, we can achieve some really good results if we manage to keep PoM on cd, spam shield before heavy raid damage, shield / hasted PoH during raid damage. The downside is that we are very dependant on good group positionning for the PoH portion.

Of those 3 different "typical" disc priest assignments, I consider number 2 - reactive healing on the tank(s) while bubblespamming and crosshealing the raid - to be the most effective on most encounters, by far. Having a disc priest smooth out the damage spikes on the tank and the raid, is our best contribution to most healing compositions. With "conventional" healing on the tank and raid, you'd need to dedicate vastly more raw healing power to prevent spike damage kills, but with a disc priest acting reactively on the tanks and the raid, you free up a good portion of that "extra" healing.

I see our role being more that of controlling the damage inputs and stabilising people so the other healers become more effective with what they do best - which is putting out the raw healing. It is my estimation that a good disc priest paired with effective raid healers can lower the amount of healers needed for most fights. This is why I actively pursue a more or less "carte blanche" healing assignment for myself. I'm well aware that my strong point is not raidhealing, and that I can more or less keep up with paladins on MT healing spam, but neither of those assignments make the best use of us. As of 3.1, with no cd on shields and targettable poh's, confining us to the "single target healer" stereotype of old, is just silly.

Hi, I've just completed a full Ulduar gear spreadsheet and I want to know if the current stat weights are still valid. Those weights that I know of are from The Doctor's original post in this thread. Please let let me know if these are still viable and if not, what the new stat weights are.... and I'm not so great at understanding equations so a boiled down version like the one I quoted would be most appreciated.
Thanks!

The trouble is, your raid will need the that portion of "extra" healing anyway, at least on the tank. Because most bosses use their "i kill you" attacks (Sartharions flame breath, Steelbreakers fusion punch, etc) more frequently then 15 seconds apart, this whole "smoothing out damage spikes on the tank" business is a fallacy. Sure you can smooth out a damage spike. But the next one will need to be healed the usual way: by raw healing output.

In that case, having enough crit to proc Divine Aegis off of every penance use is ideal. With enough crit, Penance has almost a guaranteed chance to crit with the 3 charges. So if PW:S won't do it, Divine Aegis will mitigate at least a small % of the damage. It might be enough to prevent the tank from dying.

I agree with this assessment of Disc synthesis. To me relegating a Disc priest to a standard "tank healer" role will limit the overall effectiveness of said priest in a 25 man environment. Honestly, Disc healing is insanely powerful in the hands of a player who possesses great healing instincts and understands the ebb and flow of particular boss fights. The ability to adapt to different healing demands is why I am so in love with Disc atm.

My personal healing habits in Ulduar literally change from boss to boss. Some fights my absorb hps outweighs my raw healing depending on the situation. On the whole, however, I have a few standard behaviors:

- Always keep a PW:S on yourself - besides the protection it is also 2% more mana for you when eaten
- Keep Weakened Soul and Renew on the MT for the extra healing benefit
- Cast PoM every CD on the MT - this keeps it bouncing in the melee and also will proc an extra DA shield on a crit
- Cast Penance on the MT at least every 15s - as Forces noted above, in a 25 you almost always see a crit somewhere in those 3 pulses, which means more DA shielding, but also a proc of Inspiration, and of course a triple Grace stack
- If you find yourself with nothing to do for a few seconds, cast a PW:S on a melee (ret pallies are a great choice), if for no other reason than to keep your BT buff up (and the T8 4pc bonus if you are lucky enough to have it!)

These things you can do every fight IMO no matter what your greater overall role is in the raid. With this behavior, and other healers tossing occasional HoTs/small heals to the MT, a single dedicated HL pally MT healer is typically enough. Again, it is all about awareness with Disc and being flexible enough to let your other healers shine in what they do best.

From my log the other night:

Razorscale Fireball hit Sthenno fro 568 fire damage (750 resisted) (5980 absorbed).

