I like that idea. drinking and trying to get a full bar between pulls is a pain at the moment.

I don't think it's a big deal if Barrier is tied to Rapture, the ability is obviously going to have some sort of cooldown (maybe even the same Weakened Soul).

As far as water. Something needs to be done, but I don't think Blizzard will change to percentage based increases. Not in that manner. The primary reason is because they're stubborn and want us to keep "upgrading water". Now that said, there could be a hybrid solution. Percentage based with caps.

e.g. 2% mana returned per a second up to a maximum of 1000 for 30 seconds. Or something of the sort.

This might sound whiney, but I really, really, don't want to be the guy that forces a 5-man to slow down after every pull in a heroic or 10 man. That's probably the one thing that bothers me the most, but I am going to play wait and see with the PTR.

It's looking very good, seems like I'm gonna enjoy my priest again in 3.1 (holypriest is just terrible gameplay at the moment).

The big question is what they are doing with Serendipity, I know I'm not the first one wondering about this! If they just remove the mana refund deep holy is looking at a big nerf to mana efficency on single target healing. I would much rather have a fun on demand haste-talent than a no-brainer mana refund-talent, but I thought we were kind of balanced around it now.

# Serendipity – this talent now reduces the cast time of Greater Heal and Prayer of Healing when Binding Heal or Flash Heal are cast.

Have I missed something or is this unclear?

I think you misunderstand the blue post Bjork. The way I see it is that that is going to be an addition to the talent. Getting rid of that talent or just changing it and removing the mana back from it would be a very big dent in the holy tree.

It's amazing what the word "also" or lack thereof does to a sentence. I understand it to be getting rid of the mana return, partially because it just sounds like it's getting replaced, and partially because of what they've said about priests never running out of mana.

They're trying to to nerf regen just enough and it sounds like getting rid of the current serendipity might be part of their plans. If Clearcasting becomes a direct mana gain, as they've mentioned they're thinking about, it makes a little more sense.

I don't have a clue as to how successful they're going to be with finding the sweet spot. It could be possible that the real effect would be healers are forced to go oom without enough replenishment. Whether intended or not, it seems that would be an easy situation to be in. Being efficient would mean you could get away with less replenishment givers. Being inefficient would mean bringing another healer into the raid, or bringing another replenishment buff.

I want you to imagine the wording of a talent which has one effect when you cast Greater Heal/Flash Heal, and a second, orthogonal effect when Flash Heal/Binding Heal are cast. Is your head hurting yet? If so, then you can start to see why people are suspecting it's a replacement. That, by the way, would be a bigger regen nerf than any of the spirit changes, or even any of the potential clearcast changes.

I guess you'll be better off with the [Majestic Dragon Figurine] after that change. If you don't get regen breaks then the figurine is already the better trinket, because it simply gives more Spirit.

Glass: 84 + 20/120 * 336 = 140 Spirit (if always used on cooldown)
Figurine: 180 Spirit permanently

The glass would only be better if the fight was shorter than X = 20 * 336 / 96 = 70 seconds.

Of course there may be incidents when the Figurine can't be sustained because nothing can be cast. But I haven't come across a fight like that yet.

And there everybody was laughing when I spent DKP on the Figurine.

For me with ~ 2600 spellpower raid-buffed, and 24% crit (so close to your values), I get base heals of:
Flash: 5983
Renew: 7908

so Renew heals for 32% more. It does still scale better, you're right, but it's not 50% right off the bat. I imagine this is mostly due to the Glyph being broken, but also due to the typo in your calculation for Flash heal value.

Also, no one was arguing that Renew doesn't scale better. It just doesn't scale fast enough.

I apologize for the specific notes against your post: please don't consider this trolling at all. You've just made a couple of fairly large flaws in the math, and I wanted to show you exactly where.

