According to our xG this season, we should have scored TWENTY more goals. No other club is even in double figures for negative xG let alone near 20. Is this down to management, terrible finishing from the players or shear bad luck? Or a mixture of all 3?
We have been the nearly men for many games this season. Played quite well and created chances in games, but then seemed to lose a goal from nowhere and lose the game.
We should be 9th in the table on 45 points according to xG.
Use these statistics however you wish, and I know that they do not tell the whole story. We still look terrible to the eye, and we certainly do not dominate or control games.
However, this is about measuring Dyche against something, and the statistics tell me that he hasn't done half bad.
Dyche - HE'S GONE
-
StirlingBlue
- Posts: 1741
- Karma: 895
Re: Dyche
One of the things that xG doesn’t take into account is something very specific to how Dyche has us playing, how prepared the player is for the chance.Gary1878 wrote:According to our xG this season, we should have scored TWENTY more goals. No other club is even in double figures for negative xG let alone near 20. Is this down to management, terrible finishing from the players or shear bad luck? Or a mixture of all 3?
We have been the nearly men for many games this season. Played quite well and created chances in games, but then seemed to lose a goal from nowhere and lose the game.
We should be 9th in the table on 45 points according to xG.
Use these statistics however you wish, and I know that they do not tell the whole story. We still look terrible to the eye, and we certainly do not dominate or control games.
However, this is about measuring Dyche against something, and the statistics tell me that he hasn't done half bad.
Under an xG model a cutback for a shot on the edge of the 6 yard box is where an equal amount as one of our ping pong in the box chances - our chances are super reactive and not planned.
Re: Dyche
Probably wrong but I think
- quantity of chance v quality of chance
- open play Vs set piece / immediately following set piece
- how much of chance creation and finishing relies on the individual making a decision rather than patterned and practiced play
- game state
- actual flow of the game. We don't really attack on the break isolating defenders and stuff so we, we do more kind of...Godfrey swinging a size 11 at it in a crowded box and booting it into the press section
All comes into, plus probably lots more things.
That said, we have certainly missed some proper sitters. I've seen em.
- quantity of chance v quality of chance
- open play Vs set piece / immediately following set piece
- how much of chance creation and finishing relies on the individual making a decision rather than patterned and practiced play
- game state
- actual flow of the game. We don't really attack on the break isolating defenders and stuff so we, we do more kind of...Godfrey swinging a size 11 at it in a crowded box and booting it into the press section
All comes into, plus probably lots more things.
That said, we have certainly missed some proper sitters. I've seen em.
Re: Dyche
Definitely all 3.Gary1878 wrote: ↑Wed Apr 10, 2024 9:43 am According to our xG this season, we should have scored TWENTY more goals. No other club is even in double figures for negative xG let alone near 20. Is this down to management, terrible finishing from the players or shear bad luck? Or a mixture of all 3?
We have been the nearly men for many games this season. Played quite well and created chances in games, but then seemed to lose a goal from nowhere and lose the game.
We should be 9th in the table on 45 points according to xG.
Use these statistics however you wish, and I know that they do not tell the whole story. We still look terrible to the eye, and we certainly do not dominate or control games.
However, this is about measuring Dyche against something, and the statistics tell me that he hasn't done half bad.
We wrote off the first 4 games of the season playing Maupay up front who missed several gilt edged chances, we've had other players miss a host of very presentable chances, and we've seemingly done everything possible to not score with rebounds and scrambles within 8 yards of goal.
Second half of the season though it feels a lot to do with setup and low quality chance creation. I don't feel we've missed loads of great opportunities like we did earlier in the season.
-
777Kidnappings
- Posts: 3036
- Karma: 1699
Re: Dyche
Surely lots of low quality chances and a few great chances if the XG is the same then the goals should average out too but for luck and bad finishing
1000 5% chances should average 50 and should 100 50% chances
It's luck and poor finishing why we are so behind on xg. It's all it can be. There's a separate argument of how useful xg is but you can't blame a manager for being way behind the clubs xg
1000 5% chances should average 50 and should 100 50% chances
It's luck and poor finishing why we are so behind on xg. It's all it can be. There's a separate argument of how useful xg is but you can't blame a manager for being way behind the clubs xg
-
Kerryblueboy
- Posts: 2520
- Karma: 689
Re: Dyche
I can remember dcl against villa and Beto against Fulham where were all those other great chances we missed don’t let xg mask over shite football
Re: Dyche
Even if 10x are on the manager, that’s still 10x on the players, which is ridiculous, damning and unacceptable.
Are we really surprised that McNeil, Harrison, Doucoure, Garner etc aren’t making better chances? Apart from overloading in attack - which this manage (rightly), isn’t going to do, what’s the alternative?
Are we really surprised that McNeil, Harrison, Doucoure, Garner etc aren’t making better chances? Apart from overloading in attack - which this manage (rightly), isn’t going to do, what’s the alternative?
-
Kerryblueboy
- Posts: 2520
- Karma: 689
Re: Dyche
Hard to make better chances when the ball is flying over their heads all the timeGranite wrote: ↑Wed Apr 10, 2024 7:02 pm Even if 10x are on the manager, that’s still 10x on the players, which is ridiculous, damning and unacceptable.
Are we really surprised that McNeil, Harrison, Doucoure, Garner etc aren’t making better chances? Apart from overloading in attack - which this manage (rightly), isn’t going to do, what’s the alternative?
