Monday, April 9, 2012

Haste and the Myth of Mana Inefficiency

The concept that haste is an inefficient stat is not really new - it's been around since haste first came into existence. But lately, I've heard it cropping up more and more, and it's really starting to get to me. I think it's an easy claim to make on face but a much harder one to back up when you start looking at healing holisticly rather than on a micro level.

The argument these people make is that haste negatively affects your mana pool, because you cast more spells when you have more haste. There are several problems with this theory, and I would like to address them.

Overcasting

The first glaring hole I want to punch in this theory is very simple to grasp. The idea that a player with 20% more haste - let's use a shaman as an example - is capable of casting five Greater Healing Waves instead of four in a given time period. This is, obviously, a perfectly reasonable thing to say.

But jumping to the next conclusion - that this player actually will cast five GHWs - implies that the healer is brainlessly slamming his GHW button at all times. It may be true that there are healers who do this, but if you are theorycrafting, odds are quite high that you are not talking to these healers. Those healers probably aren't interested in high-level thinking, clearly, because they have already failed to research or simply aren't putting any research into practice.

So who are you talking to, then? You're talking to healers who pay attention to whether they need to heal at any given time. To suggest that this healer will spam GHW five times is utterly incorrect if only four are needed to stabilize her raid. She isn't stupid; she's not going to needlessly waste mana.

Spell Selection

A second fallacy present in this argument is the removal of spell selection from the equation. It pretends that a healer is going to utilize the exact same ratio of spells regardless of how fast that healer is throwing them out.

Building upon the previous scenario, imagine our shaman healer has a situation where she needs to heal 20,000 damage in 30 seconds (bear with me, this is just an illustration) . Assume HW costs 1,000 mana and heals for 1,000 and GHW costs 3,000 and heals for 2,000. Both spells take 2 seconds to cast. GHW will accomplish her goal in 10 casts for 30,000 mana over 20 seconds, whereas Healing Wave will take 20 casts and is cheaper at 20,000 mana, but requires 40 seconds to meet the necessary healing totals - meaning someone dies. In this environment, she has enough time for 15 casts of a combination of these two spells. Is she going to cast 15 GHWs? Of course she isn't; she is smarter than that and knows she needs to manage her mana carefully. She can cast 10 HWs and 5 GHWs for a total of 25,000 mana and the raid is saved, as well as 5,000 mana.

Now imagine this shaman has more haste. Instead of casting those five GHWs, maybe she can squeeze in 17 casts instead of 15. In this example, she could fit in 14 HWs and only needs to cast 3 GHWs as a result. The net is that even though she had the same amount of time and has technically casted more spells, she has used less mana (23,000) than before. So tell me again how haste hurts mana efficiency?

Reacting Faster Costs Less Mana

This point is a bit more difficult for some to wrap their heads around, but it's pretty simple when you consider the consequences to reacting slowly to incoming damage, especially when it's unexpected. Faster reaction times tie in intimately with spell selection and mana efficiency, because it's very easy to get behind on healing and then you are forced to resort to spending more mana to do the same amount of healing over a specific time frame.

Also necessary to consider is the cost of a player dying if you do not have throughput quick enough to save them in a low-health emergency. Someone must resurrect that player, and if it isn't a druid, you are going to have to heal that player back up to full, generally as quickly as possible, especially if that player is a tank. That means you've spent one of your battle rezzes (possibly your only one, in 10-player raiding), meaning that later if something else goes wrong, you're hard up. On top of that, if you are a class that brings a casted buff, you need to recast that buff if you want the player back up to peak efficiency. If another healer rezzed the player, that's more healing (and thus more mana) you will be spending to cover them for a few seconds while they rez. If it was a DPS, your raid has lost a few seconds of DPS which will mean you have to heal through a little bit more of your encounter. Worst of all, you may have cost yourself another player entirely, if all of your rezzes have been used, and this is by far the worst possible mana drain you can have in a raid environment, no matter what role the player was.

