Disc Priest - Mists of Pandaria - Page 16

• Hamlet · 01/30/13, 7:20 PM

Originally Posted by Jeges

Just a few data points regarding crit, overhealing, etc., from some of my own recent logs:
Protectors of the Endless (Hard): 11.8% overhealing, mean effective crit (not counting DA) 2.06x mean effective non-crit
Sha of Fear: 24.8% overhealing (mostly due to Atonement), crit multiplier 2.04x
Shek'zeer: 14.1% overhealing, 2.16x crit multiplier

These were taken across many attempts, so should be a reasonably accurate picture of my healing. Crit is damn useful.

This is pretty fascinating actually--2.06 is dead on the expected raw healing increase from crits. Notwithstanding everything I've said, I wouldn't have guessed that. In other words I always kind of bought the logic that effectiveness ratio has to go down by some extent when you add crit healing (or indeed, when you add to heal size in almost any form). I just thought it was silly to think with no basis that any added inefficiency would devalue the stat by a factor of 2 or 3.

But that there's no detectable difference whatsoever in effectiveness--that would be an interesting thing to survey more.

Havoc12 · 01/30/13, 9:19 PM

For IF devalues crit. Actually its not. If you have 25% crit, under IF your crit multiplier raise from 1.25 to 2.25 for Spirit Shell. No loses here. Check logs if you dont trust me) IF works for 100% with any amount of crit for Spirit Shell.
Yes there is a cap for all others spell with 100% crit. But SpS cant crit, it only takes crit modifier that can be 200% (which ofc means you crit 100% times, but for Spirit Shell its 3.00 modifier)

Tested this on live sretzelek and you are wrong. Currently IF just doubles exactly the spirit shell absorb. It does not add 100% to the crit modifier. Here is the proof

59 315 normal PoH spirit shell. This agrees perfectly with the spirit shell formula 29502*1.0789*(1+0.4334)*1.3

With inner focus I get ~118600. If this was just adding 100% to the crit value I would expect 29502*2.0789*(1+0.4334)*1.3 = 114 286.73

I changed my values around and increased spellpower and crit (it is now 10%) but reduced mastery

I now get 59609 for a normal spirit shell and ~119200 for IF spirit shell. Again its exactly double, but I would only expect 29602*2.1065*(1+0.3997)*1.3 = 113 464.72, notice how IF would still devalue crit, because 210/110 < 207/107, if it worked like you thought.

However it is clear that IF just doubles spirit shell value regardless of your crit rate. This is almost certainly not going to be the case in 5.2, but I will test it once the PTR is back.

===========================================================================

Hamlet its not fascinating at all, its an artifact caused by inappropriate handling of the data.

When you average a crit you include a full non-overhealing normal spell for every crit that does not overheal more than 50%.

Think of it like this for a crit you have two parts the normal part and the crit part --> C = Hn+Hc

The value of a crit to healing is the Hc part not Hn. So take two criticals. One overhealing for 10% and another for 50%.

i.e. you have C1 = Hn+0.8*Hc and C2 = Hn

If you take the crit average you get Hn+0.4*Hn. The average value of a crit is 0.4*Hn.

If you now compare it to normal heals with an average overheal of 20% you are comparing it to 0.8*Hn. If you take the ratio you get 1.4/0.8 = 1.75*Hn, so you might think that the value of a crit is 0.75*Hn, but this would be completely wrong. If you subtracted the average heal from the average crit you would come up with a closer value, but it still very inaccurate, because to get enough statistical power you would need literally 1000s of casts and we only have 100s.

You CANNOT under any circumstances divide crit average by non-crit average and claim that this is the value of a crit. It is completely nonsense, because the value of a crit is everything above a normal heal with no overheal, it is not the entire value of the crit.

If you have no critical overheals over 50% you can simply subtract the value of a non-overhealing non-crit heal to find the total crit value, which will always be less than 100% if you have overheal > 0. If you do have lots of overheals over 50% you can't do that.

The most important thing to realise however is that you simply cannot estimate the value of crits by dividing the average crit by the average non-crit. The value that you has zero practical value.

