Table based on xG

[ExiledAjax]

57 minutes ago, Mr Popodopolous said:

Had a quick look at Experimental 3-6-1.

Actually had us actually if anything, shading it at Burnley, 1.4-1.2.

Burnley 1.20 v Bristol City 1.40

Although rounding wise and xG margin for error, that feels like draw.

Average of the 4 sources I look at was 1.39 for us and 1.30 for Burnley. So yeh both teams created roughly the same overall chance to score.

[Mr Popodopolous]

23 minutes ago, ExiledAjax said:

Average of the 4 sources I look at was 1.39 for us and 1.30 for Burnley. So yeh both teams created roughly the same overall chance to score.

Yeah, quality of chances doesn't necessarily align with baseline stats ie possession, shots etc.

Would say though that when factoring this in, his table only had us a couple of points away from our actual return to date. We have 14, think his model had us on 11 and that was pre Norwich and Burnley away.

[Ivorguy]

Meaningless stats

[CiderJar]

2 hours ago, Ivorguy said:

Meaningless stats

Miserable (rhymes with stat)

[Mr Popodopolous]

4 hours ago, Ivorguy said:

Meaningless stats

In themselves they can be. Indeed stats can be used to disprove or prove anything.

However they can also act as a bellwether. Do you remember when we went 10 unbeaten 3 years ago? Culminating in the win v Reading in October 2019 before the international break, early goal by Diedhiou.

These results were masking our performances somewhat, and lo our results took a down turn. Stats defo aren't the be all and end all and a big intangible was that:

Kalas- From match day 4 or 5 was out

DaSilva- Matchday 1 and then injured!

Nagy- Injured in late August, maybe 1.5 games in?

Afobe- Integral to the new system and I would argue, getting the best out of Palmer was injured just 4 or 5 games into his loan spell.

Palmer feeding Weimann and Afobe was very different to feeding Weimann and Diedhiou. Diedhiou for Adobe, totally different players.

DaSilva and Afobe, these two were very important to the new system IMO. Kalas and Nagy more replaceable but the former could play a higher line more readily than Williams I think and Nagy oh yes Korey was still out too with recurring long term injuries. Bad luck.

4-5 starters out by mid September..perhaps our results and underlying performances would have been better with them available! Might not have needed to overplay Masengo too.

Edited September 21, 2022 by Mr Popodopolous

[Mr Popodopolous]

For those who are interested.

Albeit this is the Experimental 3-6-1 version and there are different models, which I appreciate.

Edited October 10, 2022 by Mr Popodopolous

[joe jordans teeth]

With all these stats about that people take seriously can you tell me if we are doomed,safe or promotion contenders,save me a lot of money and worry for the rest of the season 

[Davefevs]

12 minutes ago, joe jordans teeth said:

With all these stats about that people take seriously can you tell me if we are doomed,safe or promotion contenders,save me a lot of money and worry for the rest of the season 

All three

[joe jordans teeth]

Just now, Davefevs said:

All three

Just trust my eyes and knowledge of football then I guess

[Mr Popodopolous]

27 minutes ago, joe jordans teeth said:

With all these stats about that people take seriously can you tell me if we are doomed,safe or promotion contenders,save me a lot of money and worry for the rest of the season 

Not an exact science but neither up or down IMO.

Probably some good, striking wins, interspersed with some abysmal defeats, some high scoring wins, draws and losses, some entertaining games, some shockers.

Midtable to lower midtable.

Edited October 10, 2022 by Mr Popodopolous

[The turtle]

 

A coach at a Premier League side told me a story of his manager being convinced by their data department to operate a high line against a team with a notably quick forward, despite a first-choice centre-back having to be replaced by a veteran who was just returning from injury and hadn’t been quick on the turn even in his pomp

They conceded three within 30 minutes and lost 3-0, but the analysts justified their advice by pointing out their team had won the xG. But that was because, as the coach angrily replied, having scored with three early chances, the other team had no need to attack. They sat back, conserved energy and weren’t much bothered if they conceded a couple of half-chances: the game was over with an hour remaining. That’s not to say that xG is not a very useful tool – it is – merely that it doesn’t always give the whole picture.

https://www.theguardian.com/football/blog/2022/sep/17/football-tacticians-bowled-over-by-quick-fix-data-risk-being-knocked-for-six

The problem with Xg isn't so much that fans use it. It's more it's only useful if you've watched the game and understand how the game script played out.

 

 

 

[joe jordans teeth]

3 minutes ago, The turtle said:

 

A coach at a Premier League side told me a story of his manager being convinced by their data department to operate a high line against a team with a notably quick forward, despite a first-choice centre-back having to be replaced by a veteran who was just returning from injury and hadn’t been quick on the turn even in his pomp

They conceded three within 30 minutes and lost 3-0, but the analysts justified their advice by pointing out their team had won the xG. But that was because, as the coach angrily replied, having scored with three early chances, the other team had no need to attack. They sat back, conserved energy and weren’t much bothered if they conceded a couple of half-chances: the game was over with an hour remaining. That’s not to say that xG is not a very useful tool – it is – merely that it doesn’t always give the whole picture.

https://www.theguardian.com/football/blog/2022/sep/17/football-tacticians-bowled-over-by-quick-fix-data-risk-being-knocked-for-six

The problem with Xg isn't so much that fans use it. It's more it's only useful if you've watched the game and understand how the game script played out.

