A few other random musings on Xg.
For the vast majority of players, their goals over time trend toward their Xg . In other words, there isn't a big variance between players in scoring more or less than their Xg would predict, and only a very rare and exceptional player can sustain scoring more than his Xg over time. If this is real, then it means there really isn't any such thing as expert "finishing". What distinguishes players that score a lot of goals has almost nothing to do with how well they shoot and is instead a factor of getting a lot of shots and the quality of those shots. That in turn is a factor of things like making smart runs, dribbling ability, being able to receive a pass, a good first touch, etc. But not shooting.
Most of the analytical shops will adjust a team's Xg so that no single possession can add up to more than 1.0. Think of the goal we scored that was ruled offside. It's possible that Keaton's shot and then Pellegrini's rebound goal together would have added up to more than 1.0 (assuming Pellegrini had not been offside). Of course, that doesn't make any sense, as there is no way we could have gotten more than one goal off that play.
There is a new goalkeeper stat that is based on the shot-adjusted Xg (or some such term). That's the expectation that a shot will go in based on all the things that make up Xg (position, defense, etc.)
and where the shot goes - e.g. to the corner, high/low, middle, etc. The stat then looks at how many goals a keeper gives up compared to what would be expected by his Shot-Adjusted Xg. Unlike field player, there
is a big difference among keepers on that stat.
We should consider an analytics thread for discussions like this. All I know is that
dummyrun would be very proud of us right now.