Originally Posted by Jeges

Just a few data points regarding crit, overhealing, etc., from some of my own recent logs:
Protectors of the Endless (Hard): 11.8% overhealing, mean effective crit (not counting DA) 2.06x mean effective non-crit
Sha of Fear: 24.8% overhealing (mostly due to Atonement), crit multiplier 2.04x
Shek'zeer: 14.1% overhealing, 2.16x crit multiplier

These were taken across many attempts, so should be a reasonably accurate picture of my healing. Crit is damn useful.

Jeges I have explained this before. You cannot predict the value of crits by looking at average crit vs average heal. You can get 2:1 ratio of heals to crits and still have the average crit be below a non-crit heal. Even more importantly you are averaging over all spells!! You can't do that!! You are completely obliterating any information your data had.

This value that you are quoting has absolutely no meaning. I explained why a few posts ago. Anytime you see a crit average being high its an artifact caused by non-normal crit overheal distributions, low statistical power and inner focus, with the non-normal distributions being the most important factor. You cannot under any circumstances average crits compare them with normal heal and get a meaningful value unless you can assume a normal distribution of crit overheals. The data you are posting have absolutely no meaning.

Not only they don't prove that crit has value, they don't even give us the slightest indication of what the value of crit is.

Post your logs so we can see.

Here is an example:

Details for Kelandrah - 06-01 19:33 - Deus Ex Machina - World of Logs, 10H so its comparable and a non ranking log, so we don't get artefacts.

Look at cascade. Average overheal 60%, average crit = 19237.5 , average crit = 41335.5

The average crit is barely equal to a normal heal. All that is happening is you are getting a non standard overheal distribution and that is why the average is more than 2-fold the normal hits. But the real value of crit is not 2-fold

Here are all the cascade crits

Expression Editor - 06-01 19:33 - Deus Ex Machina - World of Logs

The benefit of crit is nowhere near half the total crit heals. You can easily predict that because a crit heal is just about the same as a non ovehealing normal heal.

Look at PoH

Expression Editor - 06-01 19:33 - Deus Ex Machina - World of Logs

Its even worse, the benefit of crit heals is pretty much just from aegis and nothing else.

In both cases the ratio of average crit to average heal tells us nothing at all about the actual value of crit, because there are many crit heals with more than 50% overheal.

Expression Editor - 06-01 19:33 - Deus Ex Machina - World of Logs

Look at this maybe 1:3 spells has zero overheal. If you replaced x% of those heals with crits only 0.33*x would actually provide any benefit at all. The average crit over non-crit is 1.56

• Hamlet · 01/30/13, 9:49 PM

Originally Posted by Havoc12

Hamlet its not fascinating at all, its an artifact caused by inappropriate handling of the data.

That I can't comment on, since he didn't post his data. I'm assuming for now he did the computation correctly, but it's not one that should have been too complicated.

As to the rest of your post:

What you're stating here is the logical argument I outlined above--since crit healing is added to the "top" of normal healing, it does seems its average effectiveness would be lower. Worth noting when talking about stats though, that this applies to nearly any kind of added healing, although one of the exceptions, ironically, is Divine Aegis (since DA can be effective even if the initial heal was overheal).

However, when trying to measure just how effective crit heals tend to be compared to normal heals, I don't see why you'd do anything other than what Jeges claims he did--see what a non-crit usually (effective) heals for, and what a crit usually heals for. All that crit chance does is move heals from the non-crit into the crit column. If heals in the crit column tend to be twice as effective (for reasons that maybe aren't totally clear), then it is what it is--they really are healing for that much more.

Havoc12 · 01/30/13, 9:55 PM

Originally Posted by Hamlet

That I can't comment on, since he didn't post his data. I'm assuming for now he did the computation correctly, but it's not one that should have been too complicated.

As to the rest of your post:

What you're stating here is the logical argument I outlined above--since crit healing is added to the "top" of normal healing, it does seems its average effectiveness would be lower. Worth noting when talking about stats though, that this applies to nearly any kind of added healing, although one of the exceptions, ironically, is Divine Aegis (since DA can be effective even if the initial heal was overheal).

