Thursday, June 13, 2013

Spirit and the Prevalence of Magical Thinking

Okay, back to writing about healing (heck, back to writing anything about anything).

The biggest misconception I see about healing regards spirit. Spirit, in all honesty, is not a very complicated thing in World of Warcraft (it's arguably complicated in real life). It quantifiably increases your mana regeneration, which allows you to cast more spells. We can tabulate how much extra mana we'll get from spirit (thus, quantifiable) and we can quantify how much healing power that mana provides us.

I emphasized that second part because it seems to get lost in many/most/all healer discussions about spirit. Spirit has become a very mystical* concept in the healer world, for reasons that aren't entirely clear in this increasingly mathed-out game.

*There's some irony in decrying something as having "mystical" standing in a game about magic and dragons. But I digress.

Ask about how much spirit one should have as a healer on a major healing forum and the answer you'll invariably get is, "Enough to feel comfortable then start stacking 'throughput' stats." This is wrong on a lot of levels, but the root is that people just aren't thinking about spirit the right way.

Now, I realize that asking "How much should I have?" in a public forum is always going to be a meaningless question, for any stat, if you're looking for a number. However, we can give non-numeric answers that help the asker (or ourselves) come to good results. If someone asked "How much critical strike [or haste or mastery] should I have," I can guarantee you that the most popular answer wouldn't be "Enough to feel comfortable."

"Comfort" is such a vague, nebulous word that we're already starting off on the wrong foot. Healers seem to mean, by "comfort" with spirit, not noticing themselves running out of mana (OOM) very commonly in whatever content they're competing in. That's a really poor metric, because mana usage is fluid...not only do different encounters have different demands, but different playstyles can use mana differently. There isn't a static amount (or even a modest range) of spirit that means you're incapable of running out of mana...this isn't defense capping (for those of you who go back at least to Wrath) or full combat table coverage (for those of you who go back at least to Cataclysm). Even on a per-person basis, your spirit needs change and they should always be changing based on the situation.

Healers have, for some reason, come to regard spirit as a force outside of the optimization game, something not well understood and best left up to "feel" or one's gut. Does your current spirit total feel good? Then you have enough! Doesn't feel good? Get more! Can't say whether it feels good or bad because there's really no frame of reference and the question is a little silly? Um, well... [end transmission]


So What Is Spirit?


Spirit is a throughput stat like critical strike, mastery and haste with one difference: rather than empowering your spells, it empowers you to cast spells. This difference seems to throw a lot of people for a loop, but it's really not a huge difference.

After all, we can quantify this. Spirit, at level 90, gives you this much mana regen (per five seconds): 6,000 + (0.56435)Spirit. We can multiply that by 12 to get how much mana regen per minutes. Let's say you have 12,000 spirit. That formula yields 12,772 mana per 5 seconds (MP5), which is 153,266 mana per minute...approximately half your mana bar per minute.

What can you cast with 153,266 mana? Or, to be more fair, what can you cast with 81,266 mana, since 6000 MP5 comes for "free" per the "6000" constant in the formula, which equates to 72,000 mana per minute if you had no spirit at all. Over a 6 minute encounter, your 12,000 spirit is netting you 487,598 mana.

Is that worth more than investing those 12,000 points of stat budget into things like critical strike, haste or mastery? That's a much more complicated question, of course, and it depends greatly on class, but it's one that can be investigated.

You can do the same exercise on a smaller scale, like whether cutting down from 12,000 spirit to 11,000 spirit is a good idea. Or whether a spirit trinket is better than a haste trinket. You can calculate the numbers and then evaluate what you'd spend that mana on. That gives you a much firmer perspective on how much spirit you need than "how it feels."


What Can We Spend Mana On?


Okay, yes, heals, of course (unless you're a discipline priest, then DPS spells). But which healing spells? Ignore your spells that come with cooldowns attached. If those spells are good ones, you'll be casting them pretty much on cooldown and more mana won't allow you to cast them more because the cooldown prevents you. Mana doesn't give you more Penances or Swiftmends. What it will give you is more of your spells that have no cooldown, your "spammable" spells. These are mana-limited...you cast as many of them as your mana allows, around your cooldown-limited spells.

