3.1 PTR - Page 13

Elimbras · 03/08/09, 11:08 AM

For 1 point, I'm not sure we are likely to have short OOFSR time, neither that we will have significant HC uptime in OOFSR (both are obviously correlated).
It's true that we can construct models were it happens. But basically, these are models were you have short breaks in the fights. Without any clearcast, this means that we don't throw any heal for more than 5 seconds. You need a specific damage structure for it to be valid.

The second point is that, unlike the HC Uptime in OOFSR, which can be theoretically 100%, the OOFSR time while HC can be more than 37.5% (3s every 8s), and is almost surely lower. When you want to evaluate the impact of spirit or crit on regen, you first compute the mean total time in HC (using the formula with constant intervals between cast, or with the distribution of intervals). Then you have to attribute which portion of HC is OOFSR. For constant total HC uptime and OOFSR uptime, the difference between 37.5% HC in OOFSR and 0% HC in OOFSR is, assuming 15% OOFSR (which is likely to be overestimated without phase transition (HC uptime close to zero) nor clearcast) :

$max = regen_{OOFSR} + regen_{IFSR} + regen_{OFFSR+HC} + regen_{IFSR+HC}$
$max= C * Oofsr + C * 0.5 * (1 - Oofsr) + C * HC * (0.5 * 0.375 + 0.25 * 0.625) = C * 0.5 * (1 + Oofsr) + C * HC * 0.34375$
$min = C * Oofsr + C * 0.5 * (1 - Oofsr) + C * HC * 0.25 = C * 0.5 * (1 + Oofsr) + C * HC * 0.25$

The impact on total regen (from spirit) is basically one tenth of HC uptime in the coefficient (counting 50% uptime, this is less than 10% total impact (the difference is 0.05, and the total coefficient is more than 0.5).
In practice, for the difference to be maximal, you need a 0 OOFSR uptime (to keep the global coefficient low), and high HC uptime (to keep the difference high) and a big part of your HC uptime while in OOFSR (which contradicts both other assumptions). So it will be lower.

Sureall · 03/08/09, 11:52 AM

Originally Posted by Tweaksys

Last night on PTR since it was down for so long into the testing much of my guild gave up hope for a raid but we did get a 10 man in there. Again we were without replenishment like the last time we tested it (ironic that Blizz now says replenishment is integral and expected and we are having our worst luck with getting replenish on PTR raids). I was Healing tank on the big guy then tank on the medium guy after big guy went down as well as healing around on raid. Mana was only tight one attempt when we got unlucky on interrupts (we had resto shammies doing most of the interrupts). Our group set up was COMPLETELY unideal, so much so as to be about the worst comp we could have had really. Even with that we basically had a kill on the ten (mob got down to less then a percent an enraged on us because a mage went ADD and forgot to get outta an overload around 10%).
Anyways... even without replenish I bounced gear around a little adding some haste changing for some crit and bouncing between Soul of the Dead and Forethought Talisman just to try things out. I was never completely dry but you can very much feel the difference between live and PTR. I was using Hymn of Hope when the big dude died which helped me a bit but we had 2 moonkins specced for zero regen so they were always stopping dps because they ran oom -.-. Used shadowfiend half the attempts (only once did I need it though cause a shammy died early and couldn't stand). Overall, at least for this fight, mana feels about right for me as disc (haven't ran holy at all because I am solely a tank healer in my guild between me and a holy pally). If fights get much more mana intensive I would have to re-evaluate some gear choices and regem/chant a little probably. Do have the WWS if anyone wants to see it but it does not recognize Council as a boss yet so it all reads as a bunch of trash pulls lumped together =\ .

Edit: I was running with http://talent.mmo-champion.com/?prie...0&version=9658 as my build last night. First time trying extensive PTR healing without Imp Healing. I have also been running with Frost Wyrm because my healing is lower then it will be on live but I made 150sh dstilled wisdom for live for 3.1 release.

Edit 2: Linked WWS Wow Web Stats

Your flash heal crits seem nice and high from this parse. Did you find yourself healing riad members below 50% a decent amount of the time?