However, when trying to measure just how effective crit heals tend to be compared to normal heals, I don't see why you'd do anything other than what Jeges claims he did--see what a non-crit usually (effective) heals for, and what a crit usually heals for. All that crit chance does is move heals from the non-crit into the crit column. If heals in the crit column tend to be twice as effective (for reasons that maybe aren't totally clear), then it is what it is--they really are healing for that much more.

No that is not correct. Its an artifact caused by averaging. You cannot average crits and divide by non-crits and then say this is the benefit of crit.

You must subtract from that the value of a non-overhealing non-crit heal.

If you want some hard proof. Let us say that you have two quantities V1 = k + a, where K is a constant and a is a variable with variance x and V2 = b, where b is a variable with variance y.

If you measure V1 and take the average V1* you will get k+a*, where a* is the average value of a

If you measure similarly V2 you will get an average value b* for V2*

Now lets say that you want to know a* and you are given that the maximum value of b is k.

What Jegges is doing is assuming that a*/b* = V1*/V2* - 1

Is this true? The answer is absolutely not.
Thus V1*/V2* = (k+a*)/b* - 1 = a*/b* + (k-b*)/b*

Since k = max value of B ---> k-b* > 0, so what Jagges is doing is adding a factor (k-b*)/b* to the value of a*/b*

That means what he is doing is not calculating the actual value of crits, but he is also adding the proportion that normal heals overheal by.

This is a completely artificial value that has no meaning, and its value depends on overheal distribution and not the value of crit, basically that means it depends on X and Y

In order to average and do a calculation one MUST be certain first that averaging is a legitimate thing to do.

In this case it is absolutely not.

The only valid way to see how much your average crits contributes is to subtract from the average crit the value of a non-ovehealing non crit heal from the logs. This will give you a nonsensical value however if a significant percentage of your heals overheals by more than 50%.

You might think that its simple to calculate the crit contribution, but its anything but.

Crit value can never be more than 100%. It can be 100% only if overheal is zero or if overheal is binary i.e. either 0% or 100%, but nothing in between. It is simply impossible to get anything else.

Szeretlek · 01/31/13, 12:19 AM

Originally Posted by Polopretress

That is why , it is very difficult to find a mathematic formula to take into account all theses criterias...

Healing mathematic is the most difficult mathematic in WoW.
We have limits. Limits about mana, overhealing etc.

DPSer concerns about fight length to manage their CDs and he doesnt restrain himself with DPS output like healers restrain themselves in HPS output. Calc the dps cycle - is a meh. Piece`o`cake.

Tank concerns about survivability that just divide on magical and physical survivability, ehps, spike-resistance. Easy to calc. More - better. Their choice now usually is: more dps or more survival? Easy.

Originally Posted by Havoc12

Sretzelek..

You cant be serious to spell it wrong second time... Give me some respect please.

About SS and IF.
I took my numbers from Expression Editor - 23-01 22:59 - SevereWar! - World of Logs

1st SpS cast was 215825
2nd cast 119903

215825 / 2.25 (crit modifier with IF) = 95922
119903 / 1.25 (crit modifier w/o IF) = 95922
So I was right.

I need your logs to believe you really.

Originally Posted by Havoc12

If you are maxing intellect and you tell me mana is tough the answer is its not because you can afford to not max spirit.

Im maxing Int/Crit NOW because I dont need haste, I dont need mastery and I dont need MORE Spirit. For 5.2 I need to find that level of spirit that suits me (on progression full spirit as always) and then I again start stacking throughput stats. When I will know fights better.
Are you missed my phrase?

Originally Posted by Szeretlek

Go full Spirit/Crit in 5.2!

Szeretlek · 01/31/13, 12:50 AM

About Overheal at crit ratio.
Lets take example from Attonement because at WoL it divides on two parts: crit, non-crit - and its easy to analyze.

I take my time and smite 3 boss down at 1st LFR of MSV
Details for Нищебродка - 30-01 18:43 - SevereWar! - World of Logs

50% heal from non-crit, 25% from crit, 25% from DA.
OH from non-crit 33%, from crit 50%, from DA 25%.

