Don't know what you mean, mine are beautiful and perfect

Go on, I need to know I'm not the only one

Maybe Levy was the pain sponge all along

I just built / stole a toy model to show that at the same goal difference, lower xg has less variance. in r here: snippet.host/efpxhw Should be advantageous for better teams to reduce overall xg, all else being equal.

Have you looked at whether there are stochastic advantages to being a team that on average wins xg by say 2.0 to 0.7, rather than 3.5 to 2.0? In which case Pep's insane control Frank thing is good actually (derogatory)