Tweaksys · 03/08/09, 1:00 PM

Because of the spike that happens during a Fusion Punch and that I knew it had a random timer on it I Flashed the tank between fusion punches then Penance on Fusion Punch after the dispell. As well flash is typically what I used to heal random raid members because they were rarely that hurt, but did use Penance to heal the other OT's. No damage in the fight was THAT spikey except fusion punch really, it was all manageable with FH. I could have used Penance every timer on MT healing but again I always worried that if I didn't have penance on a Fusion punch it meant dead tank.
When the first dude died my responsibilities did shift to as much raid healing as tank healing though because the damage is so small for the rest of the fight really (minus slow people in death runes or ADD people in overloads). We had plenty of interrupt for whirl so even that wasn't a big spike.

Squeakster · 03/08/09, 2:28 PM

Originally Posted by Sureall

Your flash heal crits seem nice and high from this parse. Did you find yourself healing riad members below 50% a decent amount of the time?

Yeah his flash heal crit rate is high (41%), but so is his penance crit rate (42%) and PoM crit rate (40%), so it looks to me as if the extra crit from Improved Flash Heal is either broken or does not come into play enough to matter at all.

(Edit) I am impressed by your overheal percentages, those are significantly lower than I am seeing from all the healers in my raids on live. Tweaksys, have you noticed a significant improvement in overhealing percentages for all of your guilds healers from your live raids to the PTR raids?

Alv!ra · 03/08/09, 4:18 PM

Originally Posted by Squeakster

(Edit) I am impressed by your overheal percentages, those are significantly lower than I am seeing from all the healers in my raids on live. Tweaksys, have you noticed a significant improvement in overhealing percentages for all of your guilds healers from your live raids to the PTR raids?

I think that'll be natural once mana is an issue again. On live, there's no reason whatsoever to not be spamhealing constantly.

Tweaksys · 03/08/09, 9:31 PM

True, I think people may have to redig out the recount and work on their overhealing percentages as one thing, but I don't think that is all. I think the damage is just SOOOO much more predictable on ptr, at least so far. As far as the content I have tested that has been live on US servers (all but the forge guy we really didn't get a chance on him) has been pretty predictable damage. Typically its been one-two healers that compliment each other's heals and you are watching other parts of the raid as well. So the healing is being spread out and the bigger hits are coming at predictable intervals that you can feel out. Maybe other people have experienced this differently but for me that seems to be the flow so far. When heals aren't totally random (and of course grace lasts longer so I have less time wasting flashes to stack it as well) you do less over heal. I am also using FH WAY more then I ever did on live, it really seems such an important spell in the PTR content. Could be just some observations of mine.
If I think of more random thoughts on it I will maybe post an edit. But ya, it is definitely a collection of things both out of force and out of composition of content imo naturally working together so far in short.

AtrociousAaron · 03/08/09, 11:43 PM

Okay, so I am currently specced flash heals for raids, but in the upcoming patch I have noticed that my specc is going to get the shaft. So I am just wondering what way should I go. I can still specc flash heal, but I have to go deep into disc to get imp FH, and then I cant go into holy to get CoH which is always nice.

Should I just say heck with it and go complete into disc or holy. Im so torn between the 3 different ways.

Ive also got a somewhat decent shadow set, so shadow is also an option, but Im less inclined to go that way and my doesnt guild want me to.

So my main question is which specc looks to be the over-all best for heals in 3.1. Or do they all have their perks.

Lambi · 03/09/09, 12:18 AM

Originally Posted by AtrociousAaron

Okay, so I am currently specced flash heals for raids, but in the upcoming patch I have noticed that my specc is going to get the shaft. So I am just wondering what way should I go. I can still specc flash heal, but I have to go deep into disc to get imp FH, and then I cant go into holy to get CoH which is always nice.

Should I just say heck with it and go complete into disc or holy. Im so torn between the 3 different ways.

Ive also got a somewhat decent shadow set, so shadow is also an option, but Im less inclined to go that way and my doesnt guild want me to.

So my main question is which specc looks to be the over-all best for heals in 3.1. Or do they all have their perks.

It's actually pretty funny how it works. The deepest specialized specs (deep disc and deep holy) are the best allround specs by far. It's when you mix the trees the most you get the encounter specific talentspecs. So far I have to say I prefer disc as the most fun spec, since you can shield and get a hasted on demand heal straight after.

