For holy, PT/Lightweave or Heartsong/Darkglow? (these things change)

[zakaluka]

edit: I've run some sims, the transient int from PT/LW is still worth a little more than HS/DG

I'm curious what the current state of this is. It changes with item level - mp5 per int gets lower and lower with a higher gearpoint, and mp5 per spirit goes up by a lot. It's why I'm bringing up this long dead topic that had a pretty clear answer in 4.0.

PT and Lightweave both seem to average about 130 int.

Heartsong seems worth 180 spirit (very high uptime), and I can't find reliable numbers for darkglow. Probably about 130 comparing the tooltip to Lightweave.

So, which is better? Averaging 260 int on procs, or averaging 310 spirit from procs?
On one side, int procs weigh in less for holy than disc. Passive regen plays a much larger role. For disc, the benefits are throughput + active regen, so it's a no brainer to use PT/LW.

SpiritRegen(SPI,INT,LEVEL) = (0.001 + SPI * sqrt(INT) * BASE_REGEN[LEVEL]) * 5
dR/dSpi = 5*base*sqrt(int)
dR/dInt = ( 5*base*spi ) / ( 2 * sqrt [ INT ] )

dR/dSpi = 5 * 0.003345 * sqrt ( 6999 ) ->>>>>>> raidbuffed
dR/dInt = ( 5 * 0.003345 * 2300 ) / ( 2 * sqrt [ 6999 ] ) ->>>>>>> raidbuffed

dR / dSpi = 1.40 ( mp5 from HS+Darkglow = 434 mp5 )
dR / dInt = 0.23 ( mp5 from PT+Lightweave = 60 mp5 )

So, Heartsong and Darkglow have a significant edge on mp5 unless you can time shadowfiend with PT and LW. Consider also they're unlikely to proc near the same moment except right at the beginning of the fight. Even then, you have an extra 1080 transient int, which is an extra 16k mana. 30% of this is ~5k mana, and in a 5 minute fight that's another 80 mp5 from shadowfiend for a grand total of 140 mp5, if you can make all your procs line up. It usually won't happen, so you're more likely to use fiend with only one proc active which would make the two enchants worth 100 mp5.

It's not enough to say, really, because int procs improve efficiency. By how much? Maybe 3-5%? Does that make up the difference? is 4% overall efficiency worth losing 330 mp5? Maybe, but it's hard to think about. Same 5 minute fight that extra regen counts for 4 extra PoH. I think I have to see some logs from a holy priest in 25 man, at about my gear level, to see if that's enough.

Any thoughts on this? I might have to start playing holy again soon.

[Kelesti]

The other point of the matter is if you're "regen capped" or close to it, you start pulling away from spirit and into haste/mastery itemization where you can. But Power Torrent and Lightweave are the only places that you can actually get Intellect out of the deal, which is by far the "best" stat to have.

And of course, being a Gnome or Blood Elf skews them a little more towards the Intellect bias, especially Blood Elf (it's hard to time a Shadowfiend double proc, but Arcane Torrent only needs 1 second of twin-uptime, not 8). But it is a good point to actually work out that the value of Spirit rises as we get more Intellect. I'd almost forgotten about that since Wrath (even though it's been in the equation since 2.4)

[zakaluka]

Now that sanc is holy's most efficient spell in 25 man, giving us a super-efficient manadump, I don't see the regen cap being really attainable. Back in 4.0 when everyone defaulted to PT/LW, sanc was something to be used sparingly as (1) precast or (2) expensive spike. Now we spam it on cd and try to line up our last one with heaviest AoE.

When raid can stack anyway. I guess there's several fights where it just won't work too well (beth shan rhyo)

[Kelesti]

And here's me, being silly and running 10's.

Assuming Sanctuary ever stops being completely broken and ignoring the target cap, that would shift this entire thing around, wouldn't it?

[zakaluka]

I still run 10s But,

Have you ever been really fed up with the 10 man model because it encourages minimum roster? You're 100% reliant on nobody having a computer problem or some real life fiasco. Every night someone has a computer problem, or someone's late, or someone got tired of raiding 10 (for all these reasons). Basically the roster is unstable and it's the same thing I've experienced in every 10 man . Then people get lazy because we aren't progressing.