I had renewed hope up. The thing that struck me about this is that 750 resisted means that total hit must have been for between 7496 and 7505 damage, since resists come in clumps of 10%. But 568/.97 = 586 damage that I would have taken without renewed hope. That means the total damage would have been 586 + 750 + 5980 = 7316, which would not give the correct resist amount. However, if you also divide 5980 by .97 you get 586 + 6165 + 750 = 7501. I confirmed this with several other lines in the log: When you have renewed hope up, the shield is applied *before* the 3% reduction, but the amount listed as absorbed is then reduced by 3%.

For example, if you put a 6000 point shield on someone and they get hit for 7000, the 6k will be subtracted from the 7k, then the 3% damage reduction will be applied meaning they will take 970 damage, and it will say that only 5820 was absorbed, though the entire shield will be consumed.

The results were different for different damage reduction effects. Icebound fortitude was applied to the damage before the shield, as was one of Blade Barrier and Frost Presence (as they are both 5%, I couldn't tell which was before and which was after). The only numbers I could really confirm this on with the logs I had were magic damage that was partially absorbed, since the absorb number gives you a fix on what the total damage would have been without the mitigation, so I can't say whether this related to magic damage specifically: maybe swings use a different mechanic. Anyway, it looks like some damage prevention effects (notably Renewed Hope) are also shield mitigation effects at the moment at least.

It depends what you mean by "valid", but the short answer is yes. There have not been any huge changes. All "boiled down" stat weights make several assumptions which are pretty much guaranteed to be untrue but hopefully end up with values which are close enough. I would say they are 95% likely to be within about 25% of the true value to you. If that is not good enough, then you will have to get into the details of how stat weights are derived.

Your math is right, but it looks like your conclusion is flawed. It's easier to see what is happening if you approach it from the front, the way the server would calculate the damage. The question we're trying to address is: which of the following three models is used to resolve a simultaneous resist, absorb, and damage reduction?

1. Damage*Reduction - Resist - Absorb
2. (Damage - Resist)*Reduction - Absorb
3. (Damage - Resist - Absorb)*Reduction
4. (Damage - Absorb)*Reduction - Resist (this makes no sense, but it's possible)

Normally the problem with doing this analysis forwards is that we don't know how much base damage an attack has, because most attacks have a range. Here, however, we are in business because we know that Razorscale's fireballs do exactly 7500 damage, every time. Plugging this in, along with the given numbers, shows that the correct model is 2 as you would expect. This actually agrees with your result; you computed that Damage = (FinalDamage + Absorb)/Reduction + Resist, which is just a rearrangement of model 2 and shows that the absorption is being applied AFTER the damage reduction. I think you then got confused and interpreted this result incorrectly, or something. I would be interested to see your data for testing other reduction and absorption effects, especially if they behave differently.

If the reported Absorbed number is decreased by Reduction from what was actually absorbed, your example indicates model 3, which as you say would be sort of unfortunate for Discipline priests. I see no reason for it to lie, but we can test it. Unfortunately it's relatively difficult to construct an absorption effect of a known value, but it is possible. Consider the following lines from a recent Heroic Ignis kill:

Zagral suffers 4516 Fire damage from nil Slag Pot. (1200 Resisted)
Mygrane Glyph of Power Word: Shield heals Zagral for 1292.
Zagral gains Power Word: Shield.
nil Slag Pot was absorbed by Zagral for a moment.
Zagral suffers 2852 Fire damage from nil Slag Pot. (600 Resisted) (2229 Absorbed)
Mygrane Prayer of Mending heals Zagral for 4962. (Critical)
Zagral gains Divine Aegis.
Zagral gains 1576 health from Cruellia Divine Storm.
Zagral suffers 2463 Fire damage from nil Slag Pot. (1800 Resisted) (1488 Absorbed)

I used myself because I know I don't have any other reduction effects from talents or resilience, and I'm not sure about other raiders. We'll take this line by line. The first one was the first tick of the slag pot, so I didn't have a shield up. It did exactly 6k base damage (it always does, every single time). 1200 resist means I took 4800 damage. This was reduced by 3% by Renewed Hope, which gives 4656. This confused me for a while until I realized that Blessing of Sanctuary now stacks (multiplicatively) with Renewed Hope, so this is reduced by another 3% for 4516.32, rounded down. So our current model for a resisted hit is (6000 - resist)*.97*.97. As far as why I had Blessing of Sanctuary, we had a lot of paladins in this raid, everyone got 3 blessings, and I'm not likely to want Wisdom, am I?

