Based on xg and xga

What is xg and xga ?

That will be the point we were robbed at Leicester and two at Wolves

Absolutely, however, the issue is XG + XGA + VAR + big clubbiastakathepissa

Can someone explain what this thread is about? What exactly are XG and XGA? Are they comets that are going to collide with Earth at some point.

houseboy wrote:Can someone explain what this thread is about? What exactly are XG and XGA? Are they comets that are going to collide with Earth at some point.

Xpected goals and Xpected goals against

And Everton 5th... think it's safe to say it's a load of old tosh.

MACCA wrote:And Everton 5th... think it's safe to say it's a load of old tosh.

What do you mean?

Down_Rover wrote:Xpected goals and Xpected goals against

So how do they work that one out then and place us 6th? Surely no-one this side of Venus expects us to score many goals? It's not what we do, even when playing well. Our defence is actually pretty good normally but if the idea is to balance them out then surely the table as it is speaks for itself. As the saying goes 'the table doesn't lie'. I have always believed that, contrary to what fans and managers say, you are always pretty much where you deserve to be.

Newty wrote:What do you mean?

It says we should be currently 6th...

Everton should be 5th based on whatever system they're using.

Basically a load of rubbish

Quality of chances created. It’s not too hard to understand...we destroyed Chelsea on chances created and missed them. They then scored with a few half chances

MACCA wrote:It says we should be currently 6th...

Everton should be 5th based on whatever system they're using.

Basically a load of rubbish

That isn't what it says at all.

It's a very good statistical model, designed to provide an indication of the volume and quality of chances created and conceded by teams. It's definitely not perfect, but it's much better than just looking at the league table.

It's your interpretation of that model that's a load of rubbish. Do you really think Everton are the 5th worst team in the league?

I don't see why you think its a load of rubbish, its exactly what it says it is.

If a team is wasteful and miss chances you would expect them to score, then they wont be as high up the table as if they were very clinical and scored goals you would never expect them too. That's literally what the xgoals table is. Only more statistical and complex. Analysing every single chance and goal and based on historical similar chances saying whether the chance should have been scored or not.

It certainly doesn't mean we deserve to be 6th and doesn't claim to.

Our xg value for the 2 we scored against Chelsea would be very small. The issue was Ashley missing the first half and second half chance amongst others...

Hibsclaret wrote:Our xg value for the 2 we scored against Chelsea would be very small. The issue was Ashley missing the first half and second half chance amongst others...

The two headers that Barnes missed in the first half had a combined xG of 1.42, which is huge.

The two we actually did score were 0.06 xG in total.

When you have spent not far off £200m on strikers or attacking “goal scoring” midfielders like Everton have and they are missing good chances as the table suggests I would be a bit more worried than when your half a million striker like Barnes is missing these chances.

We do seem to be making a few more decent chances this year than last year / previous seasons in the premier league but it’s very unlikely that Burnley will have a 20 goal per season striker in this league or a 10 goal a season midfielder.

Hibsclaret wrote:Quality of chances created. It’s not too hard to understand...we destroyed Chelsea on chances created and missed them. They then scored with a few half chances

So that means we have crap strikers then? Is that what it's saying? And there's me thinking that Wood and Barnes were very good at converting the few chances we get.

houseboy wrote: So that means we have crap strikers then? Is that what it's saying? And there's me thinking that Wood and Barnes were very good at converting the few chances we get.

No, that isn't what it's saying. We've actually scored slightly more than "expected" based on the quality of chances. The reason we're lower in the actual table than the xG table is because we've conceded more goals than we "should" have.

Newty wrote:What do you mean?

The presented table was no doubt accompanied by declarations of " this team is over performing in this area" and " this teams is underperforming in this area". Taking it as a given that the expectations generated by whatever xG formulae they employed ( because there isn't a universally accepted formulae for xG and every stat bunny employs there own subjective variation ) were fundamentally accurate and objective.

However, when I look at the table I see that the formulae used only matched with what actually happened in a couple of cases. In the overwhelming majority of cases there is a statistically significant degree of divergence. To anybody with a reasonable grasp of mathematics and data analysis that can only mean one thing - the formulae is ******.