Haste Breakpoints and Efficiency

Sometimes haste only does two things - makes your casted spells faster and makes your global cooldown shorter. However, sometimes haste does another thing - adds an additional tick to a heal-over-time spell. This is when the case for the inefficiency of haste goes straight out the window. You can't argue that another tick of a spell isn't efficient, ignoring the other impacts of stacking haste in favor of other stats. There are certainly times when stacking haste to hit a breakpoint for a new tick is a bad idea - when it comes at the cost of too many other secondary stats, or when it comes at an intellect cost - but generally speaking, most specs currently aim for a haste breakpoint of some kind. Holy has their 12.5% breakpoint, resto shaman have 916 for the extra Riptide tick (some even aim for 2005, though I understand this is fringe), and most resto druids aim for 2005 for an additional Wild Growth and Efflorescence boost (hitting breakpoints for Rejuvenation and Lifebloom along the way). All of these specs do so because the net healing gained from these additional ticks end up making their overall healing stronger as well as making their signature spells more efficient.

Healing on the Move

Sometimes you don't have time to take the luxury of using long cast-time, low mana heals, so you default to casting more expensive, faster heals. This is, in a word, inefficient. When you have to move, this is especially true. Say our no-haste shaman knows that in a few seconds Goriona is going to cast Twilight Flames during phase 2 of Blackhorn, but the tank needs a heal, fast. She has the option of moving first, then casting, but during that time the tank might die. So now she has two options (yes, I'm ignoring Tidal Waves here for the sake of analogy): she can either move first and hope the tank makes it until she can stop moving and cast, or she can quickly cast Healing Surge and then move, exchanging mana for speed.

Now let's look at this occurrence again, but with our hasted shaman. She knows that she has 1.5 seconds, but luckily for her, she has enough haste that she can squeeze off a Greater Healing Wave instead of a Healing Surge, doing more healing for (roughly) the same amount of mana. Assuming her GHW doesn't overheal and accomplishes the same task of stabilizing the tank, she has come out more efficiently.

Sometimes Mana Just Doesn't Matter

Even though it doesn't necessarily help my thesis that haste isn't mana-efficient, there is a very important factor to weigh in the debate between stacking haste and stacking another stat: sometimes you need healing now and it doesn't matter how much mana you burn in the meantime.

Think of the last 30 seconds of Heroic Ultraxion. You aren't interested in mana efficiency at this point. Likely our shaman from above has the blue crystal (or perhaps green depending on group makeup) and has no concern for mana, simply needing to dump as many Chain Heals and Healing Rains onto the raid as possible before someone dies. Maybe she's finishing off the final phase of Madness and she's just pouring all of her mana into keeping the tanks alive through stacks of Tetanus. Maybe she's solo healing her side of heroic Morchok on 10-man through the last 20% and all that matters is getting the raid topped off before a stomp. In these situations, all the mana efficiency in the world won't matter if you can't put out enough raw healing.

Brass Tacks

The end result of this discussion is not to tell you to stack haste blindly. It's not to sway you that you'll be a more efficient healer if you stack haste. I'm only trying to put down a concept that is true in simple but much less accurate when viewed from a wide angle. Haste does not necessarily make you a less efficient healer, and this is especially true if you are a very proactive healer capable of taking the long view in a situation and planning out in advance when you need to forego cheap, low-healing spells for your big-heal mana-drain spells. So can we please stop acting like it does?

-Rush

Addendum

This final section is really only applicable to shaman (and maybe to a lesser extent Archangel priests), but since I've been using one as an example, I'm sure someone will make the argument. "What about Telluric Currents? That shaman could cast 10 GHWs in 20 seconds and spend the other 10sec regenerating mana with Lightning Bolt!" It's a somewhat valid response. At the moment, in fact, because of how good TC is as a mana regen mechanic, this is actually the most effective strategy. But at earlier gear levels, TC was not so strong and this would have been a bad choice.

More importantly, though, that argument makes even less sense because of one key factor: haste will make you regen more mana over a given time period by allowing additional Lightning Bolt casts anyway.

No comments:

Post a Comment