So we're talking about spells like Rejuvenation, Prayer of Healing, Chain Heal, etc. These are the spells that you should check the mana cost on, compare it to the mana you'll get from spirit and from that calculate the approximate throughput you're adding.

If you have a World of Logs record of your performance, the comparisons become reasonably straight forward. For example, you can figure out how much critical strike percentage you'd get from the same amount of stat budget and multiply your non-overhealing output by that added percentage and compare that with the added healing you'd get from the X extra Rejuvenations you could have cast, as an example.

Of course, that does introduce a tricky element: over-healing. If you critical strike more (or cast faster due to haste, or just hit harder with your heals through mastery) some of that extra healing will likely be pushed into overhealing. This might actually be a benefit to more spirit...since you control where and when the extra casts go, presumably your decision-making will bias those spells against being overhealing (though with other healers in the mix, this is far from certain).

The point is not that you can come up with clear-cut certainty as to which will give you more bang for the buck, spirit or another secondary stat. The point is that you do not need to relegate this comparison to the realm of feelings and intuition-elves. You can sit down and take a more systematic look at it.


tl;dr


You neither want to blindly stack spirit, nor blindly cut spirit. You don't want to think to yourself, "I don't run OOM, so clearly I must cut spirit." You want to consider what spirit nets you (not mana, the healing output of the extra spells you're able to cast) versus the healing output that the other secondary stats net you. Logical thinking versus magical thinking. Spirit isn't weird and magical...it's concrete and it gives you more healing spells to cast. Nothing hard to understand or calculate there.

If I were to boil this down to a principle (beyond "investigate what everything nets you"), it would be this: If the spammable spells your class affords you are powerful, you're probably better off with more spirit. If the spammable spells your class affords you are somewhat lackluster, then you're probably better off with less spirit (other secondaries will allow your cooldown-limited spells, which you rely disproportionately on, to hit significantly harder).

That said, don't let anyone tell you that spirit is touchy-feely. You don't stack it til you feel comfortable. You stack as much of it as you can effectively use. If you have strong spammable spells and don't come anywhere close to running out of mana, that may just be a sign that you should cast more (and more expensive spells).

Hey, wow, I just wrote a blog post.

Sunday, February 10, 2013

Best Practices for the Shadowy Arts

I probably still heal more than I DPS, though I DPS as a greater percentage of my raid encounters than at any time in my raiding history since I had a retribution paladin as my main. However, I have no fascinating insights about healing right now. Maybe I will if I sit and think through what I do a bit more carefully--one consequence of working hard on my DPS is that I think a lot less about healing than I once did. I think I still do a good job, but I depend a lot more on gleaning best practices from other people than I once did.

However, shadow...I have all sorts of things to say about that right now, as I suppose my recent posting history suggests. ("All sorts of things" defined as "something every three or four months or so.")

 A couple of the people I read have been doing some "best practices" posts, about restoration shamans and healing priests, respectively. If you really want to read about healing, check them out. Wait, wait...on second thought, no, don't click away from this post. They get enough traffic.


Vampiric Touch Is A Last Resort


Vampiric Touch is not a vitally important spell. In fact, there are only two situation in which you should use it. Admittedly, those two situations will ensure you use it a lot, but it's useful to understand why you're using it so that you don't overuse it.

  • Situation 1: You are fighting one or two opponents and you have nothing else to cast except Mind Flay. Getting Vampiric Touch's damage rolling is worth a Mind Flay GCD (which will, in fact, lead us to another shadow pillar in just a moment). But you should literally place every other rotational spell (Devouring Plague, Mind Blast, FDCL Mind Spike, Shadow Word: Pain refresh, Shadow Word: Death, Shadowfiend/Mindbender, the level 90 talent you've chosen) ahead of casting or refreshing Vampiric Touch.
  • Situation 2: You are fighting multiple opponents that will live for a decent amount of time and more opponents will join the fight. In this case, you do not want to put Vampiric Touch on every monster. You want to put Shadow Word: Pain on every monster and Vampiric Touch on just a few of them, before resorting to Mind Sear (and things like Mind Blast, Devouring Plague, Shadow Word: Death, etc, when possible). 