I have 25% crit chance in raid btw. And mastery 40%
If we take numbers from theory and exclude all overheals, distribution should be like this:
Non-crit 47%, 32% from crits and 21% from DA

If we take numbers from log.
My non-crit effectiveness = 1 * (1-33%) = 0.67, where 1 is normalized healing with healing power 1. and 33% is Overheal for non-crit.
Crit Effectiveness = 2 * (1-50%) + 2*0.5 * (1+40%) * (1-25%) = 2.05, where 2 is normalized crit healing with healing power 2, 50% - crit overheal ratio, 0.5 - DA coeff, 40% - mastery rating, 25% - DA overheal.

If we do it your way and exclude normal heals from crit heals, then Real Crit heal Effectiveness = Crit Effectiveness - Non-crit effectiveness = 2.05 - 0.67 = 1,38

1.38/0.67 ~ 2

So my crit heals twice as effective then non-crit heals.
It means that crit heals TWICE more then non-crit even with overhealing in mind, even in LFR low damage scenario, even in 25 man with random bubble spawn and random fading.

That point of view also proved by number of hits of critical attonement and non-critical attonement

955 noncrit attonement made 50% of hps
292 crit attonement made 50% of hps (crit heals+DA)

If I had 0% crit, so all my hits were non-crit with 33% overheal, then
1247 noncrit should make 1247/955 * 50% = 65% of my hps compared to that hps I made with 25% crit.

So 25% crit increase my hps by 1.00/0.65 = 53% in 1st LFR of MSV with attonement xD
Honestly I cant assume from that analyze that 1% crit increase hps by 2% in all situations, but I found that example very transparent and easy to understand.

No doubt Crit is ok even in 25man, even with random bubble fade. In 10 man it will be better.

P.S. I start to think that crit is much better then even I thought. =)

Szeretlek · 01/31/13, 1:49 AM

Now lets see real fight.
This is my log from 2nd phase of Sha heroic.
Details for Нищебродка - 28-01 21:05 - SevereWar! - World of Logs

SS 1st place, then DA, then Attonement, then other spells.
And I can assume that DA was not spawn from PoH, because PoH usage is minimal (2.5% total hps) without SS.
So DA was spawn totally from crit heals.

If I took summary healing from Cascade for example and divide it by Crit healing and non-crit healing with WoL data like this:
Cascade 10.1% hps. Crit amount 34.1%. Hit amount 65.9%
So Crit cascade did 10.1% * 34.1% = 3,44% of total hps
Non-Crit cascade did 10.1% * 65.9% = 6,66% hps

And then I summurize all Crit and Non-crit heals, DA goes into CritHeals. SS and PWS standalone. My hps distribution looks like this:
CritHeals 35.72%
SS 30.7%
Non-CritHeals 25.88%
PWS 7.7%

Which means that w/o 25% crit I lose 35.72% of my healing and my SS will be x1.25 lower, which means I totally lose (35.72 + [30.7 - 30.7 / 1.25]) = 41,86 % of my HPS.

So without 25% crit I should do 58,14% of hps that I did.

25% crit raise my HPS by 100/58.14 =72%
Almost 3% hps increase per 1% crit. In real fight. With 50% overheal average. Without PoH spam for DA stack.

Close to 5.2 situation.

Anyone wants say something about Crit? =)

P.S. About Mastery. If I drop my 1800 rating, I lose 7.5% of mastery. That means that PWS, DA and SS will all be nerfed by (1+40%) / (1+32.5%) = 5.7% That occasion lower my hps by 0.057 * (0.077 [PWS] + 0.307 [SS] + 0.166 [DA] ) = 55% * 5.7% = 3.1%
Which means that 1 point of mastery (600 rating) gives me ~1% hps increase.

Szeretlek · 01/31/13, 5:26 AM

Hm, I made those calcs for Shekzeer fight, where I usually heavy PoH spam... And conclusions are strange.
Crit better even on those fights. Diminishes hardly, but better then mastery.

