Disc stat weights

[Noradin]

However the question remains: Is there a maximum value for PWS which cannot be exceeded in a single cast? You say that
And since the definition of 'cap' (verb) is 'place a limit or restriction on', this means that the word "cap" is not exclusive to absorbs which can be extended.
If, as you say, there is a maximum value that cannot be exceeded, then yes, PWS has a cap.

But there is no 'limit or restriction placed on' it. The maximum value is due to the stats you can have when you cast it.
Would you say 'heal' (the spell) has a cap? I would not, but if your (re)definition of 'cap' would have you say yes, then yes, PW:S would have the exact same kind of cap. Otherwise no 'cap' applies.

[Myllior]

Incredible work Myllior!

Thanks Aparthia!

So are we going int>crit to 18%>mastery>multistrike>haste>vers now?

Well those numbers came from considering 100% of healing coming from non-Empowered, Divine Aegis-producing spells (excluding PoH), so as it seems that PW:S/CoW/EAA-PoH are going to constitute a very large part of Discipline's healing, you wouldn't want to use that. Rather, you'll go with something more like,

Intellect > Mastery > Critical Strike > Multistrike > Versatility

which agrees with what others have stated in other threads/forums. Where Spirit and Haste fall into that priority is beyond the scope of the spreadsheet, and I'm afraid I have no analysis to reasonably quantify them either.

[Jimjam38]

Basically yes. The second cast would cut it down to the cap, though. (Meaning it has a negative effect)!
That is how DA where this kind of cap was needed first started out, and I do not remember this basic behaviour getting changed. (Since Ulduar.)
It should apply to all similar cases just with different caps and I believe I saw it happen with CoW. It hit for just below 100k on someone with 180k HP. I could be mistake,so maybetest it on lowies?

I haven't seen it happening with CoW yet, and I've been watching. I've even set up TMW bars to monitor the PWS / CoW / DA values on my target and mouseover target (because its important when you get a PWS bigger than CoW, meaning that if you keep spaming CoW and it never goes over the PWS then it keeps 'sniping' the pws and you're likely to have the PWS drop off unused)

- - - Updated - - -

But there is no 'limit or restriction placed on' it. The maximum value is due to the stats you can have when you cast it.
Would you say 'heal' (the spell) has a cap? I would not, but if your (re)definition of 'cap' would have you say yes, then yes, PW:S would have the exact same kind of cap. Otherwise no 'cap' applies.

The definition requires a direct 'limit or restriction'. If there's no limit or restriction then there's no cap. A cap implies that if you theoretically had infinite spell power (only spellpower), the spell would not be infinitely large, ie, it would be capped at some value related to something other than your spellpower.

- - - Updated - - -

Thanks Aparthia!



Well those numbers came from considering 100% of healing coming from non-Empowered, Divine Aegis-producing spells (excluding PoH), so as it seems that PW:S/CoW/EAA-PoH are going to constitute a very large part of Discipline's healing, you wouldn't want to use that. Rather, you'll go with something more like,

Intellect > Mastery > Critical Strike > Multistrike > Versatility

which agrees with what others have stated in other threads/forums. Where Spirit and Haste fall into that priority is beyond the scope of the spreadsheet, and I'm afraid I have no analysis to reasonably quantify them either.

I'd still like to know how much of an ilevel upgrade we need (lets say in terms of int) for us to want to take versatility on an item.

Also, i'm thinking for raid purposes (not considering int or spirit) we want:

Mastery > Crit => Multi > Haste >>> Versatility.

This is including raid buffs.

Meaning, mastery is our best stat. Second best is crit, but its almost equal to multistrike. We want to balance the crit/multi% (not ratings) or have higher crit %, but not higher multistrike %. Haste isn't bad, but it's not as strong as the first 3. Versatility is best avoided unless it's a significant ilevel upgrade, it's quiet bad, it scales very poorly and doesn't interact much.

In choosing an upgrade this means that any item with mastery is an upgrade, unless it also has versatility, in which case it needs to be a significant ilevel upgrade.
You want to go for both Mastery/Crit and Mastery/Multistrike items, not just Mastery/Crit. Mastery/Haste items are also viable, as are Crit/Multi items. Any item without versatility is better than one with versatility. Everything is an upgrade from Versatility/Haste.