If you have mana problems I would recommend you going disc with a flash heal glyph and later when it comes out, a penance glyph.

Tweaksys · 03/09/09, 12:52 AM

It really depends on what role your guild is going to let you fill as well. I wouldn't want to be holy to tank heal after my extensive experience as disc on PTR ulduar content. That being said I also think disc feels very lack luster for much raid healing (though I have not tried holy on this PTR, I can't imagine disc competing on raid healing at all, even with the redux cd on PWS though a lot of content does allow offhealing the raid even while tank healing atm.)

If I was going to be primarily tank healing and off healing the raid when tanks arent getting destroyed I would go disc and then yes, have imp flash with flash glyph is just amazing from my testing. If I was going to be filling a raid healing spot then I would go holy because it is obviously gonna have the higher HPS raid wide compared to disc (though disc may be a little more viable in a preventative way if we ever get our PWB.)

All that being said... I really don't feel there is a place for 2 disc priests in a raid. A pally and a disc priest make up about the best combo for tank healing really, good synergy, and if need be roll some druid hots. Disc just loses a lot of purpose if u have more then one, at least imo, whereas for raid healing it is not silly to have 2 holy priests raid healing and doing a little off healing on a tank in a pinch in a 25. Holy won't last long on a tank though as these boards have shown. Hope that answers it a bit for ya. They aren't better and best really, its just two different styles of play for two different purposes imo, as distinct as different healing classes.

Observer · 03/09/09, 12:49 PM

A second disc priest could forego (most of) the PW:S talents and get Improved Renew and Greater Heal which are generally not going to be taken by heavy disc. Then perhaps you end up with that priest as dedicated tank healer (keep Renew on the tank and cancel-cast Greater Heal when Penance is on CD) and the other as tank healer/shielder, but also shield/flash heal tosser.

It would be an odd setup, but the crit from Renewed Hope still applies even if it's not your Weakened Soul effect (or should, in theory, I don't know if it's ever actually been tested).

Douldoul · 03/09/09, 1:22 PM

Question to the ptr guys (thanks for all the details you are digging) :
Contrary to the live, does the crit healing of the PWS glyph proc Inspiration and DA as a normal heal? Or is it still bug? (Or intented? but I never see things on this).

Also :
I see some Disc template with divine fury here. Does GH still usefull as we have a cheap FH with Improved Flash Heal?
For my playstyle I always keeping renew up on tank and so I think Improved renew is better than DF. What is your opinion on this with an experience on PTR?
And as a complement : what about the pushback of the fights in Ulduar ? Healing focus is still unecessary?

Tweaksys · 03/09/09, 1:42 PM

The last couple nights I have been running without Divine Fury and actually placed the points in Spell Warding. Could be gimmicky but I haven't been disappointed yet (a handful of the bosses so far have raid wide spell damage). I have been fine without any real gheal unless its a lot of downtime and I want to crank one out with an inner focus hoping for a big DA. Flashes and Penances are really the bread and butter with the content so far and lots of PWS on tank and other people to keep BT rolling. Still no sign of that pesky Penance glyph which is unfortunate since that will be a part of any disc priest once its found. I actually don't roll renews much at all atm, that could change, but when I am solo healing it just hasn't been necessary and if I am not alone its with a spell spamming pally leaving no real time for a hot to tick. I could be way off on this but I just haven't found it to be beneficial.

AtrociousAaron · 03/09/09, 2:49 PM

Ugh, okay well now I just have to wait for my guild to tell me what way they want me to specc. Whenever I ask they are usually just like "Well a disc priest would be nice, but holy is good too" and thats all they say. Thanks guys.

But there also is the fact that we have a total of maybe 3 priest healers in my guild, including me. One is out MTs alt and the other rarely plays. Our other healers consist of mainly shammy and pallies.

meddle · 03/09/09, 3:36 PM

Originally Posted by AtrociousAaron

Ugh, okay well now I just have to wait for my guild to tell me what way they want me to specc. Whenever I ask they are usually just like "Well a disc priest would be nice, but holy is good too" and thats all they say. Thanks guys.

But there also is the fact that we have a total of maybe 3 priest healers in my guild, including me. One is out MTs alt and the other rarely plays. Our other healers consist of mainly shammy and pallies.

