Expected Goals. There and back again

[Blue Hefner]

I quite like Xg. It’s a more useful measure that shots-on-target
Particularly of those who haven’t watched the match

 

[madloaf]

The xPR (expected politeness of response) of this thread has dropped significantly in recent comments. This reflects a wider drop in xPR across the whole forum during the summer.

 

[SWP's back]

I know exactly what xG is about. It uses subjective data and presents it as hard objective scientific fact. It's about as credible as Scientology and anyone who rates it is as thick as a brick.

Again you’ve absolutely no idea what you’re talking about. You’re the fucking clown posting one CL result as a gotcha to show xG is useless when the sum total of zero clubs have ever thought to use the xG of one match to prove anything.

You could do worse than read the real AI post above showing where it fits in every top club’s analytics department, especially Peps.

 

[SWP's back]

xG is just made up you thick prick.

Oh the irony. The senile old dinosaur who doesn’t understand a new part of football analysis calling others thick.

You’ve embarrassed yourself in this thread and even made someone that agrees with you on xG tell you that you don’t have to be a **** about it.

 

[Donmcfc]

Strange that half of Pep's coaching staff have left the club during the close season.
Any statistical system that simply dismisses results that don't fit the narrative is less than useless.
Why have City gone backwards in the last 12 months since xG is the dog's bollocks?
Have Liverpool and Arsenal got better statistical analysts than us?
And the Earth isn't flat by the way.

 

[Donmcfc]

Oh the irony. The senile old dinosaur who doesn’t understand a new part of football analysis calling others thick.

You’ve embarrassed yourself in this thread and even made someone that agrees with you on xG tell you that you don’t have to be a **** about it.

 

[Georgian Maestro]

I would guess that a high proportion of the XG is bollocks brigade were in the bottom set at Maths in school.

 

[Donmcfc]

Not embarrassed in the slightest. Read the following extract from Sports Gazette UK:

"According to their xG, Huddersfield Town should have finished in 16th, compared to the third-place finish they achieved.

The data shows that they managed to win 10 more games than their expected goals suggest and finished 29 points better off too.

Nottingham Forest would have finished 10 points worse off based on xG, which would have seen them finish outside the play-offs in ninth.

The biggest reason for this is that Forest only conceded 40 goals, even though their expected goals against was 54."

Now that's embarrassing for all xG advocates. Let's see one of you explain that away.

 

[Milesvk]

Im expecting city to miss a load of chances before we go 1-0 down. That stat is certain as night and day.

 

[bluesmith]

Stop feeding the troll, he's the type of person who has his name and DOB on the internet, he'd tell.you his first pets name if you asked him.

 

[Donmcfc]

I would guess that a high proportion of the XG is bollocks brigade were in the bottom set at Maths in school.

But have you got any statistics to back that up or are you just another xG blagger?
Just make up your facts to back up your argument.
This is the modern Trumpian style of debate.
The alternative is that you have to be as thick as mince to support xG statistics.

 

[Donmcfc]

Stop feeding the troll, he's the type of person who has his name and DOB on the internet, he'd tell.you his first pets name if you asked him.

Troll? For having a different opinion than you? Pathetic.
And my first pet's name was Kerry if that's of any use to you.

 

[bluesmith]

Troll? For having a different opinion than you? Pathetic.
And my first pet's name was Kerry if that's of any use to you.

Well with your dob of 4th april 1951 and your surname it probably could be.

Difference of opinions are fine, stupidity isn't.

 

[Donmcfc]

Well with you dob of 4th april 1951 and your surname it probably could be.

Difference of opinions are fine, stupidity isn't.

Well with you dob of 4th april 1951 and your surname it probably could be.

Difference of opinions are fine, stupidity isn't.

Interesting that not one of you xG advocates have responded to the Huddersfield and Nottingham Forest discrepancies posted above? Why?
Because you can't explain it so instead you resort to personal attacks. Am I bothered? Stupid weirdo dinosaur? Me?