For item stat combos, same ilevel, this means:

Mast/Crit = Mast/Mult > Mast/Haste => Crit/Multi > Crit/Haste > Multi/Haste > Mast/Vers > Crit/Vers > Multi/Vers > Haste/Vers

- - - Updated - - -

What I'm not sure about is the scaling of spirit. Is 100 spirit going to be more throughput than 100 mastery/crit/multi/haste?
If spirit scales poorly then it might be a case of quality over quantity. Of course, we're healers, we need to be able to heal when we need to heal, but if spirit scales poorly, then the couple extra casts it provides in a fight might not be worth having stronger spells for the entire fight. Idk. Fortunately it's only on a few items, so for now, get spirit on those items if you can, if you can't don't worry too much.

[Noradin]

The definition requires a direct 'limit or restriction'. If there's no limit or restriction then there's no cap. A cap implies that if you theoretically had infinite spell power (only spellpower), the spell would not be infinitely large, ie, it would be capped at some value related to something other than your spellpower.

Exactly. Thus "no cap applies".

[Haywire5714]

Multi/Haste > Mast/Vers

Hmm interesting.

[Myllior]

I'd still like to know how much of an ilevel upgrade we need (lets say in terms of int) for us to want to take versatility on an item.

I don't know of a simple way to give an accurate answer for that, but it's probably better (although more involved) to look at individual scenarios; e.g. replacing Item 1 with Item 2. The quick, approximate way to analyse this would be to evaluate the changes in the relevant stats (i.e. new item minus old item), multiply them by their respective current stat weights (for your gear) and sum the results; if greater than zero, the new item is (probably) better.

A more accurate way (particularly if neither Haste nor Spirit is present on either piece), would be to take your spell breakdown and evaluate Eq. (7) in the equations sheet with your stats before and after the item change; whichever setup gives the greater result is better. If Haste were present, one could perhaps try adding a (1+H) factor to Eq. (7), and maybe neglect both Borrowed Time and the extra mana needed to sustain higher output with increased Haste as they partially counteract one another.

All in all simple analytical solutions can only take you so far when Haste and Spirit are considered; while rough approximations for their value can be produced, the equations required for greater accuracy are better suited to numerical solution (i.e. SimulationCraft). For example, a possible, simple way to get a weight for Spirit could be to look at the extra mana available due to an increase in Spirit, relative to that already available. If Spirit is 'W', then with an increase of 'dW', for a fight of 't' seconds in length, this relative increase (RI) would be,

RI (W) = 0.412207024*dW*t/(160000 + 640*t + 0.412207024*W*t)

Note that the denominator would need to be modified to include extra sources of mana, such as Power Word: Solace or mana potions. While this does give you an indication of how much additional mana you'll have available, relative to what is already available, the question is then how will that mana be spent? Or, will the extra mana even be needed? That's where everything gets complicated.

Hamlet's done good write-ups on Spirit during MoP and also looking forwards into WoD. They're a good read to try and get a sense of the value of Spirit. The latest is: http://iam.yellingontheinternet.com/...nt-and-future/

[Evry]

Spirit is always a difficult stat to assign some sort of absolute weight. The good news is it does scale pretty well in Warlords, so it isn't much of a wasted stat. (100 spirit is ~41 mana per second, or 1.52% of the cost of a PW:S per 1.43s)

I chose PW:S because that's primarily what spirit will get you. More Power Word Shield/CoH and maybe some Prayer of Healings. If PW:S + CoW make up 66% of your healing then that spirit could represent a 1% (additive) increase to your healing.

So, I would value spirit around the same area as Crit and Multistrike; you'll have to weigh casting a few extra PW:Shields during the fight versus having more throughput for burst healing. Both are important, especially since Disc doesn't overheal as much.

[Jimjam38]

So I've been trying to work out how much spirit we need to go ham with PWS spam for an average fight length:

We have 784 base spirit, equal to 2% Mp5.

From this, we work out that you need 392 Spirit per 1% Mp5, which equals 1960 spirit for 1% mana per second (Mp1).

PWS costs 2.4% base mana, and assuming zero haste, has a 1.5 sec GCD. This means if we were spaming PWS without pause, we would need 1.6% Mp1, or 8% Mp5, to never lose any mana (ie, regen = consumption)
In a raid environment this is absurd, and also unnecessary, since we don't need to end the fight on full mana.
Assuming an average fight length of 8 minutes during progression (some shorter, some longer), we only need enough mana to go oom on the last cast of the fight. Realistically this is impossible to calculate, since spell selection, movement etc etc all affects our mana consumption and hence our required mana regen.

We have 160000 base mana, so with zero regen this is equal to 41.66 casts of PWS, assuming zero haste this is 62.5 seconds before going oom. With base 2% Mp5, we regen 25% mana in this time (62.5/5*2), or 40000 mana. This returns another 15.625 seconds of casting PWS, during which we regen another 6.25% mana or 10000 mana. This returns another 3.90625 seconds, and we regen ~1.5% or 2500 mana and are unable to cast another pws for about a second.