To state the obvious: with the advent of dual specialization, you won't have to be pigeonholed into either spec. Just change for the encounter. That's what you can propose to your guild (the only outlying issue there being that Disc and Holy have slightly different gear preferences).

I feel like the answer to a lot of healing problems in Ulduar will be to just "spec switch" from encounter to encounter, though that may be a vast oversimplification of the solution. But like many Priests in this thread have already mentioned, a "regen spec" (going down the Shadow tree for 2 minute Shadowfiends) might be a candidate for one of our specs come dual speccing, as least while we are still wearing our Naxxramas epics.

Tweaksys · 03/09/09, 4:05 PM

To be honest my two specs will be two disc variants going varying depths into holy and disc. One will probably have Imp Healing the other will go just deep enough for inspiration(which has a +1 that I will put into Desp Prayer).

Allesin · 03/09/09, 5:29 PM

With dual speccing, I am anticipating one deep holy spec, and the other a deep disc spec to help adjust to the needs of various encounters. A regen spec would be nice, but (and I may be completely off-base here) gimmick specs such as 2 minute shadowfiends are just that - gimmick specs. They may have uses for a while, but being able to tailor your spec to your assignment (raid/tank) depending upon raid balance and the particular encounter strikes me as much more advantageous.

We will have to relearn mana management at some point, whether it's because we never learned it, or simply because Blizzard has totally changed the mechanics we used and loved (goodbye IF + CC procs). A shadowfiend spec will simply ignore learning those mechanics for a while.

Havoc12 · 03/09/09, 9:05 PM

Originally Posted by Elimbras

For 1 point, I'm not sure we are likely to have short OOFSR time, neither that we will have significant HC uptime in OOFSR (both are obviously correlated).
It's true that we can construct models were it happens. But basically, these are models were you have short breaks in the fights. Without any clearcast, this means that we don't throw any heal for more than 5 seconds. You need a specific damage structure for it to be valid.

The second point is that, unlike the HC Uptime in OOFSR, which can be theoretically 100%, the OOFSR time while HC can be more than 37.5% (3s every 8s), and is almost surely lower. When you want to evaluate the impact of spirit or crit on regen, you first compute the mean total time in HC (using the formula with constant intervals between cast, or with the distribution of intervals). Then you have to attribute which portion of HC is OOFSR. For constant total HC uptime and OOFSR uptime, the difference between 37.5% HC in OOFSR and 0% HC in OOFSR is, assuming 15% OOFSR (which is likely to be overestimated without phase transition (HC uptime close to zero) nor clearcast) :

$max = regen_{OOFSR} + regen_{IFSR} + regen_{OFFSR+HC} + regen_{IFSR+HC}$
$max= C * Oofsr + C * 0.5 * (1 - Oofsr) + C * HC * (0.5 * 0.375 + 0.25 * 0.625) = C * 0.5 * (1 + Oofsr) + C * HC * 0.34375$
$min = C * Oofsr + C * 0.5 * (1 - Oofsr) + C * HC * 0.25 = C * 0.5 * (1 + Oofsr) + C * HC * 0.25$

The impact on total regen (from spirit) is basically one tenth of HC uptime in the coefficient (counting 50% uptime, this is less than 10% total impact (the difference is 0.05, and the total coefficient is more than 0.5).
In practice, for the difference to be maximal, you need a 0 OOFSR uptime (to keep the global coefficient low), and high HC uptime (to keep the difference high) and a big part of your HC uptime while in OOFSR (which contradicts both other assumptions). So it will be lower.

I really dislike the way you are doing this and I dont agree with your conclusions. I dont know how you came up with these formulas either, they dont look right to me. Especially the bit where you are trying to calculate the contribution of HC procs and the percentage of that time spend in ooFSR, makes no sense to me at all. What is important is how much mana HC returns while you are casting since the talents it replaced gave mana only while casating.

Worrying about modelling phase transition is pointless. When you are not casting and there is no DPS on the boss spirit has the best return.

Breaking each HC proc into iFSR and ooFSR makes zero sense to me. This is impossible to calculate.

Depending on your casting pattern you can expect 40-60% uptime, the exact uptime of HC during ooFSR or iFSR is impossible to calculate. For values of <10% ooFSR (which is what you can expect during combat) it is safe to assume that uptimes are the same.