The reality is that in situations like the second one, every Vampiric Touch cast is a DPS loss because you can no longer replace Mind Flay ticks with it...between all the Shadow Word: Pain casts and refreshes, important single target casts and Mind Sear, you always have a higher DPS offering. The sole reason you want to cast Vampiric Touch in this case is for the mana regeneration. Lots of Shadow Word: Pain casts is mana intensive and you don't want to run out of mana (being unable to cast is an even bigger DPS loss). A good rule of thumb is that one Vampiric Touch for every two Shadow Word: Pain casts will keep you approximately mana neutral. Of course, if you're only getting one wave of adds and they won't be around long enough to keep refreshing your Shadow Word: Pain, you don't need to be mana neutral...you just need to not run out of mana before the adds die, at which point you'll presumably return to your single-target rotation, which is mana positive.

So, in general, you'll just want to cast Vampiric Touch when one of two conditions is met: you're only replacing a Mind Flay tick with it or you need the mana regeneration from Vampiric Touch.


All GCDs Are Not Made Equal


Whenever you make a choice as a DPS, the main resource that you're spending is not mana, it's GCDs. It's still ultimately an opportunity cost decision...a healer generally has to weigh casting a spell against the possible spells they could have used the mana on, while a DPS has to weigh casting a spell against the possible spells they could have used the GCD on.

However, there is a big difference: mana is a fungible resource, a GCD is not. Every bit of mana is essentially identical in value for the purpose of casting spells, whereas a GCD's value varies wildly due to cooldowns. You can't use every GCD on Mind Blast, because the spell is cooldown constrained. So which spell you could have used the GCD on determines the GCD's value. A GCD in which you could have cast a Mind Blast is extremely valuable whereas a GCD in which you could have cast, say, a Vampiric Touch is significantly less so.

(It bears noting, of course, that healers also have cooldowns to contend with and therefore GCD analysis can come into play for them as well, especially when not mana-constrained, as was the case at the end of Wrath of the Lich King.)

All of this means that you can significantly minimize the consequences of actions by replacing Mind Flay GCDs. I like to say that casting Mind Flay is better than casting nothing, but not much better. Which means you have a lot of almost free GCDs to fit things in when necessary. One example is the above principle; it makes using Vampiric Touch a DPS gain because all you're losing to get it up and refresh it is an occasional Mind Flay tick.

Another example is (now) purely theoretical: The original patch 5.2 redesign of Shadow Word: Insanity was fairly simple: whenever you had three of your own DoTs up on a target (so all of Shadow Word: Pain, Vampiric Touch and Devouring Plague), your Mind Flay would be buffed. This means that the Mind Flay buff is constrained by Devouring Plague up-time (since the other two DoTs you can keep up virtually at all times). The natural behavior that this encourages is to cast Devouring Plague every time you have a single shadow orb. Devouring Plague scales linearly with the number of shadow orbs consumed, so you lose no Devouring Plague damage by casting it with less than three orbs. The cost of doing that is you consume more GCDs for the same Devouring Plague damage, so it's a DPS loss to cast it at less than three shadow orbs.

However, when you consider that A. each Devouring Plague gives an equally long and strong buff to Mind Flay regardless of orbs consumed and B. the lost GCDs would essentially be taken from (unbuffed) Mind Flay, it became entirely obvious that it would be a clear DPS benefit to cast Devouring Plague at one shadow orb.

Blizzard realized this and didn't want shadow priests not ever saving up shadow orbs, so they quickly changed the design of Shadow Word: Insanity such that the buff you get to Mind Flay is based on how many shadow orbs were consumed by Devouring Plague, which returns it to the original state: you don't gain any extra DPS by using less than three shadow orbs, so you only lose DPS from the extra GCDs spent on Devouring Plague.

You can see, though, this principle in action with these changes. Losing unbuffed Mind Flay GCDs is simply not a particularly large loss, so nearly any DPS gain will be worth losing them.

You can put this into effect in various ways. A big one is movement. Movement is a DPS gain in that if you die, your DPS plummets. When possible, if you can time your movement for times when you'd otherwise Mind Flay, you'll lose a lot less DPS than if you move when you could be casting, say, Mind Blast. Another example is if you want to cast a Power Word: Shield during a high damage phase, it's ideal to cast it when you're only stealing the GCD from Mind Flay.