The formulae that is being employed in this instance is obviously failing to generate an accurate model of reality. If the formulae was operating effectively then it would be generating a far greater degree of positive correlation between what was expected to happen and what actually happened in terms of the overall outcomes.

Long Time Lurker wrote:The presented table was no doubt accompanied by declarations of " this team is over performing in this area" and " this teams is underperforming in this area". Taking it as a given that the expectations generated by whatever xG formulae they employed ( because there isn't a universally accepted formulae for xG and every stat bunny employs there own subjective variation ) were fundamentally accurate and objective.

However, when I look at the table I see that the formulae used only matched with what actually happened in a couple of cases. In the overwhelming majority of cases there is a statistically significant degree of divergence. To anybody with a reasonable grasp of mathematics and data analysis that can only mean one thing - the formulae is ******.

The formulae that is being employed in this instance is obviously failing to generate an accurate model of reality. If the formulae was operating effectively then it would be generating a far greater degree of positive correlation between what was expected to happen and what actually happened in terms of the overall outcomes.

It wasn't presented with any such declaration, so most of the rest of your argument is a strawman.

You wouldn't expect a high degree of correlation with a sample size as small as ten games. However the models have been tested over bigger sample sizes and proven to be a better indicator of relative team strength than the league table.

LoveCurryPies wrote:What is xg and xga ?

Could be the Xhaka, Granit effect!

Tall Paul wrote:That isn't what it says at all.

It's a very good statistical model, designed to provide an indication of the volume and quality of chances created and conceded by teams. It's definitely not perfect, but it's much better than just looking at the league table.

It's your interpretation of that model that's a load of rubbish. Do you really think Everton are the 5th worst team in the league?

Nope, probably not the 5th worst
Probably not the 5th best either

But I'll tell you exactly what position they deserve to be in, in May.

Tall Paul wrote:The two headers that Barnes missed in the first half had a combined xG of 1.42, which is huge.

The two we actually did score were 0.06 xG in total.

Which kind of puts things into perspective.

Does it look at Var decisions, keeper saves, woodwork hit, OG etc?

So the OP saying we should be 6th based on x model.

We could easily get another dozen tables up based on wages, transfer fees, international players etc etc.

Hibsclaret wrote:Quality of chances created. It’s not too hard to understand...we destroyed Chelsea on chances created and missed them. They then scored with a few half chances

Apparently it is too hard to understand.

MACCA wrote:Which kind of puts things into perspective.

Does it look at Var decisions, keeper saves, woodwork hit, OG etc?

So the OP saying we should be 6th based on x model.

We could easily get another dozen tables up based on wages, transfer fees, international players etc etc.

Are you really struggling this much to understand it ?

Didn't MOTD show our XG at 4.5 against Chelsea and theirs was around 1.5, of I dreamed that up?

TVC15 wrote:Are you really struggling this much to understand it ?

Nope I get it, it's just pointless.
Serves no purpose and means diddly squat.

They'll be as much correlation between that table and the real table as there would be if you did a wages one, a transfer spend one etc etc.

So when people say we "should be x place based on" you could do another half a dozen tables showing us in a different half a dozen places.

MACCA wrote:Nope I get it, it's just pointless.
Serves no purpose and means diddly squat.

They'll be as much correlation between that table and the real table as there would be if you did a wages one, a transfer spend one etc etc.

So when people say we "should be x place based on" you could do another half a dozen tables showing us in a different half a dozen places.

Don’t think anyone was saying that we deserved to be 6th - the “should” was just a link based to that particular table.

Personally I think this data is pretty useful - as long as you don’t use it as gospel. When you strip it down to its bare meaning it’s basically saying that the more good / easier chances to score that you make (or concede) then the more likely it is that you will score goals / win games. So there will definitely be correlation to these statistics and league positions especially over several seasons as looking at any data over a short period of time is likely to throw up more anomalies.
Of course other tables on things like transfer fees, wages etc are also available....and again it’s logical that on more occasions than not the richer clubs will finish higher than the poorer ones and of course there are always exceptions to this.....and there’s only one table that matters etc etc

Best thing is if you don’t like this kind of stuff - ignore it rather than try and argue it’s wrong...

