I love a good stat but this has to be the most useless one available. How on earth is it calculated? We miss so many really good chances I just can't see how we could be 20 points above what we should be...
we pretty much beat every team we play on xg so how are we -20 points? name the games?
it's made up to suit.
what is one man's clear chance is another man's poor chance.
another way to look at it is city have a terrible attack who waste chances. given they have 100 goals or something so far that clear is rubbish.
Expected goals is a pile of steaming horse chit
just flicked threw them only 4 games the opposition had a better xg those were SH away Chelsea Away City Home and Wolves away.
comment by Inbefore - still yet to realise my ambition of being pranks fluffier (U20589)
posted 2 minutes ago
we pretty much beat every team we play on xg so how are we -20 points? name the games?
----------------------------------------------------------------------
Wolves, Man City, Chelsea, Southampton are the only ones I can see. They did predict us to score more goals than Man United in the draw.
I'm not sure how close the expected goals need to be for them to think a draw is expected, maybe that explains the other 8 points.
comment by Loco Liverpool (U18018)
posted 3 minutes ago
comment by Inbefore - still yet to realise my ambition of being pranks fluffier(U20589)
posted 2 minutes ago
we pretty much beat every team we play on xg so how are we -20 points? name the games?
----------------------------------------------------------------------
Wolves, Man City, Chelsea, Southampton are the only ones I can see. They did predict us to score more goals than Man United in the draw.
I'm not sure how close the expected goals need to be for them to think a draw is expected, maybe that explains the other 8 points.
----------------------------------------------------------------------
maybe u need to have double the xg to opponent to get the 3 points, so they have us at like 12 draws?
they basically turned us into arsenal
Of all the stats this one has to be the most useless.
A modern fad based on the shock effect to gain popularity, because the results of xG are often shocking, but that's because its a sheet stat which leads to shocking results. And the reason for that is because its a sheet stat that leads to shocking results.
Yeah that understat website is nonsense.
There was a game the other week when we scored three, hit the post, missed a sitter and got 1.8.
Don't really agree with the XG thing, but it's not that shocking.
Whenever I watch you it seems like you can turn it on at any point, so even if you went a goal down you'd turn it on earlier and still win, but you just never seem to go that goal down when other teams are on top.
Good job that football isn’t played on a spreadsheet.
comment by moreinjuredthanowen (U9641)
posted 33 minutes ago
it's made up to suit.
what is one man's clear chance is another man's poor chance.
another way to look at it is city have a terrible attack who waste chances. given they have 100 goals or something so far that clear is rubbish.
----------------------------------------------------------------------
No, both those statements can be correct. Just look how many Sterling alone has missed the last couple of months!
the stat is misleading as any game alisson plays the xg against should be zero.
it doesn't matter how many shots a team gets.
comment by meltonblue (U10617)
posted 20 minutes ago
comment by moreinjuredthanowen (U9641)
posted 33 minutes ago
it's made up to suit.
what is one man's clear chance is another man's poor chance.
another way to look at it is city have a terrible attack who waste chances. given they have 100 goals or something so far that clear is rubbish.
----------------------------------------------------------------------
No, both those statements can be correct. Just look how many Sterling alone has missed the last couple of months!
----------------------------------------------------------------------
yeah but hes one of life's special people who seem to get endless second chances. city create so many open goals and he contrived to miss an absolute shedload but retains his place.
when the guard changes at city with aguero, Silva and fernandinho going it will be interesting to see who stays with kdb.
my view is b.silva, kdb, rodri will be midfield but the front 3 could complete change. the back line is likely to keep laporte and cancel so 2 more there.
that'll be the 5 or 6 players imo. sterling to be benched eventually.
I don't think you're far off.
Where's Hafi with his wisdom?
Oh yeah, I filtered the sad cuuunt months ago
It's a bollox stat in pretty much every context bar one imo, assessing management performance.
City are a great example this season, Klopps final season at Dortmund a better one. They were in relegation form pretty much but completely turned it around. Strip the stats back and the team were more or less doing the same thing but football is about as lower a scoring game that exists so key moments mean more. It's still the same management.
City are really not doing much wrong or differently than thry have for the previous two seasons from what I have seen. In fact at some points this season they have been even better imo.
Football eh.
It's not dodgy, but inherent to prediction modelling. No model predicts 100% of the results (or if so it's heavily overfitted to the training dataset).
So Liverpool and Newcastle seems to be quite some outliers that don't suit the general prediction power of the model.
I can't find an overview of the model and the parameters, but if it's based on shot quality I can guess some of the parameters. Like position shot from, shot speed, how many defenders behind the ball, maybe strong or weak foot, etc.