The regen formula without HC is

$(1+0.5*ooFSR)*spi*\sqrt{int}$

The regen formula with HC is

$(1+0.5*ooFSR)(1+HC_{uptime}*0.5)*spi*\sqrt{int}$

at uptime of 50% the total contribution of HC is 25% of your total regen. 5% ooFSR time increases regen by 2.5%, so even if assuming equal uptime is not accurate deviations are small.

Thus the impact of spirit/int or crit are really very simple to calculate. Crit influences uptime and HC increases regen by 0.5*uptime. That means it increases the value of spirit regen by 0.5*uptime. The relationship between crit and uptime is explained with as much accuracy as is necessary by the (1-(1-C)^n) formula. At 25% crit average uptime should be around 50% meaning that HC increases spirit regen by 25%.

----------------------------------------------------------------------------------------------------------------------------

The main problem with not specing divine fury in my opinion is it limits both your burst and the max HPS you can keep on the tank.

constantius · 03/09/09, 10:17 PM

Did anyone notice Serendipity being bugged in the last 24 hours? I did a few pulls of the X-whatever boss, and never got a single proc, spec'd 14/57/0.

Wow Web Stats

Tweaksys · 03/09/09, 10:50 PM

Aye, Serendipity is currently broken. They nerfed it and somehow forgot to even put the effect on it... so a nerf and a break all in one day!

Deconstructor was absurdly easy... We did one attempt with 20 on PvP, then he got pulled down so we went to PvE and one shot him. All these fights are just crazy easy and I am hoping they get tuned up. I recorded WWS but its nothing fancy at all really, like there just isn't that much damage going out (unless you are healing peoplg who get grav bomb). Lotta PWS spamming in downtime before adds come out and when earthquake is going on really is all.
WWS: Wow Web Stats

Elimbras · 03/09/09, 11:18 PM

Originally Posted by Havoc12

I dont know how you came up with these formulas either, they dont look right to me. Especially the bit where you are trying to calculate the contribution of HC procs and the percentage of that time spend in ooFSR, makes no sense to me at all. What is important is how much mana HC returns while you are casting since the talents it replaced gave mana only while casting.

I realised that my post was not clear.
But I just broke the spirit regen in four part :
- normal regen OOFSR assuming 0% IHC
- normal regen IFSR assuming 0% IHC
- additional regen from HC OOFSR
- additional regen from HC IFSR.

Worrying about modelling phase transition is pointless. When you are not casting and there is no DPS on the boss spirit has the best return. Breaking each HC proc into iFSR and ooFSR makes zero sense to me. This is impossible to calculate.
Depending on your casting pattern you can expect 40-60% uptime, the exact uptime of HC during ooFSR or iFSR is impossible to calculate. For values of <10% ooFSR (which is what you can expect during combat) it is safe to assume that uptimes are the same.

That's basically all the point in the recent posts Promethia and I made.
Yes, you can assume that the HC uptime is the same IFSR and ooFSR.
But that's an assumption (but assuming zero HC uptime while ooFSR seems as valid for me). And in fact, this distribution of HC between IFSR and ooFSR do have impact on regen. Question is to quantify it, in order to know wether we can ignore it or not.

The regen formula without HC is
$(1+0.5*ooFSR)*spi*\sqrt{int}$

I guess you want to write $(0.5+0.5*ooFSR)*spi*\sqrt{int}$ .
Otherway, I don't understand your (1+0.5*ooFSR).
Put C = spi * sqrt(int), and you come back to the formula I have without HC.

The regen formula with HC is

$(1+0.5*ooFSR)(1+HC_{uptime}*0.5)*spi*\sqrt{int}$

Once again, you assume that you have the same IHC uptime iFSR and ooFSR.
In practice, it can't be the case.
Probability to have IHC up is 1 - (1-C)^n, if you have cast n spells that triggers IHC in the last 8s.
For ooFSR, there is 5 out these 8s where you didn't cast any spell. This leaves to n=3 at most, if you flash / renew with 50% haste. Or more realistically, n=2 (or even 1 or 0).
In realistic situations, iFSR, you do have cast at least 3 renew / flashs / gh in the last 8 second. Or this can be more.

at uptime of 50% the total contribution of HC is 25% of your total regen. 5% ooFSR time increases regen by 2.5%, so even if assuming equal uptime is not accurate deviations are small.
Thus the impact of spirit/int or crit are really very simple to calculate. Crit influences uptime and HC increases regen by 0.5*uptime. That means it increases the value of spirit regen by 0.5*uptime. The relationship between crit and uptime is explained with as much accuracy as is necessary by the (1-(1-C)^n) formula. At 25% crit average uptime should be around 50% meaning that HC increases spirit regen by 25%.