So, our base 2% Mp5 nets us: 62.5 +15.625 + 3.90625 = 82.03 seconds of spaming PWS.
Another way of calculating this is converting the 2% Mp5 to Mp1, subtracting that from the %mps cost of PWS, working this out as a mana cost per sec and then dividing our base mana by this mana per second cost.
ie. 2% Mp5 = 0.4% Mp1.
PWS = 2.4% per cast, 1.5 sec GCD, 2.4/1.5 = 1.6% mps
1.6 mps - 0.4 Mp1 = 1.2 mps
160000 * 0.012 = 1920 mana
160000 / 1920 = 83.3333333333333 seconds till OOM.

Working backwards, if we want to sustain pure PWS spam for 8 minutes: 8 min = 480 sec
160000 / x = 480
x = 160000 / 480 = 333.3333333
So we need a pseudo mana cost of 333.333333 mana
PWS original mps is 2560
2560 - y = 333.33333
y = 2560 - 333.33333 = 2226.6666
2226.6666 / 160000 = 0.013916666
~1.4% Mp1
= 7% Mp5
= 2744 spirit

Now, if we take into account solace.
Solace provides 2% mana per cast, 10 sec cooldown, equivalent to 0.2% Mp1 + a 'free' GCD every 10 seconds.

Accounting for the 'free' GCD's that won't be spent casting PWS. In an 8 minute fight you have 480 seconds, so 48 'free' GCDs of mana regen on top of the mana regen from the solace effect, equal to 72 seconds.
Which means 408 seconds of spaming PWS in a 480 second fight.
x = 160000 / 408 = 392.1568627...
need a pseudo mps cost of 392.157
PWS original mps is 2560
2560 - y = 392.157
y = 2560 - 392.157 = 2167.843
2167.843 / 160000 = 0.01355
= 1.355 % Mp1
Following the same process as before, we get down to 1.35% Mp1 required, which is equivalent to 2646 spirit. Solace is equal to 1% Mp5 (0.2% Mp1 * 5), which is equivalent to 392 spirit, bringing us down to 2254 spirit.
However this doesn't account for the fact that during those 'free' GCD's we're also regenerating mana.
If we let out final spirit requirement equal x, then we get: x = 2254 - (x/1960)*72
x = 2174.1
So, our spirit requirement to spam PWS with Solace on Cooldown for 8 minutes is 2175.

From my research, with the 690 legendary ring (we don't know what comes after that, but I assume it will be for blackrock, since mythic blackrock is 695, and the legendary ring wouldn't be lower ilevel than current gear) and the highest spirit items available from mythic blackrock, with the shadowmoon enchant, the highest theoretical spirit we can obtain is 2013.5 spirit. (.5 because of the averaging of the proc from the enchant) (the RPPM spirit trinket is probably slightly higher mana regen than the static one, idk)

This means we're actually not too far off a state of almost infinite mana, or at least very comfortable mana regen. I expect that by T18 mythic we'll be back to 5.4 SoO levels of mana freedom. Inb4 nerfs to Solace or PWS mana cost

Of course, this is all assuming zero haste, and that you somehow manage to use Solace perfectly on cooldown, which is impossible due to a 10 sec cooldown with a 1.5 sec global cooldown, unless you wait 0.5 seconds for it to come off cooldown without casing anything, in which case you'd have to also account for that 0.5 second wait time (essentially another 24 seconds of 'free' GCD's - following the same steps but adding in the 24 seconds gives us x = 2129.29). It also doesn't account for using other spells (everything except PoH, FH, Saving Grace and Halo will have a lower mana cost than PWS though), or for downtime for whatever reasons.
It would be possible to work it out in terms of haste as well, but that's far beyond my levels of sanity.

I'm sure there's plenty of flaws in my working, and I'd appreciate if any of the more mathy types could look over it.
Assuming my maths is correct however, you could attempt to work out a 'general' version for a typical disc spell usage where you calculate the 'average' mps of a disc's typical spell usage (including atonement, penance, PoH, CoW, dispels etc etc) and use that in the place of the "original PWS mps"

.....

On another note, you said haste is hard to calculate, but doesn't haste increase our throughput by an amount equal to our haste % ? (ignoring mana limitations of course).
As in, at 50% haste, our GCD is at 1sec. If you cast PWS for 150 seconds with zero haste you get 100 casts, and with 50% haste you get 150 casts, or a 50% increase. This is essentially the same for cast time spells too: CoW has a 2.5 sec cast time, if you cast it with zero haste for 250 seconds you get 100 casts. If you have 50% haste it has a 1.6666 sec cast, so if you cast it for 250 seconds you get 150 casts, again, a 50% increase.