Next, Mygrane shields me for about 6200. It's relatively difficult to work out the actual value of the shield because I had a number of healing-increased effects such as Tree of Life aura and Improved Devotion Aura, which affect the heal from Glyph of PW:S. Unfortunately this entirely absorbed the next tick of the Slag Pot, so we can't use it to resolve our issue in any case. On the 2nd hit, I took 5400 after resist, *.97*.97 gives 5080.86, rounded up to 5081 which is exactly 2852+2229. This confirms the previous analysis from your example: either model 2 is correct, or model 3 is combined with some combat log shenanigans. Incidentally it almost certainly means that I got an 1800 resist on the tick we can't see.

This hit also procs Prayer of Mending, which crits giving me a Divine Aegis for 4962*.30 = 1488 or 1489 depending on rounding. (The rounding rules seem pretty random to me. I suspect this was actually 4961.5 rounded up, but the 30% is taken before rounding.) The Divine Storm doesn't affect this story (but thanks anyway, Cruellia). In the next hit, I'm affected by 4200 fire damage, reduced twice again for 3951.78, which rounds down to 2463+1488. As we can see, the DA is in fact absorbing for full value. I think this conclusively proves that the correct model is 2, not 3 with the combat log lying. I sincerely hope that the fact that Slag Pot is a DoT and Fireball isn't does not cause the combat log to start lying; that would make NO sense.

Good news for priests! You can trust your combat log, and Renewed Hope stacks with Blessing of Sanctuary. Previous testing indicated that BoS and old Grace did not stack, so I'm forced to assume it broke when Grace and RH were reimplemented.

Hmm... I agree with this general approach to the problem and actually posted on this earlier (check out this post and this post, which are both kind of long but definitely related to what you are doing). However, I ended up with a significantly higher mp5 cost per point of haste:

I did not explain the details of how I got 0.5 Mp5, but the method I used to calculate the increased mana cost is based on the following:



Since changing haste (while holding everything else constant) keeps HPM constant, the change in Mp5 from changing haste is directly proportional to the change in HPS:



I used flash heal spam and ended up with roughly -0.5 mp5. Looking over your calculations, I think your units are mana per second, not mana per 5. That could explain the difference.

Also, I noticed you used a 0.5 crit multiplier. With divine aegis, it is common to use 0.95 in order to account for the added DA shield in throughput. That would increase the relative value of crit by 90% (so instead of 0.27, you'd have 0.51).

This might go better in the "Simple Q's/Simple A's" section, but I'm not sure if this is specific to the disc weighting calculation or not. I've just switched over from Rawr to the spreadsheet and I've been trying to work out the math required to get to the combined stats above (mainly because I don't plan to use a 60/40 split but I want to make sure that the split I *do* use is calculated correctly).

Here's the approach I took, (which didn't seem to work correctly). I'm hoping someone can find the flaws in my calcs.

First, I converted each set to be Int-based (since it's the only stat in common). I divided each value in each set by its value for Int. That gave me:

Thoughput stats...

• Intelligence - 1.0
• Spell Power - 6.25
• Crit Rating - 3.00
• Haste Rating - 3.69

Longevity stats...

