Thursday, May 14, 2009

Revitalize: WWS report

I've been keeping an eye on my raids' wws reports, and wanted to share some of the charts I put together using last Tuesday's raid, May 12th, 2009, lasting approximately 2.5 hours (Flame Leviathan was cut out, where no one was using spells/abilities).

First, here is a sample of the number of Revitalize procs vs. the number of times Rejuvenation or Wild Growth was "applied" as a buff by that player. I am assuming that the buff applications from WWS are the number of times that spell was given to that player edit: the applications are the number of times the spell was given FRESH to that player; the applications are not counting the number of times the HoT was renewed before a previous version wore off.

  • Wild Growth fresh-applications were pretty even across the board (among melee), with a 6-target splash across 10 people (+ pets).
  • Rejuv applications were focused on the two tanks, of course, and both tanks saw a greatly increased number of revitalize procs from this. NOTE: tanks' rejuvs were often refreshed before they ran out to keep them swiftmendable, and these refreshes are not counted in the applications numbers, but they did contribute significantly to their overall procs of revitalize.
Revitalize is definitely proccing fairly regularly. The proc chance itself is given as 15% per tick, but I cannot judge the exact number of ticks based on the HoT's application alone: though my Rejuv has 6 ticks and Wild Growth 7, each spell may be over-written by a fresh Rejuv or Wild Growth before each tick has occurred. For this reason, I don't want to try to reverse-engineer an exact procs/tick rate right now, like I did pre-3.1 :)

The data stands fairly well on its own since most of the spellcasts were (usully) made specifically to heal the targets, so there is no need to argue for its mana cost. Those times I did cast it specifically for the revitalize portion for the DK, warrior, and/or cat, I did so when I had plenty of mana to spare. I did check to make certain that the feral druid never cast Rejuvenation over the course of the raid, so his untalented spellcasts are not skewing the procs (no offense, Cel!).
  • I want to reiterate that Revitalize WILL PROC even when the healing from the HoT's tick is not applied (due to the target being full health). The game will not apply the healing to a full-health target, but the game will still give that hidden tick a chance to proc the power return.

Now I will be comparing the FRESH applications of the spells on the individuals, and their resulting revitalize procs. Note: 24% of my healing done was Rejuv, and 23% was WG.
  • When Rejuv > WG applications, procs per combined FRESH spell application (rejuv + WG applications) were roughly 60-65%.
  • When WG > Rejuv, procs/freshapplication = ~45-50%.
This makes me scratch my head, since Rejuv has 6 ticks (I have Nature's Splendor), and WG has 7 ticks (thus should have a greater number of procs per spbuff application). EDIT: those who had Rejuv > WG were tanks, and their rejuvs were very often renewed as well, so each "application" may have had 6 ticks or 20 or more. For dps/ranged, WG's 7th tick was sometimes over-written as its cooldown is up before the last tick occurs, but not to the same level as rejuv was, since rejuv is usually a permanent fixture on my tanks, while WG is not permanently upon the raiders. This constant renewal resulted in the higher rate compared to those who received more fresh WG's than fresh rejuvs.

Combining all 6 classes resulted in a 54% procs per fresh application, with WG > Rejuv. Combining every member of the raid (13 total due to some switch-outs for Emalon) resulted in a 50% procs/fresh application rate, again with WG > Rejuv. Remember that this estimate is also including every procs from the unaccounted time that rejuv or WG was RENEWED on the target, thus extending the chance for ticks. Unfortunately, given that procs are not specified as coming from rejuv or WG, and that any given cast of my WG may have hit 1 target or 6, that number of refreshes can't easily be found, even when parsing manually through the combat log.

This chance will fluctuate per person, per raid size (chance of over-writing a Wild Growth is less when there are more targets to choose from), how frequently the caster refreshs Rejuv vs letting it tick fully, and probably a slew of other variables hiding. This is just the estimate derived from my playing style in a 10-man raid on a single night.

  • Important: Any given spell application could proc once, twice, even three times or more before the buff wears off, or it may not proc at all! Also, the procs are determined for each individual's own buff ticks, so while one person may proc on a wild growth tick, the others who have the same wild growth upon them may or may not get a proc themselves at the same time. It is for this same reason that ticks can't be counted up in a combat log to compare against procs, because the HoT may proc revitalize without actually ticking a heal.

Now, let's ignore the craziness of procs per fresh applications vs refreshes, and just look at the breakdowns of individual players' power gains. I have selected a Prot warrior, Deathknight (blood), Feral Druid (cat form), Prot Pally, Warlock, and Resto Druid (me!).

The returns from revitalize are all shaded in the pale violet-blue, and I included the raw power numbers returned by each element. The prot pally had quite a few other smaller returns from misc things that I cut out to reduce clutter, all returning less than 6k mana each over the raid night; I also only counted the feral druid's energy returns, ignoring the rage and mana he gained while briefly in bear or caster. Otherwise, each of the other classes shown show their full power returns per type for the night.
  • Prot Warr: massive rage returns with rejuv kept on, taking up roughly 1/3 of their rage power gains, equalling or surpassing their rage returns from bloodrage and shield specialization.
  • Blood Deathknights: between WG and some occasional thoughtful Rejuvs, it beat out his own runic power regenerator, butchery.
  • Kitty Druid: it's no Tiger's Fury, but as any feral kitty will tell you, more energy is a wonderful thing. Surprise ticks of energy make for happy kitties (and rogues), and bloody mobs.
  • Prot Pally: matches/exceeds mana tide totem's returns, and that shaman had to intentionally put that totem down; also matches illumination. A relatively small portion of the returns, but noticeable.
  • Warlock: same relative percentage (and amount) of mana as the prot pally was returned to the lock by revitalize, though the lock received fewer ticks. This is because warlocks have larger mana pools than prot pallies, so each proc restored more mana (as 1% of mana pool). A small slice, but that's less life-tapping and thus less healing you need to give them, I suppose!
  • Resto Druid (myself!): Lifebloom's 3.1 change of refunding half its mana cost on bloom has really borked the energy meters, and innervate is not included since it only speeds up passive regen (atm). Again, though, it matches mana tide's contribution (a 5-minute cooldown), and that's pretty nice when I catch a WG splash or keep rejuv on myself for environment damage/aoe.