One thing I would never do is look up the personal details of a fellow Bluemoon member to make some kind of point. Now that is fucking weird.

 

[bluesmith]

Interesting that not one of you xG advocates have responded to the Huddersfield and Nottingham Forest discrepancies posted above? Why?
Because you can't explain it so instead you resort to personal attacks. Am I bothered? Stupid weirdo dinosaur? Me?

One thing I would never do is look up the personal details of a fellow Bluemoon member to make some kind of point. Now that is fucking weird.

I've not looked them up there in your bio, you are genuinely stupid.

As for the Forest and Huddersfield points the fact you think that proves your argument again proves you dont understand what xG is there for.

Imagine thinking you are right and every top club in the world is wrong.

 

[youngbob]

I don't mind XG. As others have said, it's often a decent indicator of how a game has ebbed and flowed and which team has really had the better chances.

Team A 1-0 Team B
Shots: 15, 8
On target: 7, 4

From that, you'd determine that Team A were pretty dominant while Team B didn't have a sniff. But all 15 of Team A's shots could be from outside the box and one of them just took a massive deflection and went in, while Team B could have missed half a dozen sitters.

If the XG stats from the same game were 1.1 vs 2.3, this would indicate that Team B created the better chances - their striker just didn't put them in the net. Goals win you games, yes, but XG just gives you a bit of an idea of how patterns of play went and which team had more of the overall share.

It should be used in the same way possession is, or shots on target, or any detailed stat in football really that isn't just goals scored. Not as hard proof of anything, but as a general indicator of which team performed best outside the 18 yard box. It gives managers a clearer picture of what really needs working on.

For example, if City were to lose a game 0-2, you'd look at the scoreline and say we need to completely change our attacking and defending set-up. But if the XG for City is 3.2 and the XG for our opponents is 0.4, you'd realise our opponents scored a couple of freak goals and that we just need to tighten up our finishing.

 

[SWP's back]

Not embarrassed in the slightest. Read the following extract from Sports Gazette UK:

"According to their xG, Huddersfield Town should have finished in 16th, compared to the third-place finish they achieved.

The data shows that they managed to win 10 more games than their expected goals suggest and finished 29 points better off too.

Nottingham Forest would have finished 10 points worse off based on xG, which would have seen them finish outside the play-offs in ninth.

The biggest reason for this is that Forest only conceded 40 goals, even though their expected goals against was 54."

Now that's embarrassing for all xG advocates. Let's see one of you explain that away.

Oh my fucking god. You still don’t get it. You think the above is proof of xG being shit when it is in fact the sort of thing analytics teams have a wet dream over as it opens up multiple avenues to work out how they’ve outperformed (usually a shit hot striker).

Posts like yours above back up every single poster that says you don’t understand the subject. At this point I actually think you’re trolling as no one’s this thick.

 

[Vialli98]

Not embarrassed in the slightest. Read the following extract from Sports Gazette UK:

"According to their xG, Huddersfield Town should have finished in 16th, compared to the third-place finish they achieved.

The data shows that they managed to win 10 more games than their expected goals suggest and finished 29 points better off too.

Nottingham Forest would have finished 10 points worse off based on xG, which would have seen them finish outside the play-offs in ninth.

The biggest reason for this is that Forest only conceded 40 goals, even though their expected goals against was 54."

Now that's embarrassing for all xG advocates. Let's see one of you explain that away.

I don't know what you think this shows? xG is very obviously different from actual results. It's just a statistic that allows you to reasonable know how a team performed, which is why xG for a single game is a pretty useless stat but over the course of 5-10 games and more, you can use it to spot a trend.

A player or a team can outperform or underperform their underlying stats for a whole season. It happens. It's not the gotcha that you think it is. Sometimes teams just have attackers that are extremely poor finishers but have great movement so they keep getting into great positions but they keep fluffing their lines. A team with those kind of attackers will routinely have an xG much higher than the actual goals scored. That doesn't mean xG is a bullshit stat. It just very clearly means this team is good a creating chances but shit at finishing them and xG clearly explains that in easy to understand metrics.