If you are telling us that ooFSR time will be low, and hence, we can ignore its interaction with IHC, it's fine. I'm just doing the same, with more flexibility in the scenario.
That's what I've said in the recent post.
Either you don't have any phase transition, and ooFSR is low, and you don't care about it's interaction with IHC.
Or you do have phase transition, and IHC has near zero uptime during ooFSR time.

constantius · 03/09/09, 11:36 PM

Originally Posted by Tweaksys

Aye, Serendipity is currently broken. They nerfed it and somehow forgot to even put the effect on it... so a nerf and a break all in one day!

Deconstructor was absurdly easy... We did one attempt with 20 on PvP, then he got pulled down so we went to PvE and one shot him. All these fights are just crazy easy and I am hoping they get tuned up. I recorded WWS but its nothing fancy at all really, like there just isn't that much damage going out (unless you are healing peoplg who get grav bomb). Lotta PWS spamming in downtime before adds come out and when earthquake is going on really is all.
WWS: Wow Web Stats

Easy mode is easy. Hard mode is ... hard. I think that's the mark of a well-designed fight, imo. At least, for the first 6-8 bosses out of 13-odd.

Tweaksys · 03/10/09, 12:55 AM

Fair enough, we didn't have the raiders or time to try hard mode sadly. They wanted heroic to get as many testing it for live prep as possible so we took some guild friends in and actually had 2 sit out afk and offline because of the impending server downage. Can't wait to test the hard though =\.

P.S. I figured I might add a small section on the new lightweave embroidery here. On my PvP server version I put it on my cloak to test proc rate and internal CD. After checking my WWS on some boss fights it seems the internal is right at 45s as I never had it proc faster then that but did have it proc right at 45s very often. During in combat time it seemed to come up right at or within a couple seconds of the 45s mark, basically saying high proc rate. So the Lightweave seemd to be up basically 1/3 of the time since it lasts 15 seconds with a 45 second cooldown (up for fifteen, down for 30 then back up). Depending on the prevalence of mana intensive fights I really think I may start with this cloak enchant in 3.1 over the mana one (of course I have 2 moonshroud spec tailor's atm so spare cloth is not really an issue for live testing when more bosses are out.)

Havoc12 · 03/10/09, 11:23 AM

Originally Posted by Elimbras

That's basically all the point in the recent posts Promethia and I made.
Yes, you can assume that the HC uptime is the same IFSR and ooFSR.
But that's an assumption (but assuming zero HC uptime while ooFSR seems as valid for me). And in fact, this distribution of HC between IFSR and ooFSR do have impact on regen. Question is to quantify it, in order to know wether we can ignore it or not.

I guess you want to write $(0.5+0.5*ooFSR)*spi*\sqrt{int}$ .
Otherway, I don't understand your (1+0.5*ooFSR).
Put C = spi * sqrt(int), and you come back to the formula I have without HC.

Once again, you assume that you have the same IHC uptime iFSR and ooFSR.
In practice, it can't be the case.
Probability to have IHC up is 1 - (1-C)^n, if you have cast n spells that triggers IHC in the last 8s.
For ooFSR, there is 5 out these 8s where you didn't cast any spell. This leaves to n=3 at most, if you flash / renew with 50% haste. Or more realistically, n=2 (or even 1 or 0).
In realistic situations, iFSR, you do have cast at least 3 renew / flashs / gh in the last 8 second. Or this can be more.

Actually I meant to write FSR not ooFSR and its - not +:

$(1-0.5*FSR)*spi*\sqrt{int}$

That is exactly where I think you are going wrong. You are looking at the proportion of HC procs spent in the ooFSR time but that is NOT HC uptime. Uptime is the % of time during which you have HC up.

