Superb again Beez. Do you think Klopp and Ljinders were supported by our data team in there low value shots approach? Or do you think the manager says "this my approach" and the data team build a model in which to execute it? Either way by the time we figure that out Erling Haaland will have earned another few million quid for scoring a few more goals.
Thanks mate. I suspect last season it was a combination of things - teams that are behind tend to shot more and from worse locations, so probably a bit of that. But also, they’d bought Gravenberch, Mac Allister and Szoboszlai, who - my gut feeling is, I’d have to check - are more inclined to shoot from long range than their predecessors were. It may all have been slightly random rather than a specific tactical choice.
It factors in the placement within the goal frame once a shot is put on target. Yesterday provides a good example:
Diaz scored the opener. He had a header which was 0.23 expected goals, but because he put the shot in the top corner, it was worth 0.92 post-shot expected goals. In other words, if a player puts that chance where he did, he can expect to score 92% of the time.
In the second half, Zirkzee had a header of broadly similar value (0.33 xG). However, he hit it close to Alisson in the centre of the goal, so the post shot xG for his effort was only 0.17. Where Diaz increased the value of the chance by placing his shot well, Zirkzee made his worse.
Really good Andrew 👍
Thanks!
Very good detail. Hopefully Mo bombshell on Sky will get Hughes off his arse
Superb again Beez. Do you think Klopp and Ljinders were supported by our data team in there low value shots approach? Or do you think the manager says "this my approach" and the data team build a model in which to execute it? Either way by the time we figure that out Erling Haaland will have earned another few million quid for scoring a few more goals.
Thanks mate. I suspect last season it was a combination of things - teams that are behind tend to shot more and from worse locations, so probably a bit of that. But also, they’d bought Gravenberch, Mac Allister and Szoboszlai, who - my gut feeling is, I’d have to check - are more inclined to shoot from long range than their predecessors were. It may all have been slightly random rather than a specific tactical choice.
I think you're right. And then there's Darwin Nunez.
Thank you for the nice piece.
Could be possible to elaborate a little on the PSxG concept..?
It factors in the placement within the goal frame once a shot is put on target. Yesterday provides a good example:
Diaz scored the opener. He had a header which was 0.23 expected goals, but because he put the shot in the top corner, it was worth 0.92 post-shot expected goals. In other words, if a player puts that chance where he did, he can expect to score 92% of the time.
In the second half, Zirkzee had a header of broadly similar value (0.33 xG). However, he hit it close to Alisson in the centre of the goal, so the post shot xG for his effort was only 0.17. Where Diaz increased the value of the chance by placing his shot well, Zirkzee made his worse.
Hope that makes sense! Figures from here: https://fbref.com/en/matches/a7ab7a12/North-West-Derby-Manchester-United-Liverpool-September-1-2024-Premier-League
Great! It makes sense indeed!
Thank you 🙏
No problem at all. I have thought about setting up a chat thread for that sort of query or general stats questions... it's on the to-do list anyway!