Some people would be awful at analysing data if they did.
The point I was trying to make, which you are deliberately missing, is that it's too simplistic to look at games he played in vs games he didn't this season because by definition they're different games and the sample size for the games he didn't play is pretty small. If he'd played 50% of games and there was a clear difference then it might be a more valid argument.
TL;DR, you need to be a bit careful with statistics.