Read my previous post. The uptime of HC during ooFSR can range from 0-100% and it depends on playstyle. What you are saying is that from every HC proc you can get at best 3 sec of ooFSR time, but that does not mean HC uptime is 3 out of 8, that is a blatant mistake. 3 out 8 is the simply the maximum possible proportion of TOTAL HC time spent outside the FSR, but this has nothing do with the uptime variable in the formula. Although you can use this method to calculate total regen, you end up with something that has no practical value, as it is completely impossible to determine the proportion of HC time spent in the ooFSR, unless you make an assuption about uptime of HC during ooFSR, which pretty much puts you back where you started. Basically the simplest expression you can use, that does not have any assumptions implicit is this

$0.5*FSR*(1+0.5*HC_{uptimeFSR})*regen + (1-FSR)*(1+0.5HC_{uptimeoFSR})*regen$

However in order to get any real values out of this you will have to make an assumption on how much uptime is. Using (1-(1-c)^n) is probably the most reasonable approximation you can make if you dont want to try and model a particular healing rotation.

If you want to compare the importance of correctly estimating ooFSR HC uptime simply stick values into this formula.

The reason why I disliked your method, was because you are making a very big assumption in a formula that effectively tries to model the impact of assumptions. You have basically assumed that the model I presented in my previous post, which was to take an ooFSR break of 3 seconds everytime you proc HC. This is not a reasonable assumption to make.

Remember that in order to get any ooFSR time during HC, you need to stop casting at least 1.5 seconds after you get the HC proc, but you can begin casting a spell before the 8 seconds of HC are up so that it lands when HC expires.

Elimbras · 03/10/09, 11:56 AM

Originally Posted by Havoc12

That is exactly where I think you are going wrong. You are looking at the proportion of HC procs spent in the ooFSR time but that is NOT HC uptime. Uptime is the % of time during which you have HC up.
Read my previous post. The uptime of HC during ooFSR can range from 0-100% and it depends on playstyle. What you are saying is that from every HC proc you can get at best 3 sec of ooFSR time, but that does not mean HC uptime is 3 out of 8, that is a blatant mistake. 3 out 8 is the simply the maximum possible proportion of TOTAL HC time spent outside the FSR, but that has no real bearing in the formula.

Basically, we do have two sums. First one is iFSR / ooFSR, second one is HC / not HC. Key point is that you can invert the summation order.

If you prefer it mathematically, you can write :
$regen = \sum_{i=0}^1 P(i) \sum_{k=0}^1 P(k|i) * regen_{i,k}$ ,
where i indicates whether we are iFSR (i=0) or iiFSR (i=1), and k whether HC is up (k=1) or down (k=0). regen_{i,k} is the regen in the considered case. P(X|Y) is the conditionnal probability of X knowing Y.
That is the way you consider the regen : first determine if you're iFSR or ooFSR, then whether HC is up or down.

But mathematically, we can invert the sums and probabilities, and the following also holds :
$regen = \sum_{k=0}^1 P(k) \sum_{i=0}^1 P(i|k) * regen_{i,k}$
That's the formula I've written.
Technically, I've considered that base regen is always there, and that HC provides only additional regen. So, it would read :
$regen = \left[1 - 0.5 * P(FSR|not HC)\right] * spi * \sqrt{int} + P(HC) * \left[ P(ooFSR|HC) * spi * \sqrt{int} * 0.5 + P(iFSR|HC) * spi * \sqrt{int} * 0.25 \right]$

And here, you can see that the proportion of HC time that is ooFSR do matter in the expression. And luckily, we can bound it, so the deviation to the mean case can be bounded.

Note that ultimately, all what we want to compute is the four values P(FSR , HC), P(FSR, not HC), P(ooFSR, HC) and P(ooFSR, not HC).
We could also just measure these values. they determine (in addition to spirit and int) everythink.

Lambi · 03/10/09, 1:59 PM

Originally Posted by constantius

Easy mode is easy. Hard mode is ... hard. I think that's the mark of a well-designed fight, imo. At least, for the first 6-8 bosses out of 13-odd.

I'm actually starting to think that Blizzard are being very smart this PTR test and they're working the fights with easy numbers just so alot of guilds can experience and test the various mechanics (which should be the hard things testing), and then they'll tune the numbers for live.