Disc Priest - Mists of Pandaria - Page 22

royalpoet · 02/15/13, 5:22 PM

Originally Posted by Polopretress

sorry,
i do not understand the answer of GC.
(all the sentence after "but")

does he mean that the ICD of the meta is 10 sec with 100% of chance to proc and then, as we use anothers spells by portion of 10sec, the meta will proc anyway on its full uptime ?

He's saying that the way the Real PPM system (that they've been moving things like weapon enchants and trinkets to) works, as long as you cast some kind of healing spell, including periodic ones, once every 10 seconds, you'll proc the metagem at its maximal rate.

Perkeyone · 02/15/13, 5:23 PM

If these changes go live, unless we cap DA, which will be at 60% max hp, the overheal considerations for crit are almost gone.

The quadratic scaling on mastery seems like something that might get out of hand like armor penetration.

kouby · 02/15/13, 5:32 PM

Same here, I don't understand the comment.
In the same topic it's stated that the meta doesn't have an ICD, but then it appears that smiting for 10second intervals won't reduce our proc chances? I'm having a major brain malfunction on this one. I thought you could chain procs with the RPPM system, and I've seen it on live with Jade Spirit unless I've started hallucinating as well, so how can doing something that can't proc it for 10seconds not be a loss?

royalpoet · 02/15/13, 5:40 PM

Originally Posted by kouby

Same here, I don't understand the comment.
In the same topic it's stated that the meta doesn't have an ICD, but then it appears that smiting for 10second intervals won't reduce our proc chances? I'm having a major brain malfunction on this one. I thought you could chain procs with the RPPM system, and I've seen it on live with Jade Spirit unless I've started hallucinating as well, so how can doing something that can't proc it for 10seconds not be a loss?

Real PPM doesn't use ICDs, it essentially uses a modified proc rate, where the proc rate is modified by the last time you had a chance to see a proc (it's also modified by haste %). The longer it's been, the greater the chance of a proc. So, unlike with an ICD, you can see back-to-back procs but over an extended period of time, the moving proc rate will even out the number of procs. With a 1.00 RPPM, you need a possible proc event at least once per 10 seconds to see that maximal over-time rate. If you spam cast instead over that 10 sec, you might see back-to-back procs over that period but then a longer time to your next proc (i.e. the gambler's fallacy isn't a fallacy with RPPM, because each event isn't independent...the system is accounting for how quickly you're seeing chances to proc and adjusting its rate).

• Hamlet · 02/15/13, 6:01 PM

Well, let's see. Heals increase like (C is crit chance, M mastery (as a ratio of 100 points of mastery, so equal itemization costs)):

(1+0.8M)*((1-C)+1.03C*(2+1.6M))
(1+0.8M)*(1+1.06C+1.648CM)
(1+1.06C+1.648CM)+(0.8M+0.848CM+1.3184CM^2)
(1+1.06C+0.8M+2.496CM+1.3184CM^2)

dH/dC = 1.06+2.496M+1.13184M^2
dH/dM = 0.8M+2.496C+2.6368CM

Graphing H vs. M with C constant at 0.2 (20% crit):
(1+1.06*0.2+0.8x+2.496*0.2*x+1.3184*0.2*x^2) for x from 0 to 0.5 - Wolfram|Alpha

The nonlinearity isn't that dramatic. This shows a range from 0 mastery to 50 points of mastery (80% new shield bonus).

------

e: Above description of RPPM is correct. Say you cast one heal every 10 seconds and atone at all other times. And say you have 20% total haste with a 3 RPPM trinket. Since the game "owes" you 3.6 procs every minute, or 0.6 every 10 seconds, then each of those heals will have a 60% chance to proc.

Polopretress · 02/15/13, 7:16 PM

Very usefull plot would be to have "Mastery points v/s %C" and be able to see when we are above the curve which mean that average heal is better on 5.2 than in 5.1 and when under the curve, average heal is better on 5.1 compare to 5.2.

Do you see what i mean ?

edit : what is RPPM ?

Havoc12 · 02/15/13, 10:00 PM

It does not matter if it procs off smite, the RPPM system seems to make stuff proc at the right frequency if you don't chain cast. Popping a couple of heals every minute should be fine.

With the formula that you have hamlet crit is better than mastery until 46% crit or so but I just had it pointed to me in another thread that only the aegis part remained the same for crits. The rest of the crit is now cut in half.