Here's my take on the HpS issue: let's assume we have 21% raid-buffed haste. This drops our cast-time of Flash to 1.24 seconds, and our GCD (or effective cast-time of Renew) to the same. We'll assume Glyph'd, and use the values I gave for real-world ~ 2600 spellpower effects (@ 24% crit):

Flash: 7908
Renew: 5983

If you assume all 4 ticks of a Glyph'd Renew tick, then you do get 7908 healing for the mana cost of 17% base (-any points spent in Mental Agility, which I'll assume are nil). Compare this to Flash Heal, and consider the cases (with base mana being set to 3863):

Renew:
1) All 4 ticks do not overheal. Total HpM is 12.0

Exclusive:
1) Flash Heal does not overheal. Total value of healing is 5983 on average, mana cost is 16.4% base (assume Glyph for -10%). HpM is 5983 / (.164*3863) = 9.44 HpM.
2) Flash Heal does overheal. Serendipity procs. Total value of healing is <5983, mana cost is 0.75*16.4%, or (ballpark) of 6.3 HpM.

Non-exclusive:
1) Flash Heal crits. This crit does not proc IHC or SoL. Same case as #1.
2) Flash Heal crits. It procs IHC or SoL. In the case of SoL, it's a nice big boost to HpM, since you get a free Flash Heal for 0 mana, which raises the average HpM by a chunk. In the case of IHC, it's even better, as you get that free Flash Heal, it can possibly crit, and you also get some haste, increasing HpS.

The cases for #2 are a bit tricky to model, but I did do it once before in this thread. I believe it worked out to ~ 10% increase in HpS on average, and a ~20% increase in HpM.

Ballpark values:
If Renew ticks all 4 times, it comes out to a HpM of 12.0. Its HpS is 6.4k (assuming 21% haste). Comparatively, Flash Heal, if it does not overheal, comes out to a HpM of ~ 11.4, with HpS of 4.9k (again, assuming 21% haste). If it does overheal, the model gets messy, but HpM depends heavily on how much it overheals, and HpS is unaffected.

So basically, if Renew ticks all 4 times, you just slightly beat the HpM of Flash Heal, and you *do* beat the HpS of Flash Heal. However, if you start assuming that instead of using IHC on Flash, you immediately switch and do 2 hasted GHeals, it blows Renew out of the water, both for HpM and HpS.

Hope the math made sense, and feel free to counter with some fresh understanding. Rolling Renews is obviously one of the highest HpS things we can do (my gut says that spamming PoH is higher, if only slightly). However, if you're just looking at the case of "let's heal one person, down 5k HP", you're almost always better off casting Flash Heal to get procs, and accept the slight loss in HpM that *may* occur, than to use Renew and almost certainly see the heal get sniped.

I'd also like to again point out to Havoc that Renew's HPS is only that high -if- you keep it rolling on many people at once (which requires constant refreshing, since the first renew stops ticking when the last one is cast), and -if- it ticks for it's full value, which isn't possible in any encounter. The 700 HPS each renew does comes at a cost of a GCD and cannot trigger any of the other holy talents, so it might as well be 200 HPS and 4.0 HPM.

To me, it's still pretty easy to see why rolling renews, even from an entirely theoretical standpoint as the one provided here, are still not desirable over use of other spells that can crit. That said, renew is still part of my arsenal, and I still use it on Twilight Zone tanks. Still only 6% on a full Naxx+Maly, though.

I checked my own WWS just to be sure. My flash heal (including crits) averages out to 6185 (that's about 5000 HPS, with efficiency I can better control), and we haven't even considered how many times we can substitute binding heal (11400 HPS) in a situation where we may even consider rolling renews on self and target.

Thanks to you guys for doing the extra math. I appreciate it. Honestly, though, I've forgotten what we're arguing for or against anymore, since I can't imagine it's the efficacy of rolling 8-10 renews.

You know what, I like that idea. Have CoH be a waterfall effect instead of a flat heal across 5 (or 6) people. Heal more the closer the people are, less the further they are away. Interesting.

-- (Ignoring this, because it's not the point of the question)

-- You just used the 4000 spellpower value of Renew healing here ...

-- and you ignored passive haste

-- Yes, if you assume all ticks, it does have much higher HpS. That's a pretty big assumption, but a valid one in context. We've always been more concerned with HpM, or, at least, I have..


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

Actually I dont agree with you, there are no significant mistakes there:

1) 11859.55/12 = 988 and NOT 676. I was copying these values from a spreadsheet and simply made a copy paste error. I used the correct value for 2500 spellpower in the calculation. No real mistake here.

