Twice as Often, Half as Much

#0 - June 17, 2009, 9:02 p.m.
Blizzard Post
Although I usually stay out of discussions on class balance and just how much DPS class x should do, I'll use the current feral DPS debate as a framework for making a few game balance suggestions..

Since the post by Ghostcrawler suggesting that feral DPS might be "a little too high", there have been lots of postings and data analysis on the topic of feral sustained DPS, and whether or not nerfs might be warranted. Just to summarize that data quickly, there are two major areas of contestation:

1). WWS scoreboards and world-best DPS measurements - These get brought up a lot. Basically, they indicate that the very best feral druids in the world, in a raid stacked to emphasize their damage, are roughly where they should be on most fights in Ulduar 25man, compared to Rogues and other 'pure' DPS classes. This type of data is useful and interesting, however it certainly has its limitations. It does not represent the 'average' or 'poor' feral druid, nor does it represent fights where the raid group was really challenged by the boss fight (i.e. first-time kills or even learning the fight).

2). Anecdotal and 1-to-1 comparisons of players - While this data is often hard to generalize, there have been some interesting trends noted. In particular, a lot of players talk about undergeared feral druids performing significantly above where other classes and specs might perform with an equal gear set.. in short, the 'skill factor'. The suggestion here is that feral DPS 'scales' well with skill, but poorly with stats. Unfortunately, skill is much harder to quantify than gear. What's really needed in terms of the data here is something like a WWS worst-25.. but not only does such a dataset not exist, screening out events like player deaths, disconnects, etc. would take a significant amount of work. That aside, I'd love to see that data, if anyone's up for a challenge ;) Basically, what we need is a measurement of the skill factor for each class, or the total range of DPS values on any given fight possible (while still killing the boss) while keeping gear averaged.

The skill factor itself is often contested as well.. many druids feel that because of how difficult the feral sustained DPS rotation is compared to other classes and specs, it is perfectly acceptable for ferals to do "very high DPS" compared to other classes with easier rotations to manage. It is definitely true that feral sustained DPS is difficult.. it's often compared to Shadow Priest DPS rotations (and no, I'm not going to debate which is more complicated or harder to do). Unfortunately, however, as many players and devs have mentioned, classes and specs which require more skill and effort lead to a wider range of overall DPS values. In short, a low-skilled player of a difficult class will perform much worse than a low-skilled player of an easier class, given the same gear level. So adjusting the skill factor of classes and specs is often just as important as adjusting their 'average' DPS. Personally, I wouldn't ever advocate for making a class or spec 'easier' to play. As many players have mentioned, even 'difficult' classes and specs aren't *that* difficult. Plus, the developers have noted on several occasions that various classes and specs don't have to make the kinds of 'choices' in their actions and rotations that they were intended to make (e.g. warriors and not being properly limited by rage).

But what was it that led to all of the debate over feral DPS being too high, anyway? After all, these posts didn't exist early in WotLK. Let's look at the changes that feral druids saw in patch 3.1:

Primal Gore - a 1-point talent, Primal Gore is a massive boost to both sustained DPS and scaling (particularly scaling with Agility and Crit Rating). This talent alone was roughly a 9-13% DPS boost to feral sustained DPS.. and is 9-13x too powerful for the developer-stated talent point budget (each talent point, as a goal, should give an approximately 1% increase in DPS).

Savage Roar (+40% AP switched to +30% damage) - This change was roughly a 2-3% DPS increase for most Druids, although it also served to slightly reduce feral AP scaling.

Glyph of Savage Roar - Not only did feral DPS druids only have two DPS glyphs pre-3.1, even at a +3% (additive, rather than multiplicative) buff to Savage Roar, this glyph is roughly twice as strong as any other feral DPS glyph, or a 2.3% DPS increase.
#13 - June 18, 2009, 12:44 a.m.
Blizzard Post
Nice thread. :)

We don't think cats are crazy overpowered and given the complexity of the rotation they certainly aren't facerollers. I also agree that there are very good fights for cats and very bad fights and the people that QQ that cats are too good tend to forget about half of them.

If we nerf damage, it won't be major, and we would like to get a lot of PTR testing so we can feel confident about the numbers (in the "real" world -- we do our internal testing of course).

You are correct that Primal Gore is a big dps boost (I'd say more like 10% but your mileage may vary) and that is more than we typically want a talent point to provide. We added it to bring up cats because they were pretty far behind. Baking the crit into the dot attacks is attractive and something we might eventually do, but at the same time one of the reasons that cat form alone is relatively weak is because it is available to Resto and Balance druids. We have to rely on the deep Feral tree to boost Feral dps a lot.