Jump to content
IGNORED

xG so far this season


And Its Smith

Recommended Posts

I’d never bother using xg to decide what the league table should look like but it can be useful for showing if a team is getting worse results than their performance suggests or vice versa. I only ever really use the xg graphs that show an overview of the game I find these give you a much better idea of how a games panned out rather than the traditional shots/shots on target/corner stats. They obviously still won’t be as good as watching a game but alongside the highlights they can give you a reasonable view of what’s happened.

 

AFBBD998-E337-4132-80EE-C92A154715AC.png

Link to comment
Share on other sites

18 minutes ago, Ohbasso said:

I’d never bother using xg to decide what the league table should look like but it can be useful for showing if a team is getting worse results than their performance suggests or vice versa. I only ever really use the xg graphs that show an overview of the game I find these give you a much better idea of how a games panned out rather than the traditional shots/shots on target/corner stats. They obviously still won’t be as good as watching a game but alongside the highlights they can give you a reasonable view of what’s happened.

 

AFBBD998-E337-4132-80EE-C92A154715AC.png

How is it useful to show if a teams performances are better than the results though if the very information to decide xG is flawed? 

Surely Kalas’s block on Tuesday to stop a near certain goal is part of our overall performance. Yet xG ignores it

Link to comment
Share on other sites

2 minutes ago, RedDave said:

How is it useful to show if a teams performances are better than the results though if the very information to decide xG is flawed? 

Surely Kalas’s block on Tuesday to stop a near certain goal is part of our overall performance. Yet xG ignores it

Yeah it’s far from perfect and you’d be daft to base your view of a match entirely on xg. Like I said the graphs like I posted above are much better than the traditional stats you normally see after a match and are good for showing spells of pressure and whether a game was end to end or attack v defence etc.

Link to comment
Share on other sites

This xG malarkey can not possibly take tactics into account. Second half at Derby we are 2-0 up and are not going to chase the game in any way. As for chances being saved by Goalkeepers that’s what they are paid well to do!!! Baker heading everything that moved......that’s his job. It clearly doesn’t work on the basis that a team doing what it’s instructed to do and win the game deserves to win it!! If you miss sitters you don’t deserve anything and perhaps it shows a lack of ability.

Sounds like software written for the Gas Hordes to cry into their cornflakes over...we deserved to win, we deserve a new ground blah blah blah

Link to comment
Share on other sites

10 minutes ago, Ohbasso said:

Yeah it’s far from perfect and you’d be daft to base your view of a match entirely on xg. Like I said the graphs like I posted above are much better than the traditional stats you normally see after a match and are good for showing spells of pressure and whether a game was end to end or attack v defence etc.

Yep, I use the Experimental 361 xG charts too....useful to see that City’s chances were less often, but more clear.  If you look at the Statszone view too.

6DE1178F-FAD6-4127-AC9C-51783A8D4E1B.thumb.jpeg.875df05e51fae0e37d60a1af55629ebd.jpeg

Derby had 11 shots outside the box....speculative in xG terms!

I think we will see a lot of this, with our players getting blocks because our 3 CBs will cover a lot of space in front of goal.

Link to comment
Share on other sites

2 hours ago, RedDave said:

My argument against it though would be that the xG percentages are complete nonsense which makes a mockery of the whole thing.

If the basis is flawed then it is useless.  How can they say we should have lost on Tuesday?  That's what I'm mainly trying to understand

xG is about how likely a goal is from a certain chance created.

7% for Brownhills goal isn't that low when you consider how often people blaze over from there

I believe a penalty has an xG of 0.7, because it's roughly how many penalties are scored.

So Lawrence's chance from about 5 yards out will have had a very high xG. Some xG models take into account where the defenders are and others don't, so you could see quite a variation in xG numbers for a game just from looking at different models.

Link to comment
Share on other sites

5 minutes ago, Numero Uno said:

This xG malarkey can not possibly take tactics into account. Second half at Derby we are 2-0 up and are not going to chase the game in any way. As for chances being saved by Goalkeepers that’s what they are paid well to do!!! Baker heading everything that moved......that’s his job. It clearly doesn’t work on the basis that a team doing what it’s instructed to do and win the game deserves to win it!! If you miss sitters you don’t deserve anything and perhaps it shows a lack of ability.

Sounds like software written for the Gas Hordes to cry into their cornflakes over...we deserved to win, we deserve a new ground blah blah blah

Dunno- maybe expected goals isn't the best example of it but data and analytics really on the rise in football. For better or worse- xG is just one element of this.

https://bleacherreport.com/articles/2659234-epl-2030-a-data-driven-future-football-dystopia-or-beautiful-holographic-game

The above is a bit of fun but also informative.

http://outsideoftheboot.com/2016/07/21/rise-of-data-analytics-in-football/

Could reach new levels of micromanagement by coaches and management of players moving forward, but at the expense of the excitement, the flair- the unexpected.

