That's fine in principle Disaster, but that is why I asked about the power analysis (I know you largely agree with my perspective here but I want to explain further my reasoning).
I did a back of the envelope calculation on how much data a typical days worth of GoI playing will generate and it is way bigger than any data set I ever get to work with. ~200 players online at a time, typical game time plus lobby ~30 mins, ~20 players per game, so ~20 games an hour makes ~500 games a day. That means a month gives ~15,000 games to work with. I wish I had those kinds of sample sizes for my work.
I'm not one of the folks who is grumpy that stats play a big role in balancing, although I share the concern that stats are being used without sufficient context. I worry if the stats are of sufficient quality and specificity to do the job being asked of them.
I also think balancing for low level play is foolhardy, but Muse tell us it is a mix. Well I would like to know what kind of mix. Is it 60-40, or 90-10? Because if the balance changes are going to be almost entirely determined by novice game statistics then my input and the input of other experienced players is mostly a waste of time. I'm happy to help pull the tractor out of the mud, but I need to know enough of us are pulling towards the road. If most of the pulling is just towards an entirely different patch of mud then I'm not sure I'm of much use.
I would like to know a bit more about what those stats actually are and how big a role they play though.