Statistics show that very few people die over the age of 99

Im_not_Robbie_Blake wrote:Statistics show that very few people die over the age of 99

That's wrong. They all die

TVC15 wrote:Don’t think anyone was saying that we deserved to be 6th - the “should” was just a link based to that particular table.

Personally I think this data is pretty useful - as long as you don’t use it as gospel. When you strip it down to its bare meaning it’s basically saying that the more good / easier chances to score that you make (or concede) then the more likely it is that you will score goals / win games. So there will definitely be correlation to these statistics and league positions especially over several seasons as looking at any data over a short period of time is likely to throw up more anomalies.
Of course other tables on things like transfer fees, wages etc are also available....and again it’s logical that on more occasions than not the richer clubs will finish higher than the poorer ones and of course there are always exceptions to this.....and there’s only one table that matters etc etc

Best thing is if you don’t like this kind of stuff - ignore it rather than try and argue it’s wrong...

I'm not arguing anythings wrong, just it's pointless and achieves nothing.

Will give you what you crave.

Your opinion is correct as usual, I've now changed my mind.

Before these stats came in, more would be looked at in terms of shots attempted / shots on target vs goals scored and take into account normalisation.

So if a team were scoring for fun at the start of the season, with few chances, you'd expect things to normalise at some point, and for that team to drop down the table X number of places, as it was very unlikely they would keep it up.

This seems to be a more advanced method, based on many more variables. It's interesting, but I think it will only be interesting to people into their stats, Excel spreadsheets and the like. If you aren't into evaluating numbers like this, it'll either go over your head or just won't be of interest. If you're playing FPL for example, they can be useful and should be taken into account, or if you like to place the odd flutter on a weekend.

I bet Sean and his staff use these stats to help their training and all sorts. He’s certainly not going to be displeased conceding less quality chances to the likes of Chelsea. Quite the contrary as he always talks about quality of performance etc and the stats prove that we are consistently performing well this season in terms of breaking teams down to get more tilt edged chances... he can also see that although conceding 4 goals they really shouldn’t under normal circumstances score 4 with those chances in a game...all this info can probably help training for defenders, strikers and the shape of the team.

For one thing he will much prefer the xg and xga stats for the Chelsea game at home this season compared to last season. Irrespective of actual score line they show we did really well this season and poorly last season which will be pleasing for all concerned whilst highlighting more improvement is possible and necessary if we are to get anything out of games against the better sides

Down_Rover wrote:Xpected goals and Xpected goals against

Surely it should be
XGF expected goals for
XGA expected goals against

As XG expected goals could be interpreted as goals in the game for either team...

Int Pedantry" wunderful

On a DUKW Tour of Lisbon today Bloody brilli ant fun

FactualFrank wrote:Before these stats came in, more would be looked at in terms of shots attempted / shots on target vs goals scored and take into account normalisation.

So if a team were scoring for fun at the start of the season, with few chances, you'd expect things to normalise at some point, and for that team to drop down the table X number of places, as it was very unlikely they would keep it up.

This seems to be a more advanced method, based on many more variables. It's interesting, but I think it will only be interesting to people into their stats, Excel spreadsheets and the like. If you aren't into evaluating numbers like this, it'll either go over your head or just won't be of interest. If you're playing FPL for example, they can be useful and should be taken into account, or if you like to place the odd flutter on a weekend.

This is a good take on it.

I've been using xG stats to help my decision making in FPL for a number of years and I've done very well, winning a lot of minleagues and finishing consistently in the top 20k.

MACCA wrote:I'm not arguing anythings wrong, just it's pointless and achieves nothing.

Will give you what you crave.

Your opinion is correct as usual, I've now changed my mind.

Aaah bless....

TVC15 wrote:Aaah bless....

Classic TVC.

You disagree with him, you're wrong.
You agree with him you're still wrong.

Some people are just never happy are they.

Tall Paul wrote:This is a good take on it.

I've been using xG stats to help my decision making in FPL for a number of years and I've done very well, winning a lot of minleagues and finishing consistently in the top 20k.

Aahh so people who like or use these stats or play fantasy football and such like.
No wonder I see them as pointless, as I dont play games like that either for similar reasons.

But if you're doing well, that's surely a good

MACCA wrote:Classic TVC.

You disagree with him, you're wrong.
You agree with him you're still wrong.