Forest only conceding 40 goals when their xGa was 54 for the season could mean one or more of several things. Maybe they had an extremely good keeper (they did, Sels is routinely mentioned in TOTS conversations), luck (opposing teams continually fluffing their lines against them, Chelsea for example, dominated them but we went 1-0 down via a set piece and had rescue the game late on, finished 1-1, Sels was incredible but we should have battered them).

Being anti xG is one of the most bizarre things in football. It's just a metric. You can absolutely have a problem with people using it wrong (of which there are many) but the stat itself is just a stat. Whether you find the information you can gain from it interesting is subjective but the stat itself is still just a stat. Raging against it is utterly bizarre.

 

[twosips]

Genuinely quite funny reading loads of people arguing against xG not realising that once again it's just their old school language put into metrics. You. Are. All. Saying. The. Same. Thing.

Just data teams look at it all methodically and come up with stats as opposed to saying 'we had loads of chances and they had a few and we should have scored a few and they were lucky to get one', because that's their fucking job - to put things into metrics that are easily referenced. Not long sentences and a gut feeling. I can't believe people are arguing against probabilities that have absolutely zero bias. Literally no one thinks it replaces the eye test or tactics, if you do you simply don't get it.

They've analysed millions of chances. The numbers might show that you're much more likely to score from position A than position B, based off the shots success rates of millions and millions of chances...so guess what? they encourage players to shoot at position A. That obviously doesn't mean NEVER SHOOT AT POSITION B or YOU CAN'T SCORE FROM POSITION B. Obviously not ffs. Just that usually Position A means youre more likely to score, so maybe consider going there instead. It's just teams looking for marginal gains by analysing data. It's common sense.

This is exactly what happened with Sterling in that training video I saw. Instead of trusting gut instinct and going 'get in the area to score', they did the research and found the exact positions of where to be based off where the cross was coming from. They worked at it intensely using positional play backed up by numerical research. Look at videos of Sterling's movement from the 17/18 season before he scored loads of backpost tap ins etc - it's incredibly specific and considered. He works hard to get in preplanned positions and he scores fucking loads as a result.

This is down to 1) great coaching and 2) data analytics...both working together in perfect harmony. That's modern football. If you're arguing against it you simply don't get it.

 

[Saddleworth2]

Genuinely quite funny reading loads of people arguing against xG not realising that once again it's just their old school language put into metrics. You. Are. All. Saying. The. Same. Thing.

Just data teams look at it all methodically and come up with stats as opposed to saying 'we had loads of chances and they had a few and we should have scored a few and they were lucky to get one', because that's their fucking job - to put things into metrics that are easily referenced. Not long sentences and a gut feeling. I can't believe people are arguing against probabilities that have absolutely zero bias. Literally no one thinks it replaces the eye test or tactics, if you do you simply don't get it.

They've analysed millions of chances. The numbers might show that you're much more likely to score from position A than position B, based off the shots success rates of millions and millions of chances...so guess what? they encourage players to shoot at position A. That obviously doesn't mean NEVER SHOOT AT POSITION B or YOU CAN'T SCORE FROM POSITION B. Obviously not ffs. Just that usually Position A means youre more likely to score, so maybe consider going there instead. It's just teams looking for marginal gains by analysing data. It's common sense.

This is exactly what happened with Sterling in that training video I saw. Instead of trusting gut instinct and going 'get in the area to score', they did the research and found the exact positions of where to be based off where the cross was coming from. They worked at it intensely using positional play backed up by numerical research. Look at videos of Sterling's movement from the 17/18 season before he scored loads of backpost tap ins etc - it's incredibly specific and considered. He works hard to get in preplanned positions and he scores fucking loads as a result.

This is down to 1) great coaching and 2) data analytics...both working together in perfect harmony. That's modern football. If you're arguing against it you simply don't get it.

great post mate.