Re: Dyche
You can absolutely say that the way a manager has a team playing is having an impact on their xg performance.777Kidnappings wrote: ↑Wed Apr 10, 2024 1:34 pm Surely lots of low quality chances and a few great chances if the XG is the same then the goals should average out too but for luck and bad finishing
1000 5% chances should average 50 and should 100 50% chances
It's luck and poor finishing why we are so behind on xg. It's all it can be. There's a separate argument of how useful xg is but you can't blame a manager for being way behind the clubs xg
Dyche for instance was renowned for allowing lots and lots of very high quality shots against his Burnley team, but defending them with two cbs on the line and the goalie in the middle, so a Hugh xg chance became a low xg chance post shot, because the goal was so small and the box was so crowded.
Re: Dyche
I'm not sure the huge significance that @Gary1878 mentioned is properly being understood here.
If we were just an 'average' finishing side, we would have scored 52 goals rather than 32. That's 62.5% more goals than what we've currently scored - equates to MANY more points. Imagine if we were a 'good' finishing side and exceeded our xG, we'd likely be in Europe if ALL ELSE remained the same yet our finishing improved.
And it's more valid now because it's xG calculated over most of a season, and it's growing, it's far from a blip.
As @777Kidnappings also states, 5 shots with 10% chance each of scoring is the same as 1 shot with 50% chance. The quality vs quantity argument from a statistical point of view is largely irrelevant as its factored into the calculations. Things like playing style etc are not in the xG argument.
I'm more than happy to point out the problems with assumptions in the xG model, when used over smaller sample sizes, but this is undeniably a huge shortfall in the 'Goals For 'column that shot takers carry the bulk of the responsibility for.
Edit: In an alternate universe, if our matches unfolded in exactly the same way, every pass, every shot, every tackle, except that our finishing had been 38% better (rather than 38% worse) than what is deemed average, then we'd certainly be in Champions League contention. 71 goals for, and 42 against, only the three title challengers would be in front of us. That is the magnitude of how bad our finishing has been this season.
If we were just an 'average' finishing side, we would have scored 52 goals rather than 32. That's 62.5% more goals than what we've currently scored - equates to MANY more points. Imagine if we were a 'good' finishing side and exceeded our xG, we'd likely be in Europe if ALL ELSE remained the same yet our finishing improved.
And it's more valid now because it's xG calculated over most of a season, and it's growing, it's far from a blip.
As @777Kidnappings also states, 5 shots with 10% chance each of scoring is the same as 1 shot with 50% chance. The quality vs quantity argument from a statistical point of view is largely irrelevant as its factored into the calculations. Things like playing style etc are not in the xG argument.
I'm more than happy to point out the problems with assumptions in the xG model, when used over smaller sample sizes, but this is undeniably a huge shortfall in the 'Goals For 'column that shot takers carry the bulk of the responsibility for.
Edit: In an alternate universe, if our matches unfolded in exactly the same way, every pass, every shot, every tackle, except that our finishing had been 38% better (rather than 38% worse) than what is deemed average, then we'd certainly be in Champions League contention. 71 goals for, and 42 against, only the three title challengers would be in front of us. That is the magnitude of how bad our finishing has been this season.
Last edited by Cods on Wed Apr 10, 2024 11:08 pm, edited 1 time in total.
Re: Dyche
The point is, that by its definition, many many low chance shots (lots of defenders defending a goal) is still used to calculate the xG, it's factored into the xG figure. Granted there are different assumption included in different models, but none can be used to explain the very large difference we are seeing. We're underscoring at a rate of nearly 40%.brap2 wrote: ↑Wed Apr 10, 2024 8:29 pm You can absolutely say that the way a manager has a team playing is having an impact on their xg performance.
Dyche for instance was renowned for allowing lots and lots of very high quality shots against his Burnley team, but defending them with two cbs on the line and the goalie in the middle, so a Hugh xg chance became a low xg chance post shot, because the goal was so small and the box was so crowded.
If a model showed in practice many 40% variances the model itself would have to be looked at. And they're not. No other team is anywhere near the huge underperformance we show. The model's not broken, we are.
Re: Dyche
The way we are set up to play can influence our TOTAL xG (52), no argument there. But it can't really explain our huge VARIANCE TO xG. That is, why we're such a huge outlier with ACTUAL goals (only 32), when we should have scored 62% more if we were only half decent at finishing.
Re: Dyche
The coach, how he's coaching, what he's coaching absolutely can impact that variance through the ways discussed above.
Are those weaknesses in xg, are they wrinkles that require more nuanced analytics beyond free shot quality / quantity models, yes probably, but there's a reason the last decade of football has moved towards systems that generate higher % shot quality in repeatable patterns of play, when they could all have just been twatting 100 40 yard shots a game to generate the same overall xg total.
Some managers are known to buck xg trends regularly. Dyche is literally one of them as mentioned above, in which is Burnley team would allow high and high quality xg but force a variance in goals against through what and how they did things on this pitch.
Lucien Favre is another who through what his teams do on the pitch regularly over different seasons in different leagues has bucked xg tables and impacted variance.
That's not to say we haven't had bad finishing / bad luck etc. to go along with it. Because we definitely have had both.
Are those weaknesses in xg, are they wrinkles that require more nuanced analytics beyond free shot quality / quantity models, yes probably, but there's a reason the last decade of football has moved towards systems that generate higher % shot quality in repeatable patterns of play, when they could all have just been twatting 100 40 yard shots a game to generate the same overall xg total.
Some managers are known to buck xg trends regularly. Dyche is literally one of them as mentioned above, in which is Burnley team would allow high and high quality xg but force a variance in goals against through what and how they did things on this pitch.
Lucien Favre is another who through what his teams do on the pitch regularly over different seasons in different leagues has bucked xg tables and impacted variance.
That's not to say we haven't had bad finishing / bad luck etc. to go along with it. Because we definitely have had both.