Football has been a slow adapter to this sort of thing, but over the next decade it'll be interesting to see how it all develops.

Link to comment
Share on other sites

27 minutes ago, RedDave said:

How is it useful to show if a teams performances are better than the results though if the very information to decide xG is flawed? 

Surely Kalas’s block on Tuesday to stop a near certain goal is part of our overall performance. Yet xG ignores it

It doesn't.

xG doesn't tell you how well a team has played.

If an opportunity with an xG rating of 0.9 is missed it could be a save or the ball hit against the post from a yard out- xG doesn't say that it's a bad miss or a fantastic block, just that there was a chance.

Just like with Brownhills goal. The fact it had an xG of 0.07 and was scored doesn't mean there was a goalkeeping howler, or that Brownhill scored a worldie. Just that goals in that situation are uncommon.

Link to comment
Share on other sites

3 minutes ago, Mr Popodopolous said:

Dunno- maybe expected goals isn't the best example of it but data and analytics really on the rise in football. For better or worse- xG is just one element of this.

https://bleacherreport.com/articles/2659234-epl-2030-a-data-driven-future-football-dystopia-or-beautiful-holographic-game

The above is a bit of fun but also informative.

http://outsideoftheboot.com/2016/07/21/rise-of-data-analytics-in-football/

Could reach new levels of micromanagement by coaches and management of players moving forward, but at the expense of the excitement, the flair- the unexpected.

Football has been a slow adapter to this sort of thing, but over the next decade it'll be interesting to see how it all develops.

The problem comes when you become reliant on this stuff. GJ used to bang on about Prozone stats all the time from my hazy memory. Even if you do agree with the principle of how it or other analyticals out there work (and LJ might, who knows) you have to trust what you SEE in conjunction with it. If it identifies a minor tweak you need to make then great but reliance on this stuff is dangerous imo.

Link to comment
Share on other sites

8 minutes ago, JamesBCFC said:

It doesn't.

xG doesn't tell you how well a team has played.

If an opportunity with an xG rating of 0.9 is missed it could be a save or the ball hit against the post from a yard out- xG doesn't say that it's a bad miss or a fantastic block, just that there was a chance.

Just like with Brownhills goal. The fact it had an xG of 0.07 and was scored doesn't mean there was a goalkeeping howler, or that Brownhill scored a worldie. Just that goals in that situation are uncommon.

How is that 0.07 judged? The probability of Josh Brownhill sticking that chance in the onion bag or does it include League 2 players. For example the probability of De Bruyne putting that away would be nearer 40% or probably higher on reflection than 7% imo.

Link to comment
Share on other sites

8 minutes ago, JamesBCFC said:

It doesn't.

xG doesn't tell you how well a team has played.

If an opportunity with an xG rating of 0.9 is missed it could be a save or the ball hit against the post from a yard out- xG doesn't say that it's a bad miss or a fantastic block, just that there was a chance.

Just like with Brownhills goal. The fact it had an xG of 0.07 and was scored doesn't mean there was a goalkeeping howler, or that Brownhill scored a worldie. Just that goals in that situation are uncommon.

I guess my issue might be more with the way that pundits are talking about xG.  As in saying one team deserved to win as they had a higher xG.

 

Link to comment
Share on other sites

1 minute ago, Numero Uno said:

How is that 0.07 judged? The probability of Josh Brownhill sticking that chance in the onion bag or does it include League 2 players. For example the probability of De Bruyne putting that away would be nearer 40% than 7% imo.

Can't be certain but think it's essentially all shots from that area and all goals from that area.

There's an almost endless amount of things that can go into an xG model

This could be worth a read (although it loaded up horribly on my phone)

http://alexrathke.net/expected-goals/

So on some models things like the weight of the pass can come into play.

 

xG is still relatively new and I expect over the next few years the models will get more and more refined, perhaps even to the point where there's a different xG for each of the divisions as more data becomes available.

Link to comment
Share on other sites

To me, all xG says is how clinical your strikers are, you aren’t in the wrong position in the table because your striker can’t hit a barn door. I seem to recall we had quite a high xG last season but, for all his positives, Fam isn’t that stone cold clinical striker like others.

Link to comment
Share on other sites

4 minutes ago, JamesBCFC said:

Can't be certain but think it's essentially all shots from that area and all goals from that area.

There's an almost endless amount of things that can go into an xG model

This could be worth a read (although it loaded up horribly on my phone)

http://alexrathke.net/expected-goals/

So on some models things like the weight of the pass can come into play.

 

xG is still relatively new and I expect over the next few years the models will get more and more refined, perhaps even to the point where there's a different xG for each of the divisions as more data becomes available.

A lot needs to go into a model like that to make it accurate. Take Brownhill’s goal at Derby......he’s running into a perfectly weighted ball, he’s in total control to the point he shouts Weimann to leave it...he knows he’s going to be ripping the onion bag (as much as you can know!!) so you could argue that the percentage chance of scoring should be far higher than 7% compared to being in exactly the same position with a defender closing him and more weight on the pass?