That means the actual formula is base*(1+0.5*mastery)*(1+crit*(1+mastery))

So comparing the old formula with the new it is (1+0.5*mastery)*(1+crit*(1+mastery))/(1+crit*(2+mastery))

At 60% of the old mastery, 38.4% for the new and 15% crit this is

(1+0.5*0.384)*(1+0.15*(1+0.384))/(1+0.15*(2+0.6)) = 1.035582158273, so just a 3.6% increase.

Crit is now dramatically weaker.

Let us compare it to what it would be with the old system and crit stacking (1+0.5*0.384)*(1+0.15*(1+0.384))/(1+0.26*(2+0.325)) = 0.897138797133

A 10% nerf, but since that assumes zero overhealing its probably no change.

tib · 02/16/13, 8:24 AM

"With the formula that you have hamlet crit is better than mastery until 46% crit or so but I just had it pointed to me in another thread that only the aegis part remained the same for crits. The rest of the crit is now cut in half.

That means the actual formula is base*(1+0.5*mastery)*(1+crit*(1+mastery))"

You wouldnt happend to have any half reliable source of data for this as frankly what hamlet wrote makes a hell of a lot more of sense when reading ghostcrawlers post.

Havoc12 · 02/16/13, 10:54 AM

You didn't read it carefully I am afraid. GC said that crits no longer heal for twice the amount as did hamlet himself. They just put a shield equal to 100% of the shield. Now stop and think before a heal would crit for 2x the normal amount and apply a shield equal to 0.5 of the crit. 0.5*2*normal heal = normal heal. So it would crit for twice the non amount and apply on top a shield equal to the non crit amount. In the formula hence each crit added 1 for each heal and 1+mastery for the aegis, this means 2+mastery total. Crits now only add aegis so its 1+mastery extra healing for each crit.

Basically what they did was keep the amount of aegis added for each crit the same and rolled part of the crit heal into the normal heal. If you had 60% mastery, 15% crit that means all heals are the same and 19% bigger instead of having 15% of your heals 100% bigger.

Overall its a slight nerf to theoretical throughput because you could achieve 10% more with crit gear, however as I have said so many times before, crit has 50-70% return in any real senario, so in truth there is little difference overall. Disc will be a lot stronger in the low damage phases but slightly weaker in the high damage phases.

The main consequence is that now holy is even better at disc at healing with high incoming damage.

tib · 02/16/13, 12:14 PM

Mye.
Lets start with some basic definitions and actual formulas and less ramblings as its a bit hard to draw any proper conclusions before those definitions and formulas are actually correct.

Assuming M denotes % increase of a shield your mastery gives you since thats what you seem to use (hamlets way of expressing it reforge neutral is alot better btw, cant really compare M and C easily this way), C denotes crit chance

Non crit heal = ncheal = base*(1 + 0.5M), 0.5 as the heal increase of mastery is now half of the shield increase.
Crit heal = cheal = base*(1 + 0.5M) + base*(1 + 0.5M)*(1 + M) , left side being the heal, right side being the DA.

Now lets remove the base part since its not really interesting.
The average heal is then modeled as follows:

avg heal = ((1-C)*ncheal + C*cheal) / ( 1-C +C ) = (1-C)*ncheal +C*cheal
= (1-C)*(1 + 0.5M) + C*(1 + 0.5M   +  (1 + 0.5M)*(1 + M))
= (1-C)*(1 + 0.5M) + C*(1 + 0.5M)*(1 + (1 + M))
= (1 + 0.5M)*(1-C + C*(2 + M))

This is exactly what halmet did above, except he also assumed the crit metagem which I was assuming wouldnt be that interesting come next patch (perhaps im wrong on this one?)

As far as i can tell from the number examples ghostcrawler been posting lately, this aligns perfectly with those aswell, if you have actual testing from PTR that shows something different then thats interesting.

Havoc12 · 02/16/13, 2:32 PM

avg heal = ((1-C)*ncheal + C*cheal) / ( 1-C +C ) = (1-C)*ncheal +C*cheal
= (1-C)*(1 + 0.5M) + C*(1 + 0.5M   +  (1 + 0.5M)*(1 + M))
= (1-C)*(1 + 0.5M) + C*(1 + 0.5M)*(1 + (1 + M))
= (1 + 0.5M)*(1-C + C*(2 + M)) = (1+0.5M*(1+C(-1+2+M)) = (1+0.5M)*(1+C*(1+M))

There I fixed it for you. Its really simple every heal both crits and non crits heal for base*(1+0.5M). In addition each crit adds base*(1+0.5M)*(1+M) on top of that.

Thus the total healing is base*(1+0.5M)+crit*base*(1+0.5M)*(1+M)

My formula is correct

Hamlet's formula is the same as mine except with (2+M) as the crit scaling factor, which is not correct.

