In the 17/18 season we were expected to score 46 (we scored 44) and concede 54 (we conceded 64). In 18/19 our xg was 52 (we scored 52) our xgA was 63 (we conceded 59). West Ham and Newcastle were significantly luckier than other teams outside the top 6 last season based on xg/xgA. One of many reasons Newcastle look like dead men walking next season.
The main thing that jumps out at me is the massive improvement from dropping Gomes and starting Foster. We went from conceding way more goals than we should have done under HG, to a handful fewer under Foster. Still, at least we played Gomes for the biggest match in our history.
Can someone explain - in layman's terms - what 'Expected goals XG/XGA' actually means & how the figures are worked out? It means nothing to me.
It's somewhat subjective but it's basically a situation where u should score. We created 52 situations where we should have scored last season. I'm not sure how many we did score but we balanced the xg's we missed with non xg's we scored such as a few of del boys wonder goals. xgA is slightly different as it's a combination of luck and pressure on the opposition plus their inability to finish. To let in 10 more than expected in 17/18 is really bad but it also meant we could improve relatively easily by tightening up a bit at the back which we did last season.
It’s an improvement on shots/ shots on target, as it allows people to see which team should have won, by looking at clear goal scoring opportunities where a team ordinarily would have scored/ conceded. To measure this, analysts looks at distance from goal, angle of shot and shot type, among The number of players in the cone between the ball and the goal, ball height etc. Out of interest, a penalty is worth 0.75 of an expected goal. Shots on target didn’t tell you enough about the quality of the shot. It would be misleading to say a team which took ten pot shots from 30 yards and barely troubled the keeper, should have won. If a team should have scored 7 goals and should have conceded 1, but lost 0-1, then you can say that either the away keeper played amazingly or, the home sides shooting was off, or both. This would have to be confirmed with additional evidence (generally by looking at a greater number of games). Data scientists have shown that this information can provide much greater insight than traditional methods. Dortmund sacked Klopp because he was performing very poorly in the Bundesliga. But his team were still creating way more clear goal scoring opportunities than they were conceding, they were generally on a really unlucky run. Liverpool’s data scientists saw this, and convinced the club to sign him. I quite liked this article on football analytics if you’re interested in reading more. https://www.reddit.com/r/FantasyPL/comments/c1qkgt/i_attended_a_statsbomb_course_about_football/
My expected goals for this season are to not get relegated , beat Everton, Muff and Spam, and not to have to wait more than 5 minutes to be served in the Rookery Bar.
Let us know your minutes per beer stats for next season and expected beers v actual beers. Last night my xb was 3 but I got 5. Was it luck? was it skill? I'm too hungover to work it out.
But then he scored vital goals from half chances so it evens out. All strikers miss chances. No Watford player is more likely to score a vital goal in the last 10 minutes than Andre.