Also Flash heal: (2040+1.006*spellpower)*1.1*(1+0.3*crit). The discrepancy is from using 0.5 for crit. I have explained in detail why using 0.5 on a crit even when assuming 0 overheal is not a valid assumption. If you extrapolated the value of crit at 0% overheal from effective healing data then you get a value that is ~0.3 and this is the value that is comparable with other spells.

2) Its a big mistake to ignore the HPS value of rolling renews, because this value determines whether flash heal can replace the HPS value of a single renew. If renew rolling HPS > flash heal spam HPS, then a cast of renew offers more HPS to the mix than flash heal.

3) I can safely ignore passive haste, because it does not change things. In fact ignoring haste benefits flash heal. I think this is a major mistake in the way people think about renew. The fact that renew has a set duration does not mean you gain no benefit from haste. When you cast renew you sacrifice a certain amount of your casting time in order to get a set boost to HPS over a period of time. The amount of time you sacrifice is a GCD every 12 seconds. Because haste affects the length of the GCD and hence the amount of casting time that you lose casting a renew. It takes EXACTLY as much time to cast a renew as it takes to cast a flash heal, the fact that the healing of the renew takes 12 seconds to resolve makes no difference, because this HPS is ADDED to the HPS you get from casting other spells after you cast the renew. Hence the basic gain you get from haste on flash heal is exactly the same as the gain you get on renew. Instead of getting 676 HPS for 12.5% of your casting time with 10% haste you get 676 HPS for 11.3% of your casting time, thus the amount of HPS needed by flash heal in order for the flash heal to gain a bigger benefit is increased by 10% which is exactly the benefit flash gets from 10% haste. In other words passive haste gives absoludely no benefit to flash heal that renew does not get. In fact if you want to be accurate anything over 5% haste actually has LESS benefit from flash heal because it reduces cast time to under 1 second when you get an IHC proc. Basically 10% haste means you can get 10% more HPS by rolling renews, which if you remember is the HPS value that a spell needs to exceed in order to replace a cast of renew without loss to HPS. Thus ignoring passive haste actually benefits flash heal. Your understanding of the mechanics of renew are the flaw here not my reasoning.

4) I think you have not read my conclusions properly. I will link them again for you

My model is not wrong at all. Glyphed and talented renew has a purpose and a niche. If a target will take repeated damage (e.g. a tank) renew is always worth casting. Renew is also worth casting if you can use it to get ooFSR time. Healing static damage with renew is a bad idea, because if the damage is too large for flash heal, then its better to use gheal instead.

Renew is unjustly maligned because of a few rather inaccurate misconceptions about it. Let me dispel them

- Renew gains the same or better benefit from haste as flash heal does.
- Renew does not automatically overheal massively
- SoL the glyph and IHC do not make flash heal much more mana efficient than renew.

Renew has its place. Its limited buts its definately there. Keeping a renew on certain targets is NOT bad healing, its smart healing, it saves you casting time and boosts your HPS. Renew is well worth glyphing and talenting for exactly these situations IMO. Especially considering that renew can be cast up to 3 seconds ahead of time and is instant is a great boost. The argument that renew does not use more holy talents, is valid but not completely. Flash heal also does not use imp renew and twin faiths, which renew does. Not every spell uses every talent, but that does not make them useless. IHC and SoL (which I dont take) are not the only holy talents we have.

I don't get it. If you assume 0 overheal a crit heals for 1.5 times more than a normal hit? Ofcourse we can all argue about your flash heals never overhealing, because surely they will many times... and if they do, it sounds alright to me that the value would be ~1.3 times a normal flash heal.

A flash heal crit heals for 1.5 times a normal hit, it gives the target inspiration (very useful on tanks, hardly useful at all on the other raid members), it gives you a chance to proc IHC which will skyrocket your HPS.

I will read through all your posts another time, and hopefully I'll understand what you mean. Now I just don't.

I think I have understood better the regeneration nerf, since I noticed blizzard was mostly concerned about clearcasting (and druids' hots, but that's another topic).

The main problem is that, as our crit rate increases, our chance to get out of FSR increases even more.

If you are not convinced about it:
Since it is really hard to model, I can only give you a situational example. (ignoring PoM potential crits.)