 

 

 

In other words it is bloody useless then

[Davefevs]

5 minutes ago, The turtle said:

 

A coach at a Premier League side told me a story of his manager being convinced by their data department to operate a high line against a team with a notably quick forward, despite a first-choice centre-back having to be replaced by a veteran who was just returning from injury and hadn’t been quick on the turn even in his pomp

They conceded three within 30 minutes and lost 3-0, but the analysts justified their advice by pointing out their team had won the xG. But that was because, as the coach angrily replied, having scored with three early chances, the other team had no need to attack. They sat back, conserved energy and weren’t much bothered if they conceded a couple of half-chances: the game was over with an hour remaining. That’s not to say that xG is not a very useful tool – it is – merely that it doesn’t always give the whole picture.

https://www.theguardian.com/football/blog/2022/sep/17/football-tacticians-bowled-over-by-quick-fix-data-risk-being-knocked-for-six

The problem with Xg isn't so much that fans use it. It's more it's only useful if you've watched the game and understand how the game script played out.

 

 

 

And you can often apply exactly that analogy to possession, once you are ahead on the game.  The team in the lead can keep shape, force the losing team into non-threatening possession.

[And Its Smith]

Steve Bruce sacked when top of the xG league suggests he’s been sacked because his strikers can’t finish chances and/or he’s got a shit keeper. Tough times 

[joe jordans teeth]

Does anyone know if Ferguson used this tosh years ago while he dominated the league because he sure knew how to get rid of players at the right time 

[Davefevs]

1 minute ago, joe jordans teeth said:

In other words it is bloody useless then

It’s useful as a trend over longer periods.  You can see that we take the vast majority of our big chances (high XG shots).  We are pretty clinical.  I’d rather one penalty (circa 0.8 XG) than 8 chances worth 0.10 XG, if that makes sense in terms of pure XG (rather than the excitement of creating chances) if you get my drift.

 

[Mr Popodopolous]

To be honest, that is an interesting story by @The turtle and data definitely isn't flawless!

To play a high line with a veteran CB, let alone one just returning from injury seems crazy to me.

Has to be applied well, I think it is useful for trends, and can be a sign of things to come but can also overshoot/excessively 'correct' itself in either direction.

[joe jordans teeth]

Just now, Davefevs said:

It’s useful as a trend over longer periods.  You can see that we take the vast majority of our big chances (high XG shots).  We are pretty clinical.  I’d rather one penalty (circa 0.8 XG) than 8 chances worth 0.10 XG, if that makes sense in terms of pure XG (rather than the excitement of creating chances) if you get my drift.

 

I’ve never claimed to be the brightest but that is head hurting stuff,I’m sure sir Alex would tell you where you could stick those stats 

[Mr Popodopolous]

I won't do it or maybe I will but not just yet- but an interesting experiment for this season would be xG sans penalties.

Given our lack and this goes back nearly 2 years, it'd be quite interesting to see how teams would improve or regress.

Would ultimately be a pointless exercise though!

Edited October 10, 2022 by Mr Popodopolous

[And Its Smith]

5 minutes ago, Mr Popodopolous said:

I won't do it or maybe I will but not her, but an interesting experiment for this season would be xG sans penalties.

Given our lack and this goes back nearly 2 years, it'd be quite interesting to see how teams would improve or regress.

Would ultimately be a pointless exercise though!

I’ve got a few issues with xG and penalties is one. You get just inside the box and create a 0.2 xG chance and get clipped and rewarded with a 0.75 xG reward which gets recorded.  In my opinion the 0.2 should be recorded not the 0.75.  Also if there’s a goal mouth scramble with, say, 3 shots, a team will be given xG for all 3 shots which makes no sense 

[Mr Popodopolous]

2 minutes ago, And Its Smith said:

I’ve got a few issues with xG and penalties is one. You get just inside the box and create a 0.2 xG chance and get clipped and rewarded with a 0.75 xG reward which gets recorded.  In my opinion the 0.2 should be recorded not the 0.75.  Also if there’s a goal mouth scramble with, say, 3 shots, a team will be given xG for all 3 shots which makes no sense 

I've actually started but not finished, on some kinda xP model although using more basic baseline stats such as shots and shots on target..could fouls also help here?

I thought the penalty xG was 0.8 but anyway, it's an interesting one and a hard circle to square.