The model probably doesn't even look at / determine clear chance or not. It probably only sees that a shot taken from in the 16 3 meters off centre with the right foot, medium speed, and several other conditions has a 45% chance of a goal. And then add up all the shots for the expected goals for/against.
You could for example have an outlier like Robben who might score 80% of his shots from his weak foot from a certain position while 20% is normal, because he only uses his weak foot when he can tap in. At the same time shots from the right area outside of the 16 only result in a predicted goal 10% of the time, but if you are a defender playing against Robben, you should bet if you let him shoot it will result in a goal 90% of the time
And while Robben is one of a kind, even that can be included to a certain extend in the model. If a player plays on the wing with his 'wrong foot' for crossing, expect the shots taken in certain area's of a higher quality than average, because most players are not inside forwards.
You could have saved a lot of typing by simply saying "it's bollox"
comment by Keep It Greasy: Reigning English, European & World Champions (U1396)
posted 39 minutes ago
Where's Hafi with his wisdom?
Oh yeah, I filtered the sad cuuunt months ago
----------------------------------------------------------------------
he must be going crazy. finally an article about the only thing he enjoys about football.
its most useless stat.
comment by Keep It Greasy: Reigning English, European & World Champions (U1396)
posted 2 minutes ago
You could have saved a lot of typing by simply saying "it's bollox"
----------------------------------------------------------------------
It's a bollox stat only for people who don't know how to use them, which is a majority of the population.
If you look at Jamie Vardy, you can see he scores an above average amount of goals from the middle distance inside the box, where expected goals is only 0.3 or even lower than <0.1. That's a very valuable piece of information for a scout or player to have.
At the same time, Mo Salah missed a surprising amount of shots straight in front of the goal, which is the sweet spot for scoring goals. He might have a weak spot in his shooting technique or it might be a natural result of being an inside forward (getting doubled up/going into the cul de sac).
Make no mistake, you want to have those outliers as you can improve or do further research on them.
Sign in if you want to comment
Expected Goals stat
Page 1 of 2
posted on 3/2/20
I blame VAR
posted on 3/2/20
I love a good stat but this has to be the most useless one available. How on earth is it calculated? We miss so many really good chances I just can't see how we could be 20 points above what we should be...
posted on 3/2/20
we pretty much beat every team we play on xg so how are we -20 points? name the games?
posted on 3/2/20
it's made up to suit.
what is one man's clear chance is another man's poor chance.
another way to look at it is city have a terrible attack who waste chances. given they have 100 goals or something so far that clear is rubbish.
posted on 3/2/20
Expected goals is a pile of steaming horse chit
posted on 3/2/20
just flicked threw them only 4 games the opposition had a better xg those were SH away Chelsea Away City Home and Wolves away.
posted on 3/2/20
comment by Inbefore - still yet to realise my ambition of being pranks fluffier (U20589)
posted 2 minutes ago
we pretty much beat every team we play on xg so how are we -20 points? name the games?
----------------------------------------------------------------------
Wolves, Man City, Chelsea, Southampton are the only ones I can see. They did predict us to score more goals than Man United in the draw.
I'm not sure how close the expected goals need to be for them to think a draw is expected, maybe that explains the other 8 points.
posted on 3/2/20
through*
posted on 3/2/20
comment by Loco Liverpool (U18018)
posted 3 minutes ago
comment by Inbefore - still yet to realise my ambition of being pranks fluffier(U20589)
posted 2 minutes ago
we pretty much beat every team we play on xg so how are we -20 points? name the games?
----------------------------------------------------------------------
Wolves, Man City, Chelsea, Southampton are the only ones I can see. They did predict us to score more goals than Man United in the draw.
I'm not sure how close the expected goals need to be for them to think a draw is expected, maybe that explains the other 8 points.
----------------------------------------------------------------------
maybe u need to have double the xg to opponent to get the 3 points, so they have us at like 12 draws?
posted on 3/2/20
they basically turned us into arsenal
posted on 3/2/20
Of all the stats this one has to be the most useless.
A modern fad based on the shock effect to gain popularity, because the results of xG are often shocking, but that's because its a sheet stat which leads to shocking results. And the reason for that is because its a sheet stat that leads to shocking results.
posted on 3/2/20
Yeah that understat website is nonsense.
There was a game the other week when we scored three, hit the post, missed a sitter and got 1.8.
posted on 3/2/20
Don't really agree with the XG thing, but it's not that shocking.