• 1 Intelligence = 1.0
• 1 Spirit = 0.45
• 1 MP5 = 1.35

So now, I've got apples to apples, right? At this point, it seems like I could just multiply the weighting by the stats and combine them, right? IOW, for a 60/40 split, sp would work out to be: (60% * 6.25) + (40% * 0) = 3.75

This converts to:

• Intelligence - 1.0
• Spirit - 0.18
• Haste Rating - 2.21
• Crit Ratng - 1.80
• Spell Power - 3.75
• Mp5 - 0.54

Then, to make it SP-based, divide each one by SP:

• Intelligence - 0.26
• Spirit - 0.05
• Haste Rating - 0.59
• Crit Ratng - 0.48
• Spell Power - 1.0
• Mp5 - 0.14

Obviously, this doesn't work -- what am I doing wrong here?

The problem is that a 60:40 ratio for throughput / endurance (based on Int as the common factor) is completely arbitrary. Try again with 80:20, or 10:90, or 1:99, and you'll end up with quite different results. Who is to say which is correct? More likely, each would be correct for different roles in different fights. 99:1 might be reasonable for tank healing on Hodir hard-mode, where it's all but impossible to go OOM.

There's no easy way to relate throughput to endurance.

Personally, I'm using roughly 15:85 ratio using your baseline calculations of throughput and endurance. 1 SP : 1.3 Int. I don't think I have any mathematical basis for this; it just *feels* right for most fights.

No, I understand that. I guess I wasn't clear about what I meant. I agree that 60/40 is arbitrary. I personally go with something closer to 75/25, but that's not important for what I'm trying to figure out.

My problem is that I can't come up with any formula that allows me to reach the values given in the post, so I'm just wondering how those values were arrived at.

I've have been using a 50:50 statweight with a bumped up version of spirit (because spirit is better than that statweight with my intellect):

100 SP
100 MP5
74 Intellect
59 Haste
48 Crit
38 Spirit

This is more or less exactly what I do in Rawr. If you use the Custom Rotation in Rawr you can also specify what spells you are using.

First, thanks for all the responses. But, that's still not my question. I'll try one more time...

I'm not looking for assistance determining the throughput/regen value. My problem is that while the spreadsheet does a great job calculating the values for throughput AND the values for regen, there is no formula built in to provide the combined value, so I'm trying to calculate it manually. However, I can't figure out how to do that.

Basically, I'm looking for help with the math, not help determining the ratio.

Don't get me wrong -- I love rawr. However, I like the ability to tweak the stats. For example, I want to be able to tell the model to not put any value on haste, since I already have 7% from gear, which is well over the soft cap for disc when we consider borrowed time.

The weighted average would work, if you had absolute values or comparable ratios.

However, these ratios are not comparable.

Assume you use the Int-based ratios where the relative value of Int is 1.0 for both longevity and throughput. In that case, though you assign a value of 1.0 to each, 10 points of Int do not have the same absolute value regarding longevity and throughput.

This is why it doesn't work.

Several spells are still not capped even with full raidbuffs and 7% from gear as Disc with Borrowed Time. Penance and Prayer of Healing are the big ones. If you wish to Optimize your gear, you should simply put in a restraint in the Optimizer to not go beyond your wanted Haste value. Like, setting Haste % <= 7.

Beyond that, I'm not sure what to tell you. Personally I prefer having around 10% haste from gear, as it (combined with my latency) Guarantees I get to cast every GCD. And it benefits ProH/Penance, perhaps an occational GHeal, and sometimes, you may not even have 5% spell haste from totem or similar. (10 man.)

Fair enough. However, The Doctor somehow got to those 60/40 values -- I still can't figure out how.

I don't want to turn this into a rawr discussion, but wouldn't that make Haste be "bad" beyond 7% instead of just useless? I'll post something over on the rawr boards so I don't derail this.

It looks like he's implicitly assuming that 1SP = 1mp5 @ 50:50, and took it from there.

Your calculations using Int as the common factor assume that 1SP = 8.4375 mp5. To convert from your calculations to his, use an 11:89 ratio (and then adjust to his 60:40 if you like).