It should also inform your gearing. Haste is very useful for reaching a haste plateau for Devouring Plague but, when you can't, it's pretty nearly worthless to gear for. Added haste that won't get you an extra tick of Devouring Plague will only shorten your spell cast times...which means more Mind Flay time. It's far better to get critical strike, which will improve the DPS of all your spells, or even mastery, which will improve the DPS of all your DoTs.  Both will benefit you much more than being able to channel Mind Flay more often will benefit you.


Master The Art Of Multi-Dotting


This tier (and quite likely the other tiers of this expansion) feature a lot of fights with adds. To get the most out of such fights, you need efficient methods to keep DoTs rolling on as many targets as possible (again...the major one to distribute is Shadow Word: Pain; only put Vampiric Touch on multiple targets when you need the mana regeneration or your only other possible action would be Mind Flay).

The techniques I've found to be most valuable in managing multi-dotting are as follows:

  • Set a focus, track your focus, make it obvious which of your DoTs your focus has (and time remaining on them) and make it easy to apply DoTs to your focus. The best way I've found to easily maintain DoTs on my focus is using shift-modifiers on my usual Shadow Word: Pain and Vampiric Touch keybinds. So, if 4 is Shadow Word: Pain and 5 is Vampiric Touch, then shift-4 puts Shadow Word: Pain on my focus and shift-5 puts Vampiric Touch on my focus.
  • Make sure frames for boss1, boss2, boss3 and boss4 are enabled and easily visible in your UI. These are unit frames that Blizzard makes available by default for major adds that pop up during a raid encounter or for each member of a council boss. For example, the active quilen in the Stone Guard encounter of Mogu'shan Vaults are represented by boss1, boss2 and boss3 frames. The two tendons (right and left) in the Spine of Deathwing encounter of Dragon Soul were represented by boss1 and boss2 frames. These allow you to quickly put DoTs on different active creatures, and track those DoTs, whether you want to use an addon like Clique to click DoTs onto those frames or use mouse-over macros to apply DoTs via those frames.

Remember that positioning doesn't matter for applying DoTs...you can cast DoTs onto even targets directly behind you. So the challenge is just in swift targeting and tracking among multiple monsters.


Survival Priests


Shadow priests have a lot of survivability utility, both for themselves and for others...make use of it. Some fights, especially on heroic mode, can be extremely healing intensive and if the raid can just survive a little bit longer, the raid can get the boss kill. Help your healers out when it makes sense, especially when healers have to be cut in order to make an enrage timer.

  • Glyph of Dark Binding: This is my least favorite option. When it affected Binding Heal (during the beta), it had a ton of potential. Now that it only affects Renew, Prayer of Mending and Leap of Faith, I think there are significantly better glyph options. I don't consider Renew or Prayer of Mending to be particularly great options for shadow priests, but if you can find ways to weave them in such that they're useful and don't hurt your DPS too much, it's definitely worth considering. One point worth making here: getting a Renew or Prayer of Mending on someone below 25% health can trigger Twist of Fate for your damage dealing.
  • Glyph of Fade: I like this glyph quite a bit, especially on a fight like heroic Gara'jal the Spiritbinder. By fading off 10% of the damage I take when I'm a Voodoo Doll, I also reduce the damage all the other Voodoo Dolls in the raid take. There are various fights where dropping your damage taken is well worth the GCD (again, remember...if you're replacing Mind Flay ticks, it's not a major DPS loss).
  • Glyph of Inner Sanctum: This is particularly nice simply because it's passive. Since you should always be using Inner Fire as a shadow priest, this is a flat reduction to magic damage at all times. Any fight in which you can expect the raid to take a lot of magic damage, this can be a very valuable glyph to slot in and it doesn't require any mental overhead once the encounter starts.
  • Power Word: Shield: The old priest stand-by. As an instant 60-70k health buffer, this can be a great use of a GCD either for yourself or for a raidmate. If you see someone on the verge of death, especially a healer or tank, giving them an immediate buffer gives the healers a chance to get them back up to safety before they actually keel over. This also has the benefit of not kicking you out of Shadowform when you use it.
  • Vampiric Embrace: This is the big one. The one you should be making your healers and/or raid leader aware of, because it can very easily be a raid-saver. Any DPS you do during the duration also provides HPS (healing per second). If you glyph it (and I strongly suggest that you do), it's like having an extra healer for the duration, while not losing any DPS at all (it's off the GCD, so you don't even lose a GCD). The reason that I suggest glyphing it is that, generally, raid health is only in serious danger for short, intensive bursts. Spiking your raid's healing as much as possible for one of those times is a bigger benefit than dribbling out a smaller amount of healing over a longer period of time.

