The key is repeatability in the stats. You can play wide, you can play with a high line, etc. But there's loads of games to look at, and you can see how one dec

The key is repeatability in the stats. You can play wide, you can play with a high line, etc. But there's loads of games to look at, and you can see how one decision works vs another, because there's repeated experiments. Granted, modelling isn't some sort of objective thing, but you have better comparability. Business is somewhere else on this scale.