(Alternatively titled: Why we throughput classes maybe, kinda shouldn’t hate Disc Priests and Paladins all that much)
It is a common occurrence on raid nights the world over—the healing team is assembled for the evening, and with every absorb class added to Group 5 (or whichever group holds your healing bundles of joy), the groans grow ever louder from your Shaman, Druid and/or Monk healers. Slowly, the raid fills and those old battle lines are redrawn in the sand, with your Discs standing resolutely next to the Holy Paladins, while the throughput classes glare at them from across a sometimes icy divide, muttering about the overhealing to come.
Those darn discs and paladins, who can take a beautiful healing meter full of competition and turn it into a one-sided argument. For this exact reason, I’m as guilty as anyone of throwing a good amount of vitriol at my absorb-centric friends. Despite the fact that I join their ranks for Promethean’s weekly alt raids, I confess that I do find it aggravating to watch the divide between classes grow ever broader as our Discs and Paladins push themselves, week after week, to beat other healers to the punch. But as frustrating as it is, I found myself wondering the other night, if taking issue with absorbs was really the right stance. Are absorbs really worthy of all the hate? Are they really to blame for the huge differential between healers, or has there been something else at work in Mists of Pandaria?
Themes and Effects
If triage was the central theme of Cataclysm, the design philosophy applied to every encounter, then I’d argue that the central theme of Mists was burst. Short bursts, rapid bursts, timed bursts and random bursts—the mechanics that we saw peppered throughput Tiers 14, 15 and 16 were designed not only to, in increasing intensity, push healers’ limits, but also test our capacity for self-tuning. And in contrast to the triage style of healing, which tempered careful resource management with self-regulation, in Mists resource management bit the dust relatively early on (for most classes, in late Tier 14), leaving us with really only two types of healing conditions: critical healing and non-critical healing.
Critical healing is something I think most healers inherently understand. It’s that time where you roll up your sleeves and dive into the thick of it. It is:
- Time-constrained and time-sensitive, meaning it needs to occur after a specific event and within a constrained window of time
- Demanding on healers’ resources, meaning it requires high mana consumption, high twitch, and precise decision-making
This type of healing is in direct contrast to non-critical healing, which would be defined as:
- Loosely impacted by time, meaning that the healing needs to occur but on a more protracted window
- Not demanding on healers’ resources, meaning it could be accomplished on a conservative and mana-neutral healing rotation, and that target selection “errors” have little to no effect
Roughly illustrated, these two types of healing represent opposite ends of the spectrum:
Now, in a triage state, for the duration of an encounter, you exist somewhere between the two extremes, fluctuating between them but tending towards a middle ground where you’re weaving inefficient spells into efficient rotations and vice versa. Triage healing is about flow, choice, and modulation. That isn’t to say that there are times in a triage-based fight where the healing window shrinks to be more immediate–even the first tier of Cata had its fair share of burst mechanics–simply that a triage game rewards finding the right mix of fast and efficient over the course of an encounter.
But, in a burst environment, that tenuous middle ground doesn’t exist because efficiency isn’t a priority. In a burst environment, you’re not focused on getting the most healing per mana, you’re focused on cramming as much healing as possible into the short window that you’re given. As a result, by virtue of the fact that you have periods of high intensity healing, balance demands that you accordingly have periods of almost pure regen. So, in a pure burst environment, the healing demands of an encounter alternate between one end of the spectrum and the other; meandering healing peppered by periods of incredibly intense activity.
You can see evidence of this all-or-nothing style of healing throughout MoP. It started in the very first instance released, in fights like Feng and Will of the Emperor. Choose almost any fight from Throne or Siege and you’ll find it there too. It’s apparent in the most dominant “heal” of the expac—Spirit Shell. It was unmistakably at the core of the major retool of Resto Druids’ Mushrooms. It is the reason that Resto Shaman were able to retain a raid spot Tier 15, despite our poor healing performance. Even the design of the legendary tank cloak acknowledges that massively unexpected burst was going to be so common that Blizzard actually designed a get-out-of-death-free card and handed it to tanks at the start of Siege. And capping it all off, the end boss of the expac, HM Garrosh, serves as a perfect example of a fight where healing is critical only in a few select moments, leaving your healers to trip over themselves for the rest of the 12-minute encounter.
Okay, okay, OKAY! WE GET IT. Burst was everywhere. Why the hell are you hammering on this point so much?
The Untruths We Believe
Because as a result of this prevalence of burst, these healing meters:
And these healing meters:
… are misleading the community into believing something that isn’t really true. (Hell, they were leading me to believe something that isn’t really true).
They are telling us that absorbs are incredibly more powerful (in some cases, twice or three times more powerful) than pure throughput. They’re telling us that the performance of throughput healers is vastly inferior to absorb classes. But, in this case, they’re wrong. And they’re wrong not because of that horrid “unseen benefits” argument that gets brought up so often during healer balance discussion, but rather, they’re wrong because they are incapable of distinguishing between critical healing and everything else. (This is actually something that GC touched on back in Ulduar, when talking about meters’ limitations. And yes, I do remember blue posts from 4 years ago!)
It’s all the non-critical healing noise, that fills in the lulls between incoming damage bursts, that’s creating such a divide. Absorbs can and will always get to this non-critical damage first, especially when mitigation classes can apply those absorbs from a mana-conservative rotation. But, when we look at meters and charts like the ones above, we forget that non-critical healing noise is exactly that—nearly irrelevant healing. It’s healing that doesn’t even need to be done by healers—it can be covered by non-healers’ splash healing (think: Shadow’s Divine Star, Enhance’s Healing Rain, a boomkin’s Nature’s Vigil, etc.), by players themselves, or even by the incidental heals that all healers are equipped with. It is healing that, but for the fact that you have an absorb class in raid, could and would be done by a multitude of other means. It is healing that has very little impact on the outcome of an encounter. (This is also why is possible that your “critical tank healer” isn’t necessarily the absorb class who absolutely dominates the healing done to a particular tank—but that’s a topic for another day).