As you say if people start believing in it they will have to refine it a lot further. But for me I wouldn’t use it to say your 7th position in the league is “lucky”....no way.

Link to comment
Share on other sites

The more I think about it the more I disagree with it. A player like Fam will hold the ball up well, bring other forwards in to play & create chances for others but will miss a lot himself, giving us a high xG and a low actual goals ratio. Whereas, a player like Tammy won’t do any of that work that Fam is so good at, but when he has a chance he’ll put it away, giving him a low xG but a high actual goals ratio. On paper you’d obviously think the low xG/high aG was the far better option but obviously that isn’t the case.

(I’m not actually comparing Tammy/Fammy, but players of their mould)

Link to comment
Share on other sites

23 minutes ago, Numero Uno said:

A lot needs to go into a model like that to make it accurate. Take Brownhill’s goal at Derby......he’s running into a perfectly weighted ball, he’s in total control to the point he shouts Weimann to leave it...he knows he’s going to be ripping the onion bag (as much as you can know!!) so you could argue that the percentage chance of scoring should be far higher than 7% compared to being in exactly the same position with a defender closing him and more weight on the pass?

As you say if people start believing in it they will have to refine it a lot further. But for me I wouldn’t use it to say your 7th position in the league is “lucky”....no way.

That I agree on, particularly so early in the season.

What xG could be used with is other statistics (and eyes) to suggest whether a team is underperforming.

An example being Juventus a couple of seasons ago. Their xG was 2+ higher that their opponents frequently, but they drew or lost a fair number of their first 12 games.

So such a consistent amount of games where they scored well below their xG says that their results aren't always reflective of the performances. They then went and steamrollered their way through the rest of the season and won the title at a canter.

 

A lot of statistics in football is knowing what to use and when.

Do multiple different stats give the same impression of a game, etc.

Link to comment
Share on other sites

3 hours ago, Unan said:

The more I think about it the more I disagree with it. A player like Fam will hold the ball up well, bring other forwards in to play & create chances for others but will miss a lot himself, giving us a high xG and a low actual goals ratio. Whereas, a player like Tammy won’t do any of that work that Fam is so good at, but when he has a chance he’ll put it away, giving him a low xG but a high actual goals ratio. On paper you’d obviously think the low xG/high aG was the far better option but obviously that isn’t the case.

(I’m not actually comparing Tammy/Fammy, but players of their mould)

This is Fammy v Tammy last season, adjusted to remove penalties.  Think Fam had 2, Tammy 7.  Something like that.

3AF30EB4-4051-4F2A-9E48-76AD8933264C.thumb.jpeg.b903c0909451d9686b55c8f9fe285a0e.jpeg

What does this show us?

Chart 1- Tammy scores more goals (non pens) per 90, but Fam is still above average for all players that played as strikers last season.

Chart 2 - Tammy gets more / better chances per 90.

Chart 3 - they actually take their chances at a similar rate (so give Fam the types of chances that Tammy gets, and he might score as many).

Interesting isn’t it.

Link to comment
Share on other sites

2 hours ago, Davefevs said:

This is Fammy v Tammy last season, adjusted to remove penalties.  Think Fam had 2, Tammy 7.  Something like that.

3AF30EB4-4051-4F2A-9E48-76AD8933264C.thumb.jpeg.b903c0909451d9686b55c8f9fe285a0e.jpeg

What does this show us?

Chart 1- Tammy scores more goals (non pens) per 90, but Fam is still above average for all players that played as strikers last season.

Chart 2 - Tammy gets more / better chances per 90.

Chart 3 - they actually take their chances at a similar rate (so give Fam the types of chances that Tammy gets, and he might score as many).

Interesting isn’t it.

I was thinking Tammy wasn’t a great example as he was in that Villa team. Saying that I can’t believe how many changes he missed! A relatively poor conversion rate all things considering.

Link to comment
Share on other sites

8 hours ago, Unan said:

To me, all xG says is how clinical your strikers are, you aren’t in the wrong position in the table because your striker can’t hit a barn door. I seem to recall we had quite a high xG last season but, for all his positives, Fam isn’t that stone cold clinical striker like others.

This, xG helps when looking at strikers and how well they finish chances, in the chart below the line runs at xG and goals per 90 being equal. In recruitment it can help find under valued strikers.

Image

Link to comment
Share on other sites

xG- puts me in mind of that long time American poster who came over to watch the City v Hull play off final.

He was put up over night by someone on the forum iirc?

Got it, x3widRacin, not very active on here since and not at all since January last year.

Can only imagine witnessing a big occasion City let down might have gradually cooled his BCFC ardour!?

Not used to it like the rest of us!!

Link to comment
Share on other sites

Archived

This topic is now archived and is closed to further replies.

  • Recently Browsing   0 members

    • No registered users viewing this page.
×
×
  • Create New...