Stat Weights for MoP Level 90

[Revulva]

I have updated my spreadsheet I was using to calculate Holy/Disc stat weights on AskMrRobot.

https://docs.google.com/spreadsheet/ccc ... Mmc#gid=16

Obviously it is still going to be a little raw... as I don't have all the information yet. I calculated some theoretical stat weights that should work pretty well for now, and then I will revise them and add multiple presets to the site based on actual raid parse information.

It's hard to use the spreadsheet... but, basically it works backwards from what happens in a fight. Enter in all the spells, hits, ticks, mana gains, etc. and then it tells you which stats will increase your HPM (Healing per Mana or Total Output) and HPS (Healing per Second) the most.

Given the data you input into it, the weights it spits out will necessarily increase your performance if you do the exact same thing the next fight (i.e., it is internally consistent). The "subjective" element to the weights is whether to care more about HPM or HPS. Both are important at different times. Right now I just average the two... which may or may not be the best solution.

In Cata, I sort of just followed the trends for stat weights, and had my calculated ones on the site as an alternate. In MoP, I am going with my calculated weights as the defaults, because I feel confident in the technique being used. What I'd like to get the community involved in is deciding what sort of spell breakdown is most realistic/optimal for Level 90 raiding. I could easily see different weights for 10/25 and AoE/Tank Healing scenarios.

Here are the priorities I am going to suggest (the actual numbers are on the "Theoretical Stat Weights" tab in the spreadsheet, in bold):
Discipline:
Spirit > Haste (12.5%) > Crit > Haste > Mastery

Holy:
Spirit > Haste > Crit > Mastery

Some interesting notes:
1.) For Holy, the choice of spells has much less impact on stat weights than Disc.
2.) Mastery is pretty weak for both specs, although better for Disc - especially if you can use a substantial amount of PW:Shield. I'm not really recommending any mastery stacking at this point.
3.) I think we'll need to stack Spirit at least for the first tier of content, so I'm weighting it high.
4.) Glyph of Renew. If you are going Chakra:Serenity, I think this will be good. If you are using it more for raid healing, you get better healing by going non-glyphed and having 5 ticks/cast. Getting 4 ticks/cast with the Glyph might be tough to achieve if you are stacking spirit, and it does the same amount of total healing.

Some questions:
1.) Anyone know how Halo works? Right now the spreadsheet just assumes max heal on everyone... which is ridiculous and kinda creates some silly results. I'll diminish it by a realistic factor - probably need some actual raid parses to do that, though.
2.) Getting 12.5% Haste is extremely valuable if you are going to use Renew to any significant degree. But, as Disc, you could just cast Renew after PW:Shield with Borrowed Time up and always get the extra ticks. I think it may be realistic to consider using this casting pattern and do away with the Haste "soft cap". Getting the second tick of Renew isn't really going to work in Tier 14 unless you stack a lot of Haste, even with Borrowed Time. As Holy, I think Haste is the best stat, so I don't really worry about the soft cap since I'm stacking haste anyway.

[Ashleycakez]

2.) Mastery is pretty weak for both specs, although better for Disc - especially if you can use a substantial amount of PW:Shield. I'm not really recommending any mastery stacking at this point.

Confused. I thought Mastery is excellent for Disc due to Spirit Shell?

Also I don't think Holy's Mastery is weak, it provides HPM which is much needed in the first tier of content.

For both specs this seems to go against what Derevka presented in his guides.
Holy: viewtopic.php?f=20&t=1613#stats
Disc: viewtopic.php?f=20&t=1837#stats

[Veiled]

Hi Revulva.. Derevka will likely chime in here in a bit. I'll be on vacation for a bit this coming week, but anything we can do to help Mr. Robot, just let us know. <3

[derevka]

Im far too tired to even look at this now, but it seems you've completely ignored spirit shell... which is a MAJOR game changer. Which calculates as:
Non-POH Shell = Average Heal * (1 + Mastery %) * (1 + Crit %) * (1 + (Crit % * 0.30))
POH Shell = Average Heal * (1 + Mastery %) * (1 + Crit %) * 1.30

And your seem to be favoring spirit a bit too much for my liking... considering you're giving it a discrete and rather generous HPS weight. This is pretty contrary to all the crunching I've done.