More importantly, the difference between absorb and throughput classes is further exacerbated by the fact that you have fights with incredible swings in incoming damage. Consider this design problem:
My answer would be: you either employ a fight mechanic to rob them of their mana cushion (mana drain, player debuff, etc.) or force them to operate at 100% as frequently as their total available mana pool will allow. Unsurprisingly, the latter is exactly what fights tended towards over the course of the expansion.
But, here’s the kicker with that solution … while healers can dump mana at a highly accelerated pace, they don’t have many means by which they can recover it at an equally accelerated pace. Yes, you can employ Mana Tides, Hymns of Hope, Innervates, Divine Pleas, Shadowfiends, and the like, but not only are they on long cooldowns (except for fiend), but they won’t return an entire mana bar, which means your healers are still constrained by time. So, the higher your burst HPS requirements (the more times you force a healer to that 100% throughput mark), the longer the time period needed to recover that mana. In other words: the higher the burst in an encounter, the more downtime is required to balance it out. And the more downtime you introduce into an encounter … dun dun dun … the more non-critical healing noise is represented on those meters.
Still a little skeptical? Still thinking I’m trying to justify my raid spot? Let’s take a practical example.
Below is a chart of Damage Taken (by ability), from this past week’s HM Garrosh kill. Each one of the solid purple humps on the log below represents a 6-second window where Garrosh is channeling Whirling Corruption. Those 30 seconds total (5 humps at 6 seconds each) represent the largest single contributor of incoming damage in the encounter, outside of the boss’s melee hits. That melee damage, charted through the course of the encounter, is shown in orange; it is 18.52% of the total damage over the course of 12 minutes and 38 seconds. In contrast, Whirling Corruption, at 18.24% of the total damage of the encounter, represents damage which occurs in 4% of the total time.
Even if you add in the time where healers are dealing with Annihilate (two 30-second segments, shown in the dashed purple), you have 36% of the total damage of the encounter being done in 12% of the total time. In actuality, this is the reason that the fight is a 4-heal encounter in 25s and a potentially 1-heal encounter in 10s. Because, for the other 88% of the encounter, your healers are tripping over themselves to handle what little healing needs to be done.
As you might expect, given what I’ve talked about in the post thus far, the chart below is an example of what the overall healing meters look like for this encounter:
Pretty discouraging picture, if you just take it at face value. Our disc positively dominates the fight when you look at overall meters. (Two discs and I start to wonder if I should just AFK until my cooldowns are back up). But, narrow the healing window to the times in which that spike damage is going out, and you see something else entirely. For reference, here’s what the incoming damage looks like during the last two Empowered Whirling Corruptions, in Phase 3:
I’d qualify that as spike damage, wouldn’t you? (lol) And here’s what the healing meters look like during that time period:
Now, keep in mind that the segment above isn’t all critical healing, as much as I’d like it to be, because it includes healing done before and after those Whirling Corruptions. But, regardless, this picture is substantially different than the overall healing meter that I showed you for that same HM Garrosh fight. The picture above actually makes healing balance look surprisingly close, and presents a good argument as to why paladins might deserve a little more help in the burst healing category. But unless you know to dial it down to this level, you’re never going to see the healer balance that is being masked by the noise present in the encounter.
Ultimately, there are a number of lessons that can be garnered from the raids in Mists of Pandaria; lessons about the impact about stat inflation (which I’ve discussed before), lessons about healer regen, etc. But I think the two most important lessons that we can take with us into the next expansion are these …
First, that context really, REALLY matters. We hear this sentiment from developers almost every expansion, and certainly every tier where one class seemingly obliterates the others on overall healing meters. It was said about Icecrown, it was repeated in Firelands, and it was dusted off again in the first tier of MoP when Spirit Shell finally found traction with Disc Priests. Class balance is not a vacuum and it is not independent of the context in which healers perform. Granted, I don’t think this absolves developers and encounter designers of the obligation to make fights in which all healers can succeed, but likewise, it doesn’t absolve the community of our responsibility, when discussing balance, to address the context of our arguments. It isn’t enough to say “absorbs are too strong”, because clearly the meter just above this paragraph demonstrates otherwise.
Second, that the presence of a middle ground (in between non-critical and critical healing) is requisite to not only the perception of healer parity but also the realization of balance. Mists, for all its successes, also did one thing horribly wrong–it polarized healers into distinct camps at almost every turn. Healers still struggling with mana and those free from its shackles. Healers with massive raid cooldowns and those without. Healers with absorbs and those without. As a result, healing became less about what healers could do with the tools they had at their disposal, and more about if they had the “right” tools to begin with. There was no positive effect of having a diverse healing team, and likewise, there was less flexibility to work with what you had available. Instead of unifying healing teams, Mists made a clear distinction between the haves and have nots.
So, although there are a number of things to argue for when it comes to adjusting healer balance in Warlords of Draenor, I don’t think that absorbs are going to rank too highly on my list any more. As much as I’m irked by a meter dominated by Divine Aegis or Illuminated Healing, I think I’ve reason enough to believe that those are only symptoms of an underlying problem. Because if there is one thing that empowers all healers to change the outcome of an encounter, if there’s one thing that normalizes the contribution of a hot versus the contribution of an absorb, if there’s something that keeps an fight from being a completely automated series of cooldowns, it is an encounter that is devoid of extremes and non-critical healing noise. It is triage. And it is time for it to be back on the menu once again.