[Chairman Mao]

42 minutes ago, The turtle said:

 

A coach at a Premier League side told me a story of his manager being convinced by their data department to operate a high line against a team with a notably quick forward, despite a first-choice centre-back having to be replaced by a veteran who was just returning from injury and hadn’t been quick on the turn even in his pomp

They conceded three within 30 minutes and lost 3-0, but the analysts justified their advice by pointing out their team had won the xG. But that was because, as the coach angrily replied, having scored with three early chances, the other team had no need to attack. They sat back, conserved energy and weren’t much bothered if they conceded a couple of half-chances: the game was over with an hour remaining. That’s not to say that xG is not a very useful tool – it is – merely that it doesn’t always give the whole picture.

https://www.theguardian.com/football/blog/2022/sep/17/football-tacticians-bowled-over-by-quick-fix-data-risk-being-knocked-for-six

The problem with Xg isn't so much that fans use it. It's more it's only useful if you've watched the game and understand how the game script played out.

 

 

 

some wag thinks it was Brendan Rodgers in a Leicester match in the comments below. 

[And Its Smith]

6 minutes ago, Mr Popodopolous said:

I've actually started but not finished, on some kinda xP model although using more basic baseline stats such as shots and shots on target..could fouls also help here?

I thought the penalty xG was 0.8 but anyway, it's an interesting one and a hard circle to square.

Yeah I think it’s recently changed from 0.75 to 0.8 but only a small movement.

[Davefevs]

56 minutes ago, And Its Smith said:

I’ve got a few issues with xG and penalties is one. You get just inside the box and create a 0.2 xG chance and get clipped and rewarded with a 0.75 xG reward which gets recorded.  In my opinion the 0.2 should be recorded not the 0.75.  Also if there’s a goal mouth scramble with, say, 3 shots, a team will be given xG for all 3 shots which makes no sense 

Dave - again the principle and application of XG is fine, it’s a score for the likelihood of each shot taken, regardless of what happened before and after.  It gets misused (imho) by too many people adding each shot together in a match to try to equate the total for and against into some form of result…then Doubling down on it by putting it into a league table.  That was not the original intention.

Your two examples are a good reason why you shouldn’t try to form a result from the sum of each shot.  It’s possible in a goalmouth scramble for the total XG > 1.  That’s a big flaw, not in XG itself but people mis-using it.

1 hour ago, And Its Smith said:

Yeah I think it’s recently changed from 0.75 to 0.8 but only a small movement.

Wyscout is 0.76

[joe jordans teeth]

1 hour ago, Chairman Mao said:

some wag thinks it was Brendan Rodgers in a Leicester match in the comments below. 

Must be this bit then 

[Mr Popodopolous]

1 hour ago, And Its Smith said:

Steve Bruce sacked when top of the xG league suggests he’s been sacked because his strikers can’t finish chances and/or he’s got a shit keeper. Tough times 

Probably a bit of both, weird outlier but they have dominated most games...the strange thing is Karlan Grant on paper is a capable striker, unsure who else they have but with Wallace and Swift able to assist and score, well they should be doing much better than they are IMO.

[HiddenGem7]

13 hours ago, Mr Popodopolous said:

Probably a bit of both, weird outlier but they have dominated most games...the strange thing is Karlan Grant on paper is a capable striker, unsure who else they have but with Wallace and Swift able to assist and score, well they should be doing much better than they are IMO.

And this highlights the other issue with xG - it takes no account of the players ability. A chance would get the same xG regardless if it fell to Haaland or Jay DaSilva (no disrespect intended to JD but he isn't a great finisher). Clearly one would have a greater likelihood of scoring any given chance than the other but the xG system simply ignores it. 

[Davefevs]

20 minutes ago, HiddenGem7 said:

And this highlights the other issue with xG - it takes no account of the players ability. A chance would get the same xG regardless if it fell to Haaland or Jay DaSilva (no disrespect intended to JD but he isn't a great finisher). Clearly one would have a greater likelihood of scoring any given chance than the other but the xG system simply ignores it. 

There is more sophistication coming over time with XG including post-shot XG (PSXG) which start to build in the fact that Haaland hits more cow’s arses with banjos than Dasilva.  So it’s coming. 

[ExiledAjax]

1 hour ago, Davefevs said:

There is more sophistication coming over time with XG including post-shot XG (PSXG) which start to build in the fact that Haaland hits more cow’s arses with banjos than Dasilva.  So it’s coming. 

It's already present in the more sophisticated models.

[Hxj]

There's only one stat in football that counts - the place in the table at the end of the season.  Everything else is attempting to explain that.  (I know I repeat myself!)

Plus if xG is so good why does it predict 25 odd goals scored more than were actually scored?  Are football statisticians as biased as everyone else?

As to @Davefevsconundrum on a penalty v 8 shots.  He would win 34/100 games, the draw would be a better bet with 40/100 games, and he would lose 26/100 games.  Much closer than I actually thought.  However he would miss out in scoring 8 goals once in 100,000,000 games