- - - Updated - - -

I don't know of a simple way to give an accurate answer for that, but it's probably better (although more involved) to look at individual scenarios; e.g. replacing Item 1 with Item 2. The quick, approximate way to analyse this would be to evaluate the changes in the relevant stats (i.e. new item minus old item), multiply them by their respective current stat weights (for your gear) and sum the results; if greater than zero, the new item is (probably) better.

A more accurate way (particularly if neither Haste nor Spirit is present on either piece), would be to take your spell breakdown and evaluate Eq. (7) in the equations sheet with your stats before and after the item change; whichever setup gives the greater result is better. If Haste were present, one could perhaps try adding a (1+H) factor to Eq. (7), and maybe neglect both Borrowed Time and the extra mana needed to sustain higher output with increased Haste as they partially counteract one another.

All in all simple analytical solutions can only take you so far when Haste and Spirit are considered; while rough approximations for their value can be produced, the equations required for greater accuracy are better suited to numerical solution (i.e. SimulationCraft). For example, a possible, simple way to get a weight for Spirit could be to look at the extra mana available due to an increase in Spirit, relative to that already available. If Spirit is 'W', then with an increase of 'dW', for a fight of 't' seconds in length, this relative increase (RI) would be,

RI (W) = 0.412207024*dW*t/(160000 + 640*t + 0.412207024*W*t)

Note that the denominator would need to be modified to include extra sources of mana, such as Power Word: Solace or mana potions. While this does give you an indication of how much additional mana you'll have available, relative to what is already available, the question is then how will that mana be spent? Or, will the extra mana even be needed? That's where everything gets complicated.

Hamlet's done good write-ups on Spirit during MoP and also looking forwards into WoD. They're a good read to try and get a sense of the value of Spirit. The latest is: http://iam.yellingontheinternet.com/...nt-and-future/

Thing is that for most people that's a very complicated way to determine if they want the upgrade or not.
Also, do we have any clear numbers for disc stats weights?

Also, what would be good is a "if it's this many ilevels higher it's ALWAYS an upgrade, regardless of stats" that people can use for a basic reference for items with versatility and other ambiguous upgrades.

[Aparthia]

We have roughly 960 base mana regen per minute.

We have our base spirit and the 2% base mana regen per five seconds.

My priest has 781 base spirit. That's 321.93 mana per second(781*2.061/5). In addition to that you regenerate 0.4% of your mana per second, which is an additional 640 MPS(0.004*160000) for a total of 961.93 MPS.

In terms of spirit that's the equivalent of 2333.65.

[Jimjam38]

I thought the 782 base spirit is what gave us our base Mp5.

Ah fuck it all then, all those calculations were pointless

[Aparthia]

It's pretty easy to check.

You can record look at the tooltip for spirit and divide by 5 for the mps

You can nuke all of your mana and engage a low-level mob to see how fast you regenerate it.

Etc

[Myllior]

On another note, you said haste is hard to calculate, but doesn't haste increase our throughput by an amount equal to our haste % ? (ignoring mana limitations of course).

That's a rough approximation for the value of Haste, yes; indeed it's the one I use in my spreadsheet for its simplicity. But it's not something I'd be comfortable with using to strictly analyse its place relative to other secondaries as, in addition to neglecting that it has a negative impact on mana to achieve the same total output, it neglects the effect of Borrowed Time and also that spells with cooldowns (say, Penance or Empowered Archangel) don't benefit from it anywhere near as much as spells without cooldowns can (e.g. CoW).

Thing is that for most people that's a very complicated way to determine if they want the upgrade or not.

True, but it's certainly more accurate to use the weights produced considering your personal gear and spell breakdown. I don't currently know how to provide a general answer for the required item level difference for a higher ilvl piece to be better regardless of the secondaries it has, but if you take a little bit of time to use either of the methods, you can be fairly confident in the result.

[Deleted]

@Atonement about basic gearing guide :

AThere are just 5 basic rules:

1) Generally pick the highest ilvl unless its something retarded. For same ilvl upgrades refer to 1-4.

2) Know the mastery/crit balance points (put formula above in excel and just input your mastery to see if your current crit and above or below the balance point.

3) If your mastery is higher than the balance point by more than 2-3% and you want to keep stacking primarily mastery (because you cast PWS/CoW a lot) Avoid both multi and versatility, since their value is quickly suppressed by mastery.