10-MAN vs 25-MAN RAIDS
Your use of WG and Rejuv will affect how useful this talent is for you. The more players that are in your raid, the smaller the proportion of players you will see this proccing on (from your own spells), due simply to the limited proportion of the targets you can cast on at any given moment. It will not, however, reduce the raw amount of power returned to your raid: it only spreads it out, just as your healing done is spread out among your targets. Having more resto druids with the revitalize talent in a 25-man raid will increase the proportion, of course.

Having more players in the raid as targets for your Wild Growth will impact how often your WGs overwrite themselves on your targets as well, which affects the "application" rate.

If you are raiding with other druids who do not have revitalize, it may throw off your own measurements of fresh HoT applications vs revitalize procs from your WWS reports. Know at least that since HoTs are not overwritten by other casters, your own spells will not be affected negatively by raiding with druids who do not have Revitalize, unless the other druids also use an unglyphed Swiftmend and thus remove your Rejuvs (fail!).

The thing to ultimately keep in mind is that if you're casting these spells anyway, this is purely passive power regen for your targets. You don't have to specifically cast it as a buff at someone, nor do you have to drop down a totem for it. You just have to heal, which you'd be doing with or without this talent.

If you're using rejuv and WG, there's no reason not to get this talent. Every bit of mana regen adds up, and for those classes with limited power regeneration options, talents like this are a blessing. Warriors/bears, rogues/feral kitties, and DKs don't have as many options for power return as mana-users, and for some of them (such as DKs and warriors) the returns can be a significant part of their overall power returns. This in turn will increase damage output and tank threat, giving you shorter combats to heal through and quicker boss/trash kills. It will also give DKs/warrs/bears that extra little bit of power to help start off threat, or to get in an interrupt early in the combat.

Modelling revitalize is a headache. You can mathematically estimate chances, but who's to tell what will happen in the real raid, when you're throwing spells about and refreshing them before they tick fully? The best way you can really determine how good revitalize is, is to look at what it did. Look at how it impacted your raid, who it boosted, how it may've even given you as much mana back as the priests' hymn of hope gave to you over the course of the raid.

I whole-heartedly support this talent.


Animagis said...

Great writeup Kae, and glad to see I'm not the only one who sees the power of the talent now that Wild Growth can proc it.

As my guild's most consistent raid healer, it would be lax of me to not get the full potential out of what this can do.

Kae said...

My husband is the DK in this study. He still requests a rejuv before going into combat on some bosses, to help tick up his runic power for the early stages of combat... like to get an interrupt off on General Vezax, or to give him a head start on threat when he's tanking. It doesn't make or break a fight, but it quite useful, and I have watched it proc 2-3 times in a row on him (and the prot warrior) as we're about to run in to start a boss fight.

Lonster said...

When you view your WWS reports, "Buff Applied" is only if they didn't have it prior. If you overwrite an existing buff, it will not appear as "applied".

That's how you can tell good rogues, they will have one application of "Slice and Dice", and it'll last 3 minutes.

What I'm supposing happened, is you clipped your Rejuv casts (since they last so much longer than WG), which will result in a lower number of listed applications.

A much simpler method of determining chance per cast of a Revitalize proc is to multiply out the chances of not getting one.

The listed chance for the proc is 15%, so there's an 85% chance of not getting it.

1-(.85^n), where n is the number of ticks, is the probability of having Revitalize proc.

Ticks : Chance of 1 proc
1 : 15.00%
2 : 27.75%
3 : 38.59%
4 : 47.80%
5 : 55.63%
6 : 62.29%
7 : 67.94%
13 : 87.91%

So, over the 6 ticks of Rejuv, there's about a 62% chance of getting a single Revitalize proc, and over the 7 ticks of WG, there's a nearly 68% chance of getting a single Revitalize proc.

Suppose both HoTs are running on a char, there'd be 13 total ticks, and there'd still be a 12% chance (that's about 1 of 8 times) of NOT getting a Revitalize proc.

Kae said...

Part of the reason I went this raid-data-derived route instead of the mathematical back-tracking is precisely because of the high chance that the spells will be over-written before the final ticks are given. Will each cast of a standard rejuv spell always have 6 ticks? Not if it is overwritten. Fewer ticks given per overwritten spellcast is something that is unique to each player, each target (tank vs dps), and raid environment, thus not something easily modeled :(

I am disappointed to hear that WWS only counts applications when they are fresh. That makes things difficult for trying to figure out Wild Growth, since the cooldown is often ready before the last tick of the previous wild growth occurs; tanks are usually refreshed on their rejuvs before it wears off, too, to keep them swiftmendable in a crutch. TY for pointing out this fault in the WWS applications, I'll have to reword parts of the post since I had stated the assumption that applications would be per cast in an effort to better model procs per person affected by WG (which isn't always the maximum number, either!).