I think there is mistake somewhere.
Details for Нищебродка - 24-01 21:10 - SevereWar! - World of Logs

HPS distribution.
30% SS
25% DA (PoH DA + Crits DA, I will divide them later)
13% PoH
8% Non-Crit Attonement
7% PoM
7% Cascade
6.5% PWS
2.5% Crit Attonement
1% ermm something. lets throw it out

1st lets divide DA for PoH and non-PoH DA.
PoH healed 6 881 395 with OH 54.2%. Thats means PoH totally (w/o OH) healed 6 881 395 / 0.542 = 12 696 300

From these 12.7kk healing, there were DA spawned for 12 696 300*0.5*(1+40%) = 8 887 410 of DA from Crit and Non-Crit PoHs.
And we know that 11.6% of DA were overhealing, so finally 8 887 410 * (1-11.6%) = 7 856 470 DA out of 13 168 084 comes from PoH and 5 311 614 comes from Crit heals.

5 311 614 [DA from crits] / 52 732 282 [Total Healing Done] = 10.07% Healing Done for DA from non-PoH crits

7 856 470 [DA from PoH] * 0.758 [Hit amount for PoH] / 52 732 282 [Total Healing Done] = 11.29% Healing Done for Non-crit PoH DA
And
7 856 470 [DA from PoH] * 0.242 [Crit amount for PoH] / 52 732 282 [Total Healing Done] = 3.6% Healing Done for Crit PoH DA

If we take Crit/Hit amount for all spells we see that:
Non-Crit heals:

PoH		13%*75.8% =  9.85%
Attonement		     8.00%
PoM		7%*65.6% =   4.59%
Cascade		7%*70.2% =   4.91%
PoH DA			    11.29%

Total 38,64% from non-crit heals

Crit heals:

PoH		13%*24.2% =  3.15%
Attonement		     2.50%
PoM		7%*34.4% =   2.41%
Cascade		7%*29.8% =   2,09%
PoH DA			     3.60%
DA from Non-PoH crits       10.07%

Total 23,82% from crit heals

6.5% [PWS] + 30% [SS] + 38.64% [Non-Crit] + 23.82% [Crit] = 98,96%
I throwed 1% and 0.04% is rounding errors, so seems right.

Lets move on.
Value of 25%crit is

[Healing with Crit] / [Healing w/o Crit-Heals and Lower SS] = 100% / (100%-23.82%-(30%-30%/1.25)) = 42.49%

It means that 25% crit makes me heal 42.49% stronger, value of 600 rating (1%crit) is 42.49/25 = 1,70% hps increase per 1% crit. In real heavy PoH spam situation with overheal.

Value of my 1800 mastery [Healing with Mastery] / [Healing with 0 mastery] = 100% / (61,46% [All absorbs] * 1.325 [No mastery] /1.4 [1800 Mastery] + 38,54% [All non-absorbs] ) = 3.4%
value of 600 rating (2.5%mastery) is 3.4/7.5 * 2.5 = 1.13%

So in that real shek'zeer fight with heavy PoH spam value of Crit is higher then value of mastery. 1.7% against 1.13%

Please tell me Im wrong...

Szeretlek · 01/31/13, 6:05 AM

Oh I found my own mistake. I forget that removing crit just convert those crit heals in ordinary heals, and not just removing them completely.

Fast recalc for Sha:
If I throw out all crit, then part of my non-crit hps will raise from 25.88% to 25.88 / (1-0.25[crit%]) = 34,51%
and my summary heal should be

34.51 [from non-crit] + 30.7 / 1.25 [nerfed SS] + 7.7 [PWS] = 66,77%

So 25% crit raise my HPS from 66.77% to 100% on Sha of Fear HM. which is 100/66.77 = 50% increase.

Or 50/25 = 2% hps increase per 1% crit (or 600 rating). Twice as higher then mastery.

Fast recalc for Shek`zeer:
If I throw out all crit, then part of my non-crit hps will raise from 38.64% to 38.64 / (1-0.25[crit%]) = 51,52%
and my summary heal should be

51.52% [from non-crit] + 30% / 1.25 [nerfed SS] + 6.5 [PWS] = 82,02%

So 25% crit raise my HPS from 82,02% to 100% on Shek`Zeer HM. which is 100/82.02 = 22% increase.