4) If your crit is higher than the balance point by more than 2-3% and you want to keep stacking primarily crit (because you cast heals a lot for some inexplicable reason, or because you want to off-DPS a lot) take multistrike and mastery in equal measures.

5) If you are near the balance point and you want to stay around there then take mastery over everything else.

Basically there are three basic gearing strategies.

a) Stack mastery for PWS/CoW and then for same ilvl upgrades pick crit over everything else.

b) Stack crit for heals/off-DPS and then for same ilvl upgrades pick multistrike and mastery on equal measures.

c) Stack balanced crit/mastery for a mixed playstyle and ignore all other stats.

Regardless of your gearing strategy, you lose very little by taking some amount of a "wrong" stat.

That's still the easiest way to handle your upgrades IMO. Since there's currently not a lot of crit/mastery items and the fact that pw:S is generally our most used spell in raids ; everybody is currently going mastery over crit. If you already have a lot of mastery (which is certainly the case), follow scenario a).

A more accurate way of gearing is using myllior's sheet if you have figured your spell usage balance.

[Jimjam38]

From the two bosses we managed to do last night amongst all the server bullshit my healing was coming out at ~70%+ from PWS/CoW and a large portion of the remaining non PWS/CoW healing was from Penance.

Since for PWS/CoW, Multistrike and crit are equally valuable, I don't think it's too problematic that there's an abundance of multistrike and a lack of crit on the available gear, but it probably means that we want to gem / enchant crit to balance the multistrike?

- - - Updated - - -

Also, what are the current secondary stat budgets for Highmaul norm, heroic and mythic?

- - - Updated - - -

I've made a spreadsheet calculating the approximate secondary stat budget for Highmaul Norm and Mythic:

https://docs.google.com/spreadsheets/d/1jCyse5SzlIgXW2Ous8_TvdstPH-_4u0C4RDYWL2nj6k/edit?usp=sharing

It's worth noting that for some reason the stat budget on items can vary (in some extreme cases) by up to 10 stats between armour types and even within the same armour type for different options. There doesn't appear to be any obvious trend or pattern behind this, but my suspicions are that it's actually related to which boss drops the item, although I haven't tried to conform this suspicion.

Anyway, point is that from items, even at mythic highmaul there's only a total secondary stat budget of ~5000. Enchants and gems and warforged can elevate this, and assuming that the stat budget increase is linear with itemlevel, i predicted that mythic blackrock will only have a 400 stat budget increase on this, however again i didn't check it.

At normal highmaul levels there's only ~3800 available, and with a 70% absorb, 30% normal heals, Myllior's spreadsheet basically says mastery and crit only, with multistrike barely appearing until. The majority of high ranking logs so far show between 70% and 80% absorbs (although, Divine Aegis is often part of this).


So, essentially, Mastery > Crit > Multi (crit not available) > Haste > Versatility

[Aparthia]

I doubt we'd want to gem or enchant crit to balance MS/Crit when we can gem/enchant for Mastery. Mastery is still ahead when most of your healing comes from absorbs(on our 7/7 normal clear yesterday, 77.2% of my healing was absorbs of which PWS was 47.3 was PWS, DA being 16 and rest CoW).

[Deleted]

What's the best weapon enchant for Disc ?

For the moment, I have all rings/neck/cloak with spirit, 3/3 DMC, Everburning Candle. I don't feel like I need more mana return at the moment.

I use mastery gems and enchants. Should I go for the mastery weapon enchant or the crit one ?

For the moment, I have a 2H staff with haste/multistrike and multistrike enchant (I play Holy too but I hope I will get a weapon for the disc spec).

Crit weapon enchant seems nice but mastery is the stat we want thanks to overuse of PWS and CoW.

Any advices ?

[Aparthia]

Mastery enchant if you don't need the mana, otherwise spirit one.

[Nadone]

I'm trying to get some stat weights to better evaluate the gear choices available to me, I stopped playing the game for over a year now and bought the expansion late so I'm a bit overwhelmed by the amount of information and misinformation out there.

I ran Simc but I got Crit coming out on top with a really big margin to Mastery, I assume from this Simc doesn't simulate Disc priest how it is being played atm.

Back to the question, How can I get numerical stat weights so that I can evaluate the choices better?

[Aparthia]

Use the spreadsheets linked to in this thread.

[Nadone]

Use the spreadsheets linked to in this thread.

Yes, terribly sorry I thought I had read the thread properly turns out I was just scanning through.

Thanks to Aparthia, Myllior and all others in the thread, keep the discussion alive