What's interesting to me is looking back on how we got to this point. Blizzard doesn't hate us (people who claim this confuse me, to be honest), and I literally cannot think of a single nerf to moonkin dps in wotlk. So how, with fairly positive attention, did we end up near the bottom of the dps charts with two soft caps to juggle? The answer, I think, is to look at something comparable to moonkin dps - a national disaster. In this case, Three Mile Island.
I'm not going to go over that here, but suffice to say that the disaster wasn't caused by someone accidentally hitting the big red button marked, "Cause A Nuclear Meltdown". Disasters rarely are - instead they are caused by multiple small decisions/events that compound upon each other, and before you know it you're knee-deep in radioactive waste (or haste, as the case may be).
So what decisions started the current moonkin chain? Let's go back to TBC...
I blame Destruction Warlocks. Not their fault, but two tiers of shadowbolt spam really made Blizzard want everyone to have interesting rotations. For Balance Druids (who ignored Wrath almost entirely at that point) this meant coming up with a talent or spell change that would force us to use both at some point during the fight. Blizzard sat down at the design table and after much thought came up with the core idea of Eclipse - one spell (Starfire or Wrath) would proc an effect to buff the other spell.
Now, eclipse is a great idea with poor execution. The original version was a dps loss, so it was buffed repeatedly before WotLK until it was a pretty ridiculous amount of moonkin dps.
2) Nature's Grace Changed To 20% Haste
Nature's Grace was always a very strong dps talent. Changing it from a flat .5 second reduction to 20% haste was a welcome change (letting Wrath scale a bit better with haste), and the 3-second duration was also a nice dps buff.
The problem with this change was it didn't go far enough. 20% haste is certainly better then .5 seconds for Wrath, but it still leaves us soft haste capping at only 400 - a number easily reachable in even entry-level raid gear. More importantly the 3-second duration means that a single Wrath crit now caps out three Wraths, aggravating the issue significantly.
3) Separation Of Lunar/Solar Cooldowns.
This was an interesting change because for the most part it was unasked for - the only real conclusion I can make is that Blizzard felt we were falling behind from internal data, and they decided to correct it.
This is a pretty positive change in two respects - it smoothed out Moonkin dps and it was a straight dps increase. Instead of having 15 seconds of burst followed by 20-25 seconds of "normal" dps, we got to rotate between burst phases. The downside to this change is it forced the soft haste cap (400) onto every moonkin. Before you could use a Lunar rotation and cast Wrath relatively rarely, which meant haste didn't effectively soft cap at any reachable point. Separating the cooldowns means Wrath will always be cast nearly as much as starfire, which makes the soft cap more important.
4) Trial of the Crusader
There is nothing directly wrong with this instance, but it did almost single-handedly cause a pretty big issue: item level inflation.
TotC was "unplanned" - that is, it's a filler instance so people had something to do before Icecrown. However, in order for people to feel it was valuable the items had to be worth it, which meant introducing a new tier. This was the first inflating step. The second was that in order to further differentiate ToC items from Ulduar HM ones, Blizzard added another half tier. I'm not sure why they did this, but it has pretty significant implications because the same pattern was kept in ICC.
The end result is that current Icecrown Gear is two full tiers ahead of where we (and Blizzard) thought we'd be. That is a massive amount of stats, and as a class that doesn't necessarily scale as well with them that's a problem.
5) Buffing Eclipse
When Blizzard decided to fix WiseEclipse (which was both fair and expected) they decided to buff moonkin dps at the same time in order to prevent a nerf. While this was nice, I'm going to look the gift horse in the mouth and state that they did it in almost exactly the wrong way - by buffing Eclipse. With this change the first point in Eclipse - that is, when I go from 0/3 to 1/3 - is worth over 1,000 dps, almost as much as 3/3 Nature's Grace or 5/5 Starlight Wrath.
Now, look back at all of these changes - nothing in there looks like a nerf. They are all either buffs or (in the case of a new instance) something that should still increase our damage. Nothing is drastic or massive, either - they are all relatively small changes.
The commulative effect, however, was to place in inordinate amount of dps into a single talent, and one that has significant issues. Especially since as raids progress, the one constant that always increases on average is movement - you move more in Icecrown then you did in Ulduar, and you moved more in Ulduar then Naxx.
Secondly, item inflation and repeated buffs to Eclipse left us hitting a soft crit cap in Icecrown, an issue that's only important because we are two tiers ahead of where we should be.
Third, the legacy talent Nature's Grace has enforced a soft haste cap practically from Naxx, and one that we can't avoid because of other changes.
Hopefully Blizzard has already learned all of this - they have the data, and I am confident that they are mostly smart people. But a few things should be obvious:
- Too much dps in a single talent is bad. Eclipse and Nature's Grace are the core cause of almost all moonkin scaling issues, simply because they are too powerful. Eclipse doubles up and also hurts us more then most other classes when we are forced to move.
- Itemization scaling is important to control. Blizzard has already said they'll address this in Cata, so I am not too worried that it will repeat itself as a problem.
- All buffs are not created equal. Buffing Eclipse may have increased Moonkin dps, but it also aggravated the issues we were having or introduced new issues. Buffing weaker talents may have solved those same issues without these complications.