Or 22/25 = 0,88% hps increase per 1% crit (or 600 rating) sligthly lower then mastery. But Crit buffs your DPS too!

Ok, now it seems absolutely right.

Havoc12 · 01/31/13, 6:22 AM

I am afraid what you are doing is

V1* - V2* = k+a* - b* which is not equal to a*, because k is the maximum value of b and thus k != b*

You have so subtract k, a non-crit heal with zero overheal. I will do it for your logs.

Also the problem with this method is that you might have more than 50% overheal, in which case it is more difficult to predict what will happen.

You are again overestimating the value of crit by a factor which depends on how accurate your estimates of V1* and V2*

This does not look like a normal overheal distribution either.

I will look over your data when I come back

Don't forget that if you are using throughput +healing CDs like proc intellect or archangel during the low overheal phases it creates an artificial bias.

Szeretlek · 01/31/13, 6:30 AM

V1* - V2* = k+a* - b* which is not equal to a*, because k is the maximum value of b and thus k != b*

You have so subtract k, a non-crit heal with zero overheal. I will do it for your logs.

I dont understand what those values mean =(

Also the problem with this method is that you might have more than 50% overheal, in which case it is more difficult to predict what will happen.

Overheal is included by WoL. There is nothing I should do with it, except spliting DA for PoH and non PoH.

Havoc12 · 01/31/13, 6:51 AM

0	93685		0
0	98135		0
0	97523		0
0	107946		0
0	107218		0
0	107674		0
32685	43440		0
0	85192		0
0	88107		0
0	87744		0
80602			40301
0	84737		0
36126	48434		0
0	84533		0
99576			49788
0	100109		0
61592	48227		6682.5
99810			49905
0	79868		0
0	88468		0
110383			55191.5
55217	21032		17092.5
0	83479		0
0	93859		0
0	84805		0
0	76760		0
64124	20101		22011.5
29184	47118		0
0	76049		0
76321			38160.5
0	83976		0
66224	18269		23977.5
85013			42506.5
0	93990		0
26514	58435		0
0	85511		0
37788	47052		0
0	93824		0
0	85302		0
0	84593		0
76398			38199
24329	51745		0
76205			38102.5
75944			37972
76818			38409
33039	52141		0
31059	79510		0
0	122235		0
53797	56740		0
87938			43969
80455			40227.5

First column is effective heal, 2nd column is overheal, 3rd column is the healing you received from that crit compared to what you would get from a normal heal.

You have a lot of 100% overheal crits, which would also give 0 with a normal heal and that sort of skews the distribution, so averaging will give you a bogus result for sure.

The total amount of healing that you got from crit is 582495.5 + aegis

In terms of only healing, crits contributed 8.4% of your total healing done by PoH and you have what 25% crit? You need to subtract every inner focus crit from there to find out the true contribution of crit.

Here is the page showing all your crits: Expression Editor - 24-01 21:10 - SevereWar! - World of Logs

Assuming you have 40% mastery your aegis created by these heals is 3181294.9 so 1.64m of that comes from crits.

In this fight PoH aegis has tiny overheal, because its easy to keep it alive and get it almost all absorbed, I suspect most of your aegis loss comes from atonement during transitions. Atonement crits during p1 just get rolled into PoH, so they are also kept alive and have low overheal.

As you can see the benefit of crit here is mostly due to aegis and thus mastery beats the pants off crit like everyone and their brother knows from experience.

This does not reflect the same situation in 5.2, where you have no way to keep aegis alive. During the first part damage is just random, so you have no clue whether you are going to get your aegis absorbed.

Here are screenies for spirit shell

ImageShack� - Online Photo and Video Hosting

59609 for normal spirit shell and 119219 for inner focus spirit shell.

Exactly twice, it should be 210/110 = 1.909090909 times which is 113799

Inner focus doubles the healing from your spirit shell exactly regardless of crit. So it amplifies the value of crit on live.

