The Level 90 Mage Talents

The level 90 Mage talents — Invocation, Rune of Power, and Incanter's Ward— were controversial in beta, and they continue to be controversial in the Mage forum now that people are hitting level 90 in the live game.

For a bit of history, the level 90 talents were introduced in the second major version of the Mage Mists of Pandoria talent grid. The Polymorph tier had been widely panned when the original grid was introduced, and many Mages had also made clear that we preferred to have thematic options available; many Frost Mages wanted Frost talents, many Fire Mages wanted Fire talents, etc. However, there was also general agreement that certain Arcane utility spells felt sufficiently universal that they didn't present a thematic problem for any spec. Polymorph was one such spell, but very few people liked those talents. A tier of talents modifying Evocation was a popular suggestion, and Blizzard ran with that idea.

A fair number of people have expressed displeasure at the result. There are a few recurring arguments being made about them, and in an effort to shed a little light on the discussion, I'd like to address some of their points.

“They’re No Fun”

I'm not really going to argue this one. Fun is a subjective thing, and if you don't find any of these talents fun, I'm not going to tell you that you're wrong. However, it is my hope that some people will look at them differently if the rest of this article helps them understand a couple of things about the talents better, and sometimes a new viewpoint is a pathway to enjoyment.

“They Limit Mobility”

There's no denying that — at least in the cases of Invocation and Rune of Power. Invocation is going to require that you stand still and get off an uninterrupted channel lasting about 4.5-5.5 seconds every 40 seconds or so, and Rune of Power makes you want to spend as much time as possible not moving. And there is no question that both modern encounter design and PvP demand a lot of movement, so anything that limits movement can be a source of frustration.

But while Mages are intended to do excellent damage, and to be highly mobile, they are not intended to do both at the same time. This was perhaps stated most clearly in a developer response to Hunters during beta testing:

Movement should be terrible for a ranged spec. Full stop. All of the various mechanics we put in from Spiritwalker's Grace to Aspect of the Fox are to make moving less terrible, but it should still be pretty terrible. (And if you think there are specs not penalized enough by moving, please let us know, though probably not in this thread. We felt that ranged movement got a bid out of control in Cataclysm so we've definitely made an attempt to scale it back.)

What Invocation and Rune of Power do is emphasize these two key design elements of the class: they improve your burst damage (more on that later, because I know there are already people shaking their heads at it), and they increase the difference between the damage we can do while being highly mobile and the damage we can do when we can plant our feet.

If a hindrance to mobile damage-dealing is simply intolerable to you, there is still the option of Incanter's Ward, which has no effect on your mobile damage whatsoever. It does, however, have its own trade-offs, in that maximizing its damage potential is not entirely in your control and requires that you soak damage.

“We Can’t Choose the One We Like, We Have to Switch for the Encounter”

It's true that each of these options will be better suited to some encounters than others. And for those of us character-design purists who think of talents as inherent traits of our characters, things that can't simply be swapped around like gear, that can be a significant annoyance. But the truth of the matter is that the talents have been designed for easy switching, and the devs see them as being used that way; one of the goals of the new design was that people would interact with the talent system on a regular basis.

That's not a design failure, it's a design goal that unfortunately happens to be incompatible with the way some of us prefer to play the game. If, like me, you refuse to play the game that way, you'll need to pick the one you want to stick with, and do your best.

“We’re Balanced Around Using Them”

Or, to put it more completely, the argument usually goes something like: "We're balanced around maintaining near-perfect uptime, so all we get for performing these extra tasks is the same damage everyone else is already doing without the hassle."

There are a couple of things wrong with this, the first of which is: even if true, that's how specs are built. You're gradually given additional abilities, and you need to use those abilities with a reasonable degree of competence in order to produce your expected damage. And I very much doubt that if you stack up the number of abilities that we need to use in our DPS process, including the L90 talents, against those of other classes, that you'd find the Mage specs to be an outlier with excessive buttons to push.