[Revulva]

I will add in some calculations for Spirit Shell and let you know what I find. Totally forgot about that.

I think it is being over-estimated, though. Basically, what you will see is that for 25% of the time, your Heal, FH, GH, and PoH gain a bonus equal to your mastery instead of a bonus equal to Crit Damage*30%*Mastery (Divine Aegis). It isn't clear to me just by looking at that whether or not it will definitely bump up the weight on Mastery or not.

Quick Napkin Math Example for PoH:
Say it heals 5 targets for 30k each, total of 150k and you have 50% mastery.

Divine Aegis: 300k*30%*1.5 = 135k Shield + 300k Healing.
Spirit Shell: 300k*1.5 = 450k Shield.

That is a 3.45% increase in total output 25% of the time, or a total of 0.86% increased throughput.

The effect on single target heals will definitely be greater from Spirit Shell:
Greater Heal for 100k, 50% mastery, 20% crit

Divine Aegis: 100k*2*30%*1.5*20% = 18000 Shield + 100k heal, on average, or, 118,000 average throughput.
Spirit Shell: 100k*2*20%+100k*80%*1.5 = 180000 average Shield 25% of the time, 120000 average heal 75% of the time = 135000 total average throughput.

That is a 14.4% increase in throughput.

So, Spirit Shell really doesn't do anything for PoH, but makes a nice difference for single-target healing. Based on your spell usage, that could put mastery up near the other stats, maybe even higher, depending on how you are healing - it is not clear-cut at all, though. This, of course, assumes you get max up-time on Spirit Shell. Especially in small groups, this won't happen - sometimes you actually need to heal, not shield.

The point of the spreadsheet is to actually figure these things out mathematically instead of just guessing at them. Once I add Spirit Shell in, we can figure out exactly what kind of spell breakdown you would need to make mastery better than Crit or Haste.

My calculations still show the HPM from EoL as being weak. It is obvious that the HPS is low, but the HPM is also extremely underwhelming. If anyone sees any other omissions, please let me know. The main goal is to have as accurate a model of the class as possible to do theory with. I made it in google docs so that anyone who wants can use it, mess with it, whatever. I'm not particularly concerned with ownership, just correct theory.

[Revulva]

I added Spirit Shell to the calculations. As I expected, it didn't really change the results significantly - I still think Mastery stacking is not going to be most optimal from a HPM or HPS viewpoint as Disc. Now, there is the consideration of the utility of putting huge pre-shields on lots of people with PoH, but that isn't something we can quantify.

Thinking about it in broad terms, this makes sense. Spirit Shell is only going to be up 25% of the time at the absolute best. So, Mastery is still only going to affect Heal, FH, GH, and PoH to some fractional degree, where Haste affects them all fully, and Crit does as well (with the added bonus of making Divine Aegis proc more, and therefore your mastery affects these spells more).

As far as Spirit is concerned, I think the best way to look at Spirit for healers is similar to the Hit Cap for DPS classes. It has very high value until you have enough, then it has no further value. Once we get into MoP raiding, we can find the "comfortable" value for Spirit (and each player will probably have to adjust this for themselves) and then set a cap on it in the stat weights. The optimizer on AMR allows for Spirit caps. This is why you will always see a high pre-cap weight on Spirit in my theory. Saying... Haste > Spirit will never allow an optimizer to get you to a specific value of Spirit which is necessary. You really need to weight it above other secondary stats until you get as much as you need.

[derevka]

I'll agree that its impossible to easily quantitatively assign a value to absorbs>raw healing; heck or even the typical fat that the overheal of raw healing is substantially more than the 'overheal' that happens with absorbs... but that has always been the trick with bridging TC and live results for Disc. There's not much you/we can do about that except go with more empirical results.

Regarding SShell's uptime, how are you accounting for the tiered approach of hastes interaction since you have to complete the cast for Shell to apply? Generally you'll have 6 or 7 major casts (2.5 sec baseline) during SShells uptime... haste doesnt increase shell until you break through a full cast during shells uptime.