Situation 1) 25% crit, situation 2) 35% crit

We get a clearcasting proc, then cast a CoH.
- Chance to get a SoL proc = 1) 55.1% 2) 68.5%
Then cast a BH to use the clearcasting effect.
- Chance to get a SoL proc = 1) 23.4% 2) 31.9%
- Chance to get a cc proc = 1) 21.2% 2) 29%

Chance of all of this happening = 1) 2.73% 2) 6.3%
And we have: A SoL proc, a free BH, another SoL proc, and another free GH or BH. That's enough to get ooFSR.

Note: The initial chance of clearcasting is not taken into account.
The numbers are low, but I am only interested into their comparison, not the actual chance of this happening.
35% crit rate does not seem unreasonable considering end of Ulduar gear. Crit is a good stat to stack indefinitely.

For a crit increase of 40% (or roughly 67% if we consider the base crit chance without buffs.) we get more than double chances to get out of FSR.
Now, the regen from ooFSR is certainly a small account of our overall regen. But this shows that it will get more powerful as our procs go up.

I am not trying to advocate blizzard's decision. I am only trying to guess what happened in their minds.
- Yes they put way too much weight upon intellect with their change.
- Yes playing the ooFSR is (to most of us) an interesting mechanism. (Else I would advise reroll shaman or paladin.)
- Yes replenishment is too powerful imho.

Now, as this has been already stated, they are also probably walking down a path to remove completely the ooFSR regen mechanism at the end. To that I can only say : time to make a world of starcraft maybe?

Edit: numbers...

I don't think your using 0.3 instead of 0.5 has been sufficiently justified.

For one thing, and perhaps most importantly, it is confusing everyone. So at the very least, it deserves more explanation.

If you are really assuming zero overhealing, then non-crits land for 100% of their base value and crits land for 150%. So one thing that is definitely confusing is that it looks like you are implying that non-crits land for 100% of their base value but somehow crits only land for 130%. That's a seeming contradiction. Why you apparently assume no overhealing for normal heals but overhealing for crits is unclear. Please explain.

The 0.3 might then come from data showing that the average effective heal (i.e. excluding overheal) is 1.3 times bigger for crit heals than for normal heals. However, that value would only be valid if you also assume a "normal" level overhealing as well. Extrapolating the value of crit at 0% overhealing from effective healing data doesn't make a lot of sense since at 0% overhealing, a crit heal is 1.5 times a normal heal by definition.

I agree if you are using the 130% factor for flash heal to account for the average healing effect you need to be clear on where you got the .30. In addition using such a factor will be different per player. While one person might overheal a lot with flash another may not. Also, if you factor any healing down of flash it will comparably make renew look better due to the fact that you have a really high estimate for the healing from renew and you aren't penalizing it any for overhealing.

This post got me interested in modeling the effect of crit on overhealing using a mixture model. One can classify heals as follows:

• One that completely overheals
• One that partially overheals
• One that does not overheal

The classification is important when talking about the effect of crits, since in the first case (i.e. a complete overheal), making the heal bigger has no effect on overhealing. So if x% of your heals completely overheal, then x% of your crit heals will also completely overheal.

However, a crit may make a heal that wouldn't overheal at all into one that partially overheals. Figuring out how many heals would change from not overhealing to overhealing depends on the number and distribution of partial overheals. That distribution may be complex, but in a 25 man raid the health bars are often changing rapidly enough that partial overheals may be nearly uniformly distributed. By that I mean that if you look at all partial overheals, it is just as common to see a 1% overheal as a 99% overheal or any other % overheal (except 0% and 100%). I checked a set of over 1600 heals from a Naxx 25 run, and indeed the distribution of partial overheals appears fairly uniform and doesn't cluster like normally distributed data would.

If you make that simplifying assumption, then the average overhealing of partial overheals is 50%. So if y% of your heals are partial overheals and x% are complete overheals, then your overhealing is:



So what happens when you increase the heal size by 50%? This is where it gets a little subtle, but I think the easiest way to think about this is to consider the reverse question: what happens when you decrease the size of your heal from 150% of normal to 100% normal size? If partial overheals are uniformly distributed, then one third of your crit heals would not have overhealed if they had not crit, right? Specifically the crit heals that overhealed from 1% to 33% would have been too small to overheal if they had not crit.