I will be really surprsied if they don't fix that in 5.2

=========================================================

Hamlet this is an excellent example for why the data posted by Jegges are entirely meaning less.

In the log I am using PoH no crit average = 20700.4, PoH crit average 32644.8, so crits are 1.6 times large and you might think they contribute 0.6% times crit rate, which is nowhere near the actual value.

You can by accident get things close, depending on the heal distribution and the percentage of over 50% overheals on crits, but there is no way of knowing. The ratio of crit to non-crit is a non-sensical value that yields no information, especially when averaged over all spells. A ratio of 2:1 or more means you can be 100% sure you have an artifact.

Havoc12 · 01/31/13, 7:46 AM

I am making a new post for this because it is important. I tested inner focus with spirit shell on the PTR and it is fantastically bugged.

What happens is spirit shell PoH does not consume inner focus but benefits from it. All the other spirit shell PoHs don't benefit from inner focus but dont consume the buf. After spirit shell ends all PoHs you cast benefit from inner focus without consuming it. Greater heals also benefit but don't consume the buff but flash heal does....

This is really broken you can pop IF during spirit shell and then have all your non spirit shell PoH and Gheals crit.

but IF spirit shell poh = 71k, non spirit shell PoH = 35782, so it still doubles it!! I need to go file a bug report....

Szeretlek · 01/31/13, 7:50 AM

First of all, I totally forget that I go with mastery build for SheekZeer HM, so it was likely 15% crit and 57.5% mastery.
That explain so big PWS and DA. It again values crit higher then mastery =(
Sorry for misleading here.

Havoc12, So I worked with that piece of data myself, your numbers are true, but you forget DA. You mention it, but dont include for some reasons.

Crit profit from non-crit PoH is 1 577 141 (sum of 1st column), or 582 495 if we substract normal heals.
But lets add DA with 11,6% OH,

Total CritHeals (with overheal) = (sum of 1st and 2nd column) = 4 544 707
DA that created by that criticals = 4 544 707 * 0.5 * 1.575 = 3 578 957
DA with 11.6% OH = 3 578 957 * (1-11.6) = 3 163 798

So totally 3 163 798 DA was consumed by all that crits.
Totally crit contribution = 3 163 798 + 582 49 = 3 746 293
Or 7.1% total healing done
PoH total heal (with DA from non-crit) = 27,89% of healing done

Crit contribution for PoH = 7.1/27.89 = 25.46% healing done by PoH were made by crit heals.
Or 25.46/15 ~ 1.7% hps increase per 1% crit.

Crit still prevail!

Btw, I use IF only under SS. So no need to count it.

Assuming you have 40% mastery your aegis created by these heals is 3181294.9 so 1.64m of that comes from crits.

why drop from 3.2kk to 1.64kk??? Is there any reason, because we already analyze only critical heals.

Havoc12 · 01/31/13, 7:57 AM

Originally Posted by Szeretlek

Oh, I see. So I worked with that piece of data myself, your numbers are true, but you forget DA. You mention it, but dont include for some reasons.

Crit profit from non-crit PoH is 1 577 141 (sum of 1st column), or 582 495 if we substract normal heals.
But lets add DA with 11,6% OH, 3 163 798 DA was created by all that crits.

Or 3 746 293 summary crit contribution in PoH. Or 7.1% total healing done
PoH total heal (with DA from non-crit) = 27,89%

Crit contribution for PoH = 7.1/27.89 = 25.46% healing done by PoH were made by crit heals.
Or 25.46/25 ~ 1% hps increase per 1% crit. So it prove my calculations.

Btw, I use IF only under SS. So no need to count it.

Nope only half of that aegis is due to crits, I added a note on the amount healed by aegis (no need to remove overheal) some time ago

. I didn't include it because its difficult to quantify wastage on it.

There is no reason to remove overheal because most the aegis overheal is artificially buffed from aegis wastage during transitions.

Its about 0.8% more healing per crit point. But tbh you picked the wrong fight for this. You should really look at garalon hc.

This does not reflect the situation in 5.2 though, as I explained in the post above.