Say it heals 5 targets for 30k each, total of 150k and you have 50% mastery.

Divine Aegis: 300k*30%*1.5 = 135k Shield + 300k Healing.
Spirit Shell: 300k*1.5 = 450k Shield.

Why'd you assume all POH targets crit? That completely makes the crit scaling for SShell moot, since its guaranteed scaling while on POH/DA it isnt; which then multiplicatively improves Mastery's scaling.

Holy you're assuming baseline no interaction with Lightwell/spring or either holy word from your results; which surprises me some. That said, of course, only one of those 4 interacts with mastery at all. As holy a 5% total output contribution of PWS seems overstated for a baseline/default spell breakdown, while POM seems understated... particularly since Divine Insight is what most holy priests should be taking if in a raid healing role, which unfortunately has 0 impact from haste aside from GCD minimization, and increases the amount of POM healing done just by nature of not triggering the CD.

This has always been my biggest "issue" (if we can even call it that) with simming results, and what I always have said: your sim results should be certain that the spell breakdown matches YOUR spell breakdown. (eg. is spirit shell X% of your output, is COH X%, etc... then it should represent similar percentages in sim numbers) Which is why i always wind up making my own Spreadsheet, and simcraft results with an actionlist that mimmicks MY spellbreakdown from WOLs for the bosses I am analyzing.

To answer your question, regarding Halo--- I dont understand the question. How it works? In what capacity? It captures targets at time of cast, and then layers in DR based on that snapshot. That said, its almost a fools errand to chase halo too much: Halo is less likely to be the 'default choice' over cascade in most encounters. Halo and Divine Star have specific niche fights where they time well with boss abilities, while Cascade is a better choice for most fights giving the CD alignment and mana cost.

I would, however, strongly recommend making your statement of "The "subjective" element to the weights is whether to care more about HPM or HPS. Both are important at different times." clear in the results. T14, HPM is going to be king since regen is going to be the real name of the game in T14. I think people are going to be in for a shocker. In all my raid testing, I wound up running in the 10-12k spirit range (scaled); granted I have a tendency to be a fairly aggressive healer. Of course some of this was done while they still fiddled with Solace/MB's values.

[Revulva]

That napkin math for spirit shell was just done quickly at the time. for PoH, I didn't assume any crits or non-crits - I just threw out a number. Since DA procs on all PoH hits, I didn't need to include a crit % for that calculation. PoH with and without Spirit Shell scale the same with crit.

The implementation of it in the spreadsheet uses a factor based on uptime, which should be a pretty accurate reflection of its contribution to HPS and HPM. (I assumed 75% of the max possible uptime during a fight, since you won't realistically use it on CD - sometimes you need to heal, not shield.) The question that I am trying to answer is... which stat improves output the most? For Heal/FH/GH/PoH - Crit improves HPM and HPS more than Mastery. Try it out for yourself in the spreadsheet.

The question about Halo was regarding how much the heal diminishes for targets closer and further than 25 yards. I'm not making a judgement on which spell priests should use, I just want to model them all realistically.

The "theoretical stat weights" that I calculated are just something to get us started in MoP before actual empirical data is available. This whole method for calculating stat weights is really based on taking observed results and figuring out what is the best way to improve the HPM or HPS. Every spell is modeled. I didn't include spells that are typically lower total healing % in the theoretical calculations. The cooldowns on Lightwell and PoM stop them from being huge portions of your total spells cast and healing done.

So, I wouldn't get too hung up on these theoretical weights. I just did a sampling of the spells to get a rough priority for each spec to start MoP with. I think the priority will end up being optimal, but the weights will change depending on spell selection. If we see Disc starting to use a lot of PW:Shield again, Mastery would jump above Crit.

Once raid parses are available, the plan is to take common healing patterns and analyze them. In Cata, I did this for 25 man raids - looking at dozens of the top raid parses with priests in them. A couple of different healing patterns emerged and I was able to calculate which stats would maximize those healing styles. The results were interesting because it usually wasn't what the priest was geared for.