But more signifcantly, the statement isn't actually true. Oh, certainly, if you're only looking at sustained DPS, these abilities allow you to produce your expected sustained DPS and very little more than that. But as we all know, sustained DPS is not the only kind of damage that matters. Bosses have vulnerability phases, adds spawn that need to be burned down quickly, and so forth. This is why we value burst damage, and burst damage is exactly what these abilities provide us.

Incanter's Ward is pretty obvious: it provides a continuous passive buff, and then has a trinket-like active buff on a very short cooldown with a short duration. This is easily recognizable as a DPS cooldown; just like a trinket you pop it whenever you can, but if a burst damage window is coming up, you might delay using the cooldown to ensure that it's available for the entire burst window.

Now this is what many people are having difficulty with: Invocation and Rune of Power are also burst cooldowns. It's not easy to recognize them as such, because they have been turned on their heads: they are up more than they are down, and your timing decisions are more about when they are down, as opposed to a trinket or Incanter's Ward, where such decisions are about when they are up. But as I'm going to illustrate with some simple numbers, burst cooldowns is exactly what they are.

Uptime

Assuming you have 20% haste, your cast time on Evocation is 5 seconds (6 / 1.2 = 5). Perfect uptime would therefore involve evocating for 5 seconds, dealing damage for 40 seconds, and then repeating the cycle, resulting in an uptime percentage of 40 / 45 = 88.89%.

Now, the devs are not stupid or sadistic. Their description in beta of how these talents were balanced was as follows:

Are Mages going to be balanced under the assumption that they have perfect 100% uptime of their level 90 talents? No. Are they going to be balanced under the assumption that they take one of them and use it reasonably well? Yes.

Some people have argued that it's still going to be near 100%, because any clod can watch a timer and know that it's time to Evocate. But it isn't just about being smart enough to use the abilities, it's also about how practical it is to use them under various situations. The devs are fully aware that there are encounters that will reduce the amount of time you can spend standing in one place, or may randomly interrupt a channel. They don't ignore things like this while they design abilities; they account for them when deciding how much uptime qualifies as using the ability "reasonably well." We don't know what that value is — and indeed, there probably isn't a hard and fast value, since there are so many variables.

I'm going to assume for the sake of this discussion that they expect an Invocation Mage to be able to use the ability to 80% of its potential maximum. If you feel that's unreasonable, you can follow along with my math here and revise it using your own assumptions.

At optimal usage, as I said above, Invocation uptime is 88.89%. This means you're spending 88.89% of the time dealing 125% damage, and 11.11% of the time doing no damage at all. Your average damage is therefore:

125 * 0.8889 = 111.1125% of the damage you do without the buff

So effectively, the talent is increasing your damage by 11.1125%. If you're using it at 80% of its potential, you're seeing an increase of:

11.1125 * 0.8 = 8.89%

Now I'm going to do Rune of Power a little differently. Let's assume Blizzard wants the two talents to produce a similar sustained benefit. How do we get an 8.89% DPS increase (or close to it) out of Rune of Power? Well, first, let's assume you need to use two runes, so you're casting the spell about twice per minute. With 20% haste, it has a cast time of 1.25 seconds, so perfect two-rune uptime would be:

58.75 / 61.25 = 95.92%

So you're spending 95.92% of your time doing 115% damage, and 4.08% of your time doing no damage at all. Average damage is therefore:

115 * 0.9592 = 110.308% of the damage you do without the buff

And you therefore need to use Rune of Power to:

8.89 / 10.308 = 86.24% of its potential

Sustained DPS and Reward for Optimal Play

So let's pretend you're in Tier 14 Normal gear, and balanced DPS is in the neighborhood of 85,000 DPS. Here's what we get from those uptime assumptions:

So far, it looks pretty much like the complaint says, doesn't it? You need to use the talents pretty well just to do your expected level of sustained DPS, and using them perfectly only lets you squeeze out a very small bit extra. So what's my problem with the argument?