I don't know what ghostcrawler thinks but tbh, I don't think he has seriously thought this through. These changes are not rolled in the PTR yet, probably something yet to come?

• Hamlet · 02/16/13, 5:36 PM

I made a post on the last page with a comment that had the (2+M) mistake, but didn't do anything with it. The one on this page (post 320) is correct and uses a mastery convention that makes it a little easier to compare to crit.

tib: I'd definitely stick with crit meta even after patch (and I'm actually in favor of the Int flavor, but that's a separate topic). Ember meta is, to put it simply, pretty bad for all healers. This is especially true in 5.2 when Disc will strongly favor Solace over Mindbender (Mindbender is all %-based regen, while Solace's mana benefit is partially %-based and partially through nullifying the cost of Holy Fire).

IIRC, Shadowfiend is 9% mana every minute on average, and Solace is at most 6%. 6000 added mana, at 15%/min of %-based regen, is only 75 MP5.

tib · 02/16/13, 6:54 PM

It's not the ember meta I'm thinking of using but rather the new legendary Courageous Primal Diamond - Meta - Gems - Items - WowDB (PTR). Without having done any numbers it strikes me as pretty unlikely its not intended to severely beat the current ones if nothing else due to its color.

About the latest 5.2 changes though.
Below is a picture of the heal scaling coefficient function we been discussing under following assumptions:

baseline crit is 14% (5% + 5%(spell crit buf) + 4% (int on gear)).
baseline mastery is 13 points (8 + 5 (might))

Y-axis is the coefficient.
X-axis is the mastery rating. For any given line the sum of mastery rating and crit rating is fixed to K.
So for the K=9900 line if X=4500 then you have 4500 mastery rating on your gear and 5400 crit rating. This way of thinking should fairly well capture the essence of ilvl.

While in theory this looks like you want to hit the sweet spot i dont think so, the difference between the minimum coefficient (all mastery reforged) and the maximum (some kinda middle ground) is very very low.
Almost always you'd want as much mastery as possible in practice as its more reliable (and then there's PW:S)

Reading the latest posts about atonoment healing and spirit shell it seems to me they should scale the same way as normal healing with respect to crit and mastery effectively making these curves covering all healing except PW:S.

• Hamlet · 02/16/13, 8:35 PM

Oh haha, I forgot about that. (...I'm so used to reiterating the argument that Ember meta is bad for everyone that I just spit it out reflexively now).

Yes, assuming legendary meta will be used. I haven't mathed it out in detail, but basic first look is that, while it's not as dramatic an effect as it is for DPS classes, will still be easily usable.

Atonement and Spirit Shell do scale coherently with crit/haste/mastery (as of 5.2), so yes, same math for all non-PWS spells.

Perkeyone · 02/17/13, 12:19 PM

Supposedly there were a few goals in mind when they came up with the ideas for these changes.

- Makes Disc still awesome at bubbles, but not quite so weak at heals. (Holy will cast bigger heals than Disc, but not 50% higher.)
- Making crit a good stat, because it benefits most of the toolbox (including PW:S) and causing Divine Aegis bubbles, but also keeping mastery a good stat, because the bubbles are large, and even when you don't get a bubble, it will still help your heal.

On the first goal, technically our average raw healing would be nerfed more than it is buffed. Average raw healing (nonabsorb) used to be roughly (1+c)/h and will now be (1+m/2)/h ignoring meta gems. So we went from 1% per 600 crit rating down to .8% per 600 mastery rating. The average bubble will be a roughly the same, maybe a tiny bit bigger, but you will be reforging out of crit and proc the aegis less often.

On the second goal, they definitely didn't make crit a good stat, in fact, it lost about 1/3 of its value (slightly less than a third at higher levels of mastery).

I feel like a better way to address the unreliability of crit would be to give priests back renewed hope in some form or another. Renewed hope was a 10% (if i remember correctly) boost to crit chance on targets with weakened soul or grace. Renewed hope gave PW:S a bit more synergy with the rest of our tool kit by making it easier to land crits. An alternative to the old renewed hope would be a critical mass or shatter type buff or a recklessness type cooldown.

On the good side, crit affecting PW:S does fix the one weird scaling issue in our tool kit, and the DA change does make mastery much more viable than the previous ptr set up, but they gutted crit to salvage mastery, which seems a bit counter productive.

This also makes us the only spec where our critical heals are only useful if the target takes damage again soon. Where previously they were equal without extra damage and better with extra damage. The only spec where our critical heals can "fade" or "expire" and overheal without actually filling the hp bar.