Pictomancer is insanely overtuned and needs to be fixed before savage drops.

[Lyth]

IQR

I'm also interested in the use of descriptive statistics to understand how player skill is distributed, but I think it's important to reflect on the limitations of this approach. I've tried using your approach previously, but I don't think it works quite the way that you'd expect.

First, there is no direct measure of player skill. What you can measure is DPS. Different jobs award DPS differently. While you can compare the performances of two players on the same job, it becomes much harder to correlate two performances on two different jobs. How does a 90th percentile PCT correlate to a 90th percentile GNB, for example? You're essentially comparing apples and oranges.

The next issue that arises is that the IQR is really only interested in the 25th and 75th percentiles. You have no idea what happens outside of those percentile ranges. What it tells you is how average performances vary with internal job changes, rather than letting you compare high skill play on two completely different jobs. For example, let's say that you increased the benefit of a job's damage up buff. That's probably going to benefit a player who knows their optimal opener and burst moreso than a player playing freestyle. So you'd increase the IQR without changing the difficulty of the job. What you've changed is the reward structure. You'd also increase the degree of positive skew and positive kurtosis, both of which are worth considering in your analysis as well.

But the conclusions in here precede the statistical analysis anyways.

Either way, I think the point here is not actually about comparing personal biases about relative job difficulty, but rather the fact that PCT has a lot of utility. I don't really mind if you want a high mobility, freeform job as your flagship caster instead of BLM. But you should pay a price for that utility and versatility, or lose it.