I would be interested to see your calculations that lead you to the stat priority that you are currently suggesting. I'm a little surprised that your initial reaction to this spreadsheet is a bit defensive... seems like the perfect theorycraft tool for folks here to make use of. It's NOT a simulation. Simulation does not work for healers. By working backwards from the results of a fight, this model is able to take into account a lot of the varied factors that affect healing. If there were a better model for healing out there, trust me, I'd rather use that than spend time making my own ;)

Any thoughts on the idea of ignoring the renew 12.5% haste threshold on discipline because Borrowed Time gives it to them for free? Seems like Disc. never raid-heals with renew, so it would be realistic to suggest refreshing it right after pw:s most of the time.

[derevka]

That napkin math for spirit shell was just done quickly at the time. for PoH, I didn't assume any crits or non-crits - I just threw out a number.

Ah Gotcha. However, except you did assume a crit it seems. You said 150k, and then used 300k in the calculation. I assumed you just doubled it bc of a crit assumption-- maybe I missed something. That said, ignoring crit when calculating spirit shell underplays the stat (and its multiplicative effect on Mastery for spirit shell) and skips a step in the formula's calc. Take a look at the calc formula again, it scales directly with crit %, guaranteed on SShell. You might have just missed it, which can affect the result.

The cooldowns on Lightwell and PoM stop them from being huge portions of your total spells cast and healing done.

Unfortunately, that is simply untrue in practice. Read: Divine Insight, Lightspring for holy; which actually dramatically affect the output of the spells. DI, particularly. Lightspring, shines as a top spell in a LOT of lvl 90 raids.

Disc and renew changed dramatically with 5.0.X despite not having Rapid Renewal as a baseline ability renew will be much more regular in Disc's toolkit; particularly given the mana efficiency and the new mana game at 90. Outside of leveraging BT, you'll see some fights with more regular casts of renew, so I would cautiously suggest the 12.5 provided you're casting Renew.

Halo Calcs as follows:
0.5 * (1.01) ^ (-1 * (((Distance - 25) /2) ^4)) + 0.1 + 0.015 * Distance

http://us.battle.net/wow/en/forum/topic ... ?page=5#97

[Revulva]

Regarding the napkin math on PoH and Spirit Shell... I'm not even sure we're talking about the same thing anymore ;)

I implemented Spirit Shell in the spreadsheet this morning and it takes into account everything: crits, non-crits, uptime, etc. You can see how it affects the scaling of different spells if you mess around with stuff. I didn't ignore Crit when calculating spirit shell... The point I was making is that when SS is up, you don't get DA procs. So, what I was really looking at, as far as mastery is concerned, is if you get a lot more out of mastery when using PoH because of SS. The answer is: not really. DA procs off every use of PoH, not just crits. So, the calculation is different for that spell than the others. How often it crits doesn't really matter for that comparison. I was comparing:
Average Heal + DA proc to Average Shield with SS. That 300,000 base number I used for PoH can be treated as the average heal (which includes crits).

I did calculate theoretical scaling for Prayer of Mending in my initial weights on the spreadsheet.

I didn't do that with Lightwell because, in a 300 second fight, you can only ever put down 1 before the fight, 1 at 3 minutes in, for a total of 30 charges, 90 ticks. That can do a lot of healing for sure. I added it in to the theoretical weights... but it didn't change anything. Lightwell doesn't scale at all with mastery - since it doesn't proc Echo of Light (is that correct, it was in cata), it doesn't scale at all with haste - since the tick period doesn't scale with haste. So really, it just makes mastery an even weaker-looking stat overall and reduces the value of haste somewhat, but not to where crit is better for holy.

So I guess I'd need my two assumptions from Cata re-affirmed for MoP:
Lightwell ticks don't proc Echo of Light and they don't benefit from your Haste.

Regarding Renew and Disc... I guess I'd wonder if you cast it often enough outside of BT in order to warrant getting the extra haste rating at the expense of more Crit, which is a stronger stat overall. This would probably be a personal healing style issue.