What that means is that the size of the partial overheal population (i.e. the percentage of heals that partially overheal) is directly proportional to the size of the heal itself. So if you decrease the size of the heal by 1/3, you decrease the size of y in the above equation by 1/3. Importantly, this means that if you increase the size of a heal by 1/2 by critting, you increase the proportion of heals that partially overheal by 1/2 as well.

That also gives us a way to test our theoretical model. If the model is wrong (or inappropriate for a certain data set), then we'll be able to tell by looking at the percentage of heals that partially overheal for both crit and non-crit heals. That percentage for crit heals should be about 50% bigger than for noncritical heals. I just did that, and here's what I got:

• Partial overheals, non-crits: 24.38%
• Partial overheals, crits: 35.17%

Maybe I got lucky, but that is pretty close to the predicted ratios. The most obvious practical application of all this is that you can estimate the percentage of heals that change from not overhealing to overhealing on crits, if you know your overhealing rate on crit versus non-crit heals because if you assume that the percentage of heals that completely overheal is the same for crit and noncrit heals (*), then:





In other words, the percentage of heals that change from not overhealing to overhealing from critting is twice the difference between the overheal rate of crit and noncrit heals. From that one could estimate an answer to some otherwise difficult questions such as "In what fraction of critical heals does this wasted healing occur?"

One could also "take the crits that don't overheal and the crits that overheal for less than 33% of the total, and get the ratio of overhealing crits to the number of all the heals in the set", but that would probably take a little longer than just getting the overhealing rate for crits and noncrits.

(*) Note: although it seems reasonable to assume the percentage of heals that completely overheal is the same for crit and noncrit heals, I did measure a 6% difference. I assume it's random variance, but in any event it throws the estimated change in the number of heals overhealing off by 12%.

3.0.9 coming this week. Only priest change:

* Inner Fire duration has been increased to 30 minutes and can no longer dispelled.


Nice quality of life change but will have pretty much zero effect for PvE.

Hey, at least I won't be doing Gluth without my free spellpower buff that I forgot to refresh anymore.

11859.55/12 = 988 and NOT 676. I was copying these values from a spreadsheet and simply made a copy paste error. I used the correct value for 2500 spellpower in the calculation. No real mistake here.

Dude its really simple just read the post: I have two calculations on my spreadsheet for a SINGLE renew one using the 4k spell power value which is 988 HPS and one using the 2500 value which is 676 HPS. In my post I uses the 2500 value which is 676 HPS for a single renew. By accident however I posted the 4000 value for total healing instead of the 2500 value. Its nothing but a typo. The calculations use the correct values and the final numbers I post are correct. Adding haste is really a non issue, it does not change things much.

The takehome message is that flash heal is not automatically better than renew, and that renew glyphed and talented can result in very high HPS, (PoH is better) which can be very important, especially given that you can roll 2 renews ahead of time.
===============================

Lambi the reason why I use 0.3 and not 0.5 is that I am interested in how much crit boosts effective healing not total healing. The reason why I want that is because I want a value that will actually scale better into effective healing calculations. Even though crit boosts raw healing by 1.5 it boosts effective healing by a value that is between 1 (no benefit) and 1.5 depending on the amount of overheal you have. The reason is that the higher your overall overheal the less the benefit from crit. Effective healing from crit loses its value very quickly with increased overheal.

For clarification in reference to discipline priests, I am still presuming spirit> mp5.
If we compare both of the above bracers we lose the 11.2 spellpower but still gain roughly 43 spirit if we assume BoK? Thus giving us more Mp5 overall?

If you penalize the crit of Flash Heal to focus on effective healing. Where are you doing the same for Renew, it will likely have ticks that are not fully effective or not effective at all?

I used reduced crit values myself (I use .36, not .3, but whatever) but if you are comparing a renew that does not overheal to a flash heal that does not overheal to figure out their theoretical throughput you run into problem here. But assuming the flash heal does not overheal, you factor out serendipity when calculating it's HpM. If a flash heal overheals for more than 0 and less than 25%, it becomes more, not less efficient. If you use a .3 value to account for the fact that crits tend to create more overheal, then you can't ignore the fact that crits, then, would also be saving you mana. How much mana exactly would be hard to model (best to just gather the data from WWS logs empirically) but it would not be zero. We could ignore this complexity to get rough numbers, but we can't ignore the complexity of crit proccing Serendipity while not ignoring the complexity of crit causing overheal, it skews the numbers.