Some people are just never happy are they.

Classic Macca
Thick as mince

Fails to understand some pretty basic stuff that a nine year old could get...then thinks they are pointless.
Then dismally fails at trying to be clever with a sarcastic post.

MACCA wrote:Aahh so people who like or use these stats or play fantasy football and such like.
No wonder I see them as pointless, as I dont play games like that either for similar reasons.

But if you're doing well, that's surely a good

What similar reasons?

And it's not just fantasy football it's useful for, they're useful for finding value in games that the bookies have mispriced. For example, we look good value for a result at Bramall Lane on Saturday based on the xG models.

TVC15 wrote:Classic Macca
Thick as mince

Fails to understand some pretty basic stuff that a nine year old could get...then thinks they are pointless.
Then dismally fails at trying to be clever with a sarcastic post.

Already told you, you are right.

Already told you I fully understand it, but deem it pointless.
Offers me no purpose what so ever, it means nothing factual to me.

Football isnt determined, or trophies/leagues won on things that are expected to happen.

The only table that I take note of is the one that shows our position based on points gained.


Cheers for the name calling though, and to think you compared me to a 9 year old... ironic.

Tall Paul wrote:What similar reasons?

And it's not just fantasy football it's useful for, they're useful for finding value in games that the bookies have mispriced. For example, we look good value for a result at Bramall Lane on Saturday based on the xG models.

Dont bother with FF for various reasons

Too time consuming
Not what it used to be
Dont like the way many of my mates are watching games in the pub, cheering goals from other teams/players
They are on their phones all game ( in pub or on Turf ) checking who's scoring/assisting in all the other matches being played.

So I just give it a miss.

Can just about fit doing super 6 in once a week, and even then I miss about half a dozen weeks a season

MACCA wrote:Already told you, you are right.

Already told you I fully understand it, but deem it pointless.
Offers me no purpose what so ever, it means nothing factual to me.

Football isnt determined, or trophies/leagues won on things that are expected to happen.

The only table that I take note of is the one that shows our position based on points gained.


Cheers for the name calling though, and to think you compared me to a 9 year old... ironic.

Not a matter of who is right or wrong - it’s the fact that you see it as pointless I don’t get.
These type of statistics are used by all clubs - and our manager uses them as much as anyone.
To think they are pointless or only used for fantasy football only reinforces you don’t understand them.

If all that you are interested in is the league table that’s absolutely fine....I’m not interested in brain surgery or astrophysics but I can see they are not pointless.

MACCA wrote:Already told you, you are right.

Already told you I fully understand it, but deem it pointless.
Offers me no purpose what so ever, it means nothing factual to me.

Football isnt determined, or trophies/leagues won on things that are expected to happen.

The last line indicates that you don't fully understand it at all.

TVC15 wrote:Not a matter of who is right or wrong - it’s the fact that you see it as pointless I don’t get.
These type of statistics are used by all clubs - and our manager uses them as much as anyone.
To think they are pointless or only used for fantasy football only reinforces you don’t understand them.

If all that you are interested in is the league table that’s absolutely fine....I’m not interested in brain surgery or astrophysics but I can see they are not pointless.

You're right

Tall Paul wrote:The last line indicates that you don't fully understand it at all.

Nope dont understand it at all, the only reason I find it pointless is because I dont have a clue about it.

If I fully understood it, I'd find it extremely informative and probably look at the updated version weekly to help me in my gambling exploits.

When do we get our extra points credited then for having such good xg and xga? Fail to see how this could be of use to any manager or coaching staff tbh.What can it tell you that watching a few videos of their games can't?

jrgbfc wrote:When do we get our extra points credited then for having such good xg and xga? Fail to see how this could be of use to any manager or coaching staff tbh.What can it tell you that watching a few videos of their games can't?

If you’ve nothing constructive to add, probably best to say nothing

Dumb and dumber

Tall Paul wrote:What similar reasons?

And it's not just fantasy football it's useful for, they're useful for finding value in games that the bookies have mispriced. For example, we look good value for a result at Bramall Lane on Saturday based on the xG models.

Is it just a case of comparing our expected goals for against their expected goals against that makes you say that?
Never looked at these before but always good to have an edge when placing a bet.