Figuring out how and when to use abilities, talents and glyphs to help keep you and your raid alive, without hurting your DPS much, is a very important part of mastering playing a DPS role in raids.

Friday, January 11, 2013

Do Priests Need Mind Spike In Their Spellbooks?

I suppose this blog was in danger of becoming abandonware, considering my last post was before Mists of Pandaria launched. It wasn't that I didn't want to write, nor was it that I wasn't enjoying World of Warcraft: Mists of Panadria, nor was it even that I didn't have time.

What I didn't have was anything interesting to say. I suppose readers will judge whether I've ever had anything interesting to say, but I don't do recapitulation of news, nor do I tend to do standard healing/DPS guides, nor do I tend to document my own adventures. I try to write when I think I have a point that I haven't seen represented elsewhere and isn't (from what I can tell) common knowledge/common sentiment. In other words, dear reader, I try to shut up when I don't believe I have something worthwhile to say, much as I'd like to push my blog back up people's readers with a new post.

That's just an aside explanation for why so long between posts...this isn't a stealth "let me get a new post up by explaining how I don't just write filler posts." That would take chutzpah, and I don't have the chutzpah necessary.

No, this post is about Mind Spike. Or, at least, primarily about Mind Spike. Rather, we could say this is about Mind Spike's role in the shadow arsenal leading into a broader (or maybe narrower) point about shadow as a whole. Though, ultimately, it's about something else entirely. It's complicated, but you'll see if you read on!

One more piece of business to attend to before I get to the main event: The last post was about shadow and this post is about shadow, but I am not now a pure shadow priest and this is not now a pure shadow priest blog. I still heal a lot...in fact, I think I still heal more than I DPS. I DPS any fight in which we only need two healers, and we're three-healing most progression fights right now, especially as we push further into heroic raid encounters. I DPS a lot of farm encounters, though! (And even a few progression fights. I consider it important and even necessary to be on top of my shadow game.)


Why Mind Spike?


Mind Spike joined the priest toolbox with the advent of Cataclysm. When it was added to the class, the rationale for it made perfect sense: it was a situational nuke for a specific situation--an add that needed to be burned down quickly, too quickly for damage-over-time (DoT) spells to be useful.

Shadow priests have traditionally been a long ramp-up-time DPS spec; that is, it takes them a while to go from zero to 60k DPS, because they need to get DoTs up and get shadowfiend out and the whole thing is a lot of up-front GCDs when you're fighting something that needs to die quickly. It's the difference between sustained DPS and burst DPS...a spec that can do a good amount of DPS over an 8 minute fight may not be able to effectively contribute to a 30 second burn phase and vice versa.

Both are important. It goes without saying that one's total DPS over a fight is generally the most crucial aspect, but hitting "burst windows" can often be just as, or more, crucial in certain fights. The Spine of Deathwing fight in the Dragon Soul raid had a tendon that became available to destroy for short periods of time, and getting it down within the necessary time frames was required to defeat the fight, especially on heroic mode. In this tier, the Elegon encounter in Mogu'shan Vaults has waves of adds that must be burst down before they reach a certain place in order to stack an increased-damage-taken debuff on Elegon that's necessary to meet the enrage timer.

Going into Cataclysm, shadow priests really didn't have much burst potential. Mind Blast had a cooldown and DoTs and Mind Flay were too slow. Mind Spike was supposed to fill that niche: a damaging spell that would allow priests to start blasting something immediately but, thanks to its DoT-extinguishing mechanic, couldn't be used rotationally. This suggested that Mind Spike could be a really powerful but really situational spell, useful for adds that you didn't want to DoT because it wasn't supposed to survive that long.

