Burnwinter Burns, at the end of the day, humans aren’t ones and zeroes. We do not reduce them to that, and I fear that maybe I have given the impression that we look at data as an end, not as a way to understand the world. The xG qmodels we look at are a way of understanding a very specific part of football - over the long run, what is the volume and quality of chances a team generates, and how well is a team executing against them.
Of course, based on that, one always has to go back to the people and fix them. Fix how they play, and how they interact with each other. The data just reveals what you need to do.
Typically what’s evident is that finishing isn’t the big driver of value - it’s just a variance and doesn’t have separation between great and good players. It lies in everything before you shoot - some of it captured by xG, some by other metrics. Did you as a group of players do enough to create enough chances in a good enough place. As you work backwards from would be goals, what goes wrong? It’s things like Nelson not taking a damn shot after that superb pass from Ramsdale early vs Liverpool. It’s Odegaard slightly over hitting a pass to Havertz who in turn over dribbles it, leading to a very wide shot. There’s a lot of stuff we are doing wrong, a lot of it can be measured.
And we should focus on the right things - better chance creation (ie, 20 shots instead of 10, 20% shots instead of 10% shots); not the variability of scoring or not scoring a 10% chance
What’s clear is that the team is largely fine. It’s literally just final third stuff that’s killing us. Creation and execution. I’d be extremely worried if we had big issues all over. We don’t.