Sustained DPS vs Burst Windows

Here's the basic thing people are missing: by focusing on sustained numbers, they're ignoring the fact that while the buff is active, your damage is significantly higher than the sustained average at which you're balanced.

To illustrate, I'm going to add a row to that table:

By way of comparison, a Mage with Incanter's Ward would be doing 82,744 DPS (97.35%) while running on the passive buff alone, but 101,478 DPS (119.39%) for the duration of her active buff. After using the active buff, there would be ten seconds of dealing only 78,060 DPS (91.84%) until the passive buff kicked back in.

So as you see, while your average, sustained DPS is balanced just as any other class's at 85,000, at any given moment when you are actually dealing damage, you are putting out significantly more than that average sustained value.

Now I want you to think about what you do when you're approaching a burst window. Just like any other non-Mage class, you're preparing to use your trinkets, you're getting ready to use your DPS cooldowns (Arcane Power or Icy Veins, maybe if you're Fire you're holding on to your next Combustion). And then when the burst window starts, you cut loose with that stuff.

With these talents, you're doing something else as well: you're making sure they're going to be up for the whole window. If you're approaching a 15-second burst window, and you have 10 seconds left on your Invocation buff or a Rune you may need is about to fade, you're going to refresh those buffs early. This costs you a bit of sustained DPS, but burst windows tend to be more important, so it's worth the trade-off. You do exactly the same thing with trinkets and Icy Veins, trading sustained DPS by activating them late so you can use them for burst.

And then, when you hit that burst window, you're not starting at 100% of your sustained DPS like other classes are. Before you even activate those trinkets or cast your cooldowns, you're hitting the burst window with 114.79% or 105.61% of your sustained DPS.

What the developers have effectively done with these talents is given Mages higher damage, without giving us higher DPS. This is useful, because there are times (such as burst windows) when short-term damage is more important than long-term DPS.

Another Way to Look At It

Some people had trouble seeing how this works; they see DPS, and their heads automatically go to sustained numbers. From their perspective, this doesn't qualify as burst because when you're not refreshing the buffs, it's the damage you're always doing. Some, I think, found it helpful to look at it not in terms of DPS, but it terms of actual spell damage.

So imagine for a moment that you play an incredibly simplified version of the Mage that casts a single spell. This spell deals an average of 156,120 damage without any of the Level 90 talents, and it takes two seconds to cast, so you deal 78,060 DPS.

Now, if you think of the level 90 talents purely in terms of sustained DPS, you might prefer that they were simply removed, and replaced with a passive 8.89% increase to damage. This would increase the damage of your spells to 169,999, and you would be dealing 85,000 DPS — the same sustained DPS as the talents currently give you, but without the extra hassle, right?

But let's look at what happens to a 20-second burst damage window with those three options:

As you can see, while all three designs produce the same sustained damage of 85,000, our L90 talents produce significantly superior results in a burst window. These gains will be amplified if there are raid effects or other buffs that further increase damage in the burst window, since the higher multipliers will apply to all such increases.

Conclusion

It may be difficult to see an always-on buff as a burst damage increase, but as you can see, that's what they are — strange, inverted burst cooldowns. Just like any other cooldown, we plan for windows of time when we need them, and adjust our usage times accordingly, sometimes sacrificing a bit of long-term sustained DPS in exchange for increased short-term burst. But unlike any other cooldown, we do so not by delaying usage after an item comes off cooldown, but by sometimes needing to refresh a buff early to ensure that its duration is sufficient to be up when we need it. The end result — higher per-cast damage in those windows when that matters more than sustained DPS — is the same.

While I realize this is not the case for everyone, and I respect the fact that some people simply have different tastes, I think there are some people who dislike these talents because there is no actual gain from using them, just an added chore that permits you to perform up to par. I hope this explanation has cleared up that misconception, and if so, that it helps more people find enjoyment in these talents.