Now, I'll admit it: I have no idea if it was successful in that role in Cataclysm, because I barely played shadow in Cataclysm. I was a committed discipline priest going into Cataclysm and I spent the first two tiers switching around my off-spec several times a week, trying to find a spec I was comfortable with alongside discipline. I eventually settled on holy in T13. I followed the news generally about shadow (as I do about a lot of classes and specs, because I'm a bit of an obsessive geek), but I didn't have any serious experience.

That has changed with Mists of Pandaria. I was tired of being one-dimensional--just a healer. Not only did I feel limited in what I was capable of, I felt like it constrained our raid team, not to have a proper swing-DPS. So I picked up a shadow spec and committed to learning it fully, tailoring a UI for it and both figuring out and researching tricks to optimize my play as a shadow priest.

All of which has led me to one question: Why Mind Spike? The above rationale for it makes sense--it just doesn't fill that role as we stand in Mists of Pandaria.

As a hard cast, Mind Spike's damage is virtually never worth casting. The damage is uninspiring and it removes the DoTs that you have rolling on your enemy. So it isn't worth using in your rotation and the low damage makes it a poor tool to spam on adds that won't be around long enough to be worth putting DoTs on. It's better than nothing, but not much better than nothing.

It seems, then, that when you consider that you can't use it rotationally nor can you use it a short-term nuke, Mind Spike really doesn't occupy a useful niche...except for one.


From Darkness, Comes Light (And Beyond?)


The one time that Mind Spike is worthwhile to cast is when a shadowpriest has the From Darkness, Comes Light talent and Surge of Darkness procs. This makes Mind Spike instant (the value of which is that it allows you to use it when moving, as the cast time is the same as the GCD consumed by an instant), significantly increases its damage and causes it to not extinguish the DoTs you have rolling.

This is a fairly potent ability (made more so by the Glyph of Mind Spike) but considering that this is the only time that it's worth using, it seems a little unnecessary to make it a spell in one's spellbook...why not simply make it a button you can use when you take the talent, like many other talented abilities? Not that it's a particularly big deal that it resides in the spellbook, but it seems like a rather wasted spell in the arsenal.

Unless...we extended the concept of From Darkness, Comes Light a bit and also fixed another problem with the shadowpriest toolkit: lack of a burst DPS cooldown. As I noted above, it's important to have a way to occasionally shift into another gear when necessary and shadowpriests don't have that. If we consider a single-target fight, we have one speed: normal. Add that needs to be burst down right now? Normal single-target rotation. Execute phase with the enrage timer seconds away? Normal single-target rotation. It not only takes away versatility and decision-making, it's also unsatisfying.

However, what if we repurposed Mind Spike to fill the niche that it was originally supposed to: burst DPS? But we want to make it a cooldown, something we can only use once in a while, but is quite powerful those times. Well, shadowpriests used to have a bit of a cooldown in Cataclysm (of varying power)--Dark Archangel. But Dark Archangel is gone, leaving only the nifty wings graphic in the form of a minor glyph that appears when you use Devouring Plague.

It would be nice to have Dark Archangel back in some form, it would be nice to have some sort of DPS burst cooldown and it would be nice to use Mind Spike a bit more, considering it's a real spell in our real spellbook.

With that in mind, I present an entirely new spell (in concept, numbers could be tweaked because hell if I know how to balance numbers across all classes):



This is obviously built upon the existing functionality of From Darkness, Comes Light but would not require that talent. It steals the mechanic (because it's a very good mechanic, in my opinion) but is independent. You could have both, but even if you took another talent from the tier, you'd still have this Dark Archangel cooldown.

This would provide a significant opportunity for burst DPS. You would still weave in Mind Blast (and Shadow Word: Death if the target were at 20% health or less) because you'd still want the resource generation, and you'd still use Devouring Plague to dump your shadow orbs, so that future orb generation wouldn't be wasted but, otherwise, you'd use Mind Spike as filler during the duration of Dark Archangel instead of Mind Flay, which would be a major increase in DPS for the 15 seconds (or whatever professional class designers considered balanced).

Ultimately, I think that shadowpriests need a burst DPS cooldown in order for the spec to have the needed variability in play. That's the foremost concern here, even if the post title is misleadingly about Mind Spike. However, with the Mind Spike spell going to waste outside of one talent, and a great graphic that used to be tied to a great concept being wasted on a minor glyph, it seems like the opportunity exists to synthesize these issues into a single solution.