Whenever I watch you it seems like you can turn it on at any point, so even if you went a goal down you'd turn it on earlier and still win, but you just never seem to go that goal down when other teams are on top.
posted on 3/2/20
Good job that football isn’t played on a spreadsheet.
posted on 3/2/20
comment by moreinjuredthanowen (U9641)
posted 33 minutes ago
it's made up to suit.
what is one man's clear chance is another man's poor chance.
another way to look at it is city have a terrible attack who waste chances. given they have 100 goals or something so far that clear is rubbish.
----------------------------------------------------------------------
No, both those statements can be correct. Just look how many Sterling alone has missed the last couple of months!
posted on 3/2/20
Very interesting
posted on 3/2/20
the stat is misleading as any game alisson plays the xg against should be zero.
it doesn't matter how many shots a team gets.
posted on 3/2/20
comment by meltonblue (U10617)
posted 20 minutes ago
comment by moreinjuredthanowen (U9641)
posted 33 minutes ago
it's made up to suit.
what is one man's clear chance is another man's poor chance.
another way to look at it is city have a terrible attack who waste chances. given they have 100 goals or something so far that clear is rubbish.
----------------------------------------------------------------------
No, both those statements can be correct. Just look how many Sterling alone has missed the last couple of months!
----------------------------------------------------------------------
yeah but hes one of life's special people who seem to get endless second chances. city create so many open goals and he contrived to miss an absolute shedload but retains his place.
when the guard changes at city with aguero, Silva and fernandinho going it will be interesting to see who stays with kdb.
my view is b.silva, kdb, rodri will be midfield but the front 3 could complete change. the back line is likely to keep laporte and cancel so 2 more there.
that'll be the 5 or 6 players imo. sterling to be benched eventually.
posted on 3/2/20
I don't think you're far off.
posted on 3/2/20
Where's Hafi with his wisdom?
Oh yeah, I filtered the sad cuuunt months ago
posted on 3/2/20
It's a bollox stat in pretty much every context bar one imo, assessing management performance.
City are a great example this season, Klopps final season at Dortmund a better one. They were in relegation form pretty much but completely turned it around. Strip the stats back and the team were more or less doing the same thing but football is about as lower a scoring game that exists so key moments mean more. It's still the same management.
City are really not doing much wrong or differently than thry have for the previous two seasons from what I have seen. In fact at some points this season they have been even better imo.
Football eh.
posted on 3/2/20
It's not dodgy, but inherent to prediction modelling. No model predicts 100% of the results (or if so it's heavily overfitted to the training dataset).
So Liverpool and Newcastle seems to be quite some outliers that don't suit the general prediction power of the model.
I can't find an overview of the model and the parameters, but if it's based on shot quality I can guess some of the parameters. Like position shot from, shot speed, how many defenders behind the ball, maybe strong or weak foot, etc.
The model probably doesn't even look at / determine clear chance or not. It probably only sees that a shot taken from in the 16 3 meters off centre with the right foot, medium speed, and several other conditions has a 45% chance of a goal. And then add up all the shots for the expected goals for/against.
You could for example have an outlier like Robben who might score 80% of his shots from his weak foot from a certain position while 20% is normal, because he only uses his weak foot when he can tap in. At the same time shots from the right area outside of the 16 only result in a predicted goal 10% of the time, but if you are a defender playing against Robben, you should bet if you let him shoot it will result in a goal 90% of the time
And while Robben is one of a kind, even that can be included to a certain extend in the model. If a player plays on the wing with his 'wrong foot' for crossing, expect the shots taken in certain area's of a higher quality than average, because most players are not inside forwards.
posted on 3/2/20
You could have saved a lot of typing by simply saying "it's bollox"
posted on 3/2/20
comment by Keep It Greasy: Reigning English, European & World Champions (U1396)
posted 39 minutes ago
Where's Hafi with his wisdom?
Oh yeah, I filtered the sad cuuunt months ago
----------------------------------------------------------------------
he must be going crazy. finally an article about the only thing he enjoys about football.
its most useless stat.
posted on 3/2/20
comment by Keep It Greasy: Reigning English, European & World Champions (U1396)
posted 2 minutes ago
You could have saved a lot of typing by simply saying "it's bollox"
----------------------------------------------------------------------
It's a bollox stat only for people who don't know how to use them, which is a majority of the population.
If you look at Jamie Vardy, you can see he scores an above average amount of goals from the middle distance inside the box, where expected goals is only 0.3 or even lower than <0.1. That's a very valuable piece of information for a scout or player to have.
At the same time, Mo Salah missed a surprising amount of shots straight in front of the goal, which is the sweet spot for scoring goals. He might have a weak spot in his shooting technique or it might be a natural result of being an inside forward (getting doubled up/going into the cul de sac).
Make no mistake, you want to have those outliers as you can improve or do further research on them.
Page 1 of 2