So now I'm shopping for a 25 man guild that overbooks its roster. If I have to sit for a third of the night, but then lose less hair staring at the portal to FL, worth it in my book. Someone asked me what I think about switching to Holy because their existing heal team already has 2 discs.

[Kelesti]

My guild's two seperate 10 mans each have a 13 person roster, plus an overlap officer that could fill in a geared alt for either. There are 10s out there that have an overbooked roster and make it work. But that still doesn't change the fact that Sanctuary isn't as good in 10s as it is in 25s! Which of course, skews the need to gear for the "mana dump" that it is. And 25s allow a Priest to make a little more liberal use of Heal over other spells, too, so that could very well come into effect.

At this point, I don't know, I'm just not ready to give up Power Torrent and how awesome lining up my cooldowns can be.

[Ultima]

-interrupts two person conversation-

being a Gnome or Blood Elf skews them a little more towards the Intellect bias, especially Blood Elf (it's hard to time a Shadowfiend double proc, but Arcane Torrent only needs 1 second of twin-uptime, not 8). But it is a good point to actually work out that the value of Spirit rises as we get more Intellect. I'd almost forgotten about that since Wrath (even though it's been in the equation since 2.4)

What about Human towards Spirit?

[zakaluka]

I calculated mp5 without holy concentration, whoops >:< So that'd mean 434 * 80 / 50 = 694 mp5 for heartsong. Starting to look more appealing...

I thought about asking simcraft, but simcraft's chakra implementation isn't quite right. It'll give me screwed up results.

Ultima, regarding your question I'd say yes. Having higher spirit doesn't increase your gain per point of spirit. It does, however, increase your gain per point of int. We're talking about a pretty small spirit bonus humans have, though. It shouldn't really make a noticeable differene on passive mp5 per point of int.

How much extra spirit does the human racial actually give you? For some reason a really low number is coming to mind. I also calculated against 2300 spirit and came up with 0.23 mp5 per point of int. Let's assume you have 2900: 2900 * (0.23) /2300 = 0.29 mp5 per int. Again comes down to whether the effic gain from the int beats out the passive regen from heartsong. Need more data. I have a feeling I'll be testing this soon, so I'll let you know what I decide.

[Wuga]

Your general result that dr/dspi > dr/dint is correct, although this has pretty much always been the case even at very low gear levels (setting dr/dspi = dr/dint and solving for int gives you int = spi/2, which is unreasonably low). Some comments on the specific numbers:

Heartsong seems worth 180 spirit (very high uptime), and I can't find reliable numbers for darkglow. Probably about 130 comparing the tooltip to Lightweave.

180 spirit (90% uptime) sounds too high for heartsong. IIRC it's a 20s icd so the max potential uptime is going to be about 75%, or 150 spirit. As holy you probably have enough random hots rolling that actual uptime would be 70%+ (I forget if it procs off echo or not).

dR/dSpi = 5 * 0.003345 * sqrt ( 6999 ) ->>>>>>> raidbuffed
dR/dInt = ( 5 * 0.003345 * 2300 ) / ( 2 * sqrt [ 6999 ] ) ->>>>>>> raidbuffed

dR / dSpi = 1.40 ( mp5 from HS+Darkglow = 434 mp5 )
dR / dInt = 0.23 ( mp5 from PT+Lightweave = 60 mp5 )

Your base coefficient of 0.003345 is for out of combat regen, so you need to multiply these values by 80%. This doesn't change the ratio of the two but will reduce the magnitude of the mp5 difference. In particular, the 694 mp5 you calculated later in the thread is too high (by a factor of 2).

[zakaluka]

cool, thanks

Edit: (edited rather than replying so I won't bump)
Figured out how to make simulationcraft calculate this. Int, even transient int, is still worth more than spirit. The original point wasn't way off. In fact, right now simulationcraft gives Int a scale factor of 5 and Spi a scale factor of 3 (making int/spi gems finally better for sure in blue sockets with a 10 int bonus, yay!) Revisit this once again in 4.3 gear, we will see if Spi ends up being worth even more.