xG and finishing under ETH

It's not, though. The proof is that, throughout a player's career, it all evens out. A player may be over performing for a particular period, but it becomes pretty balanced in the end. Yes, this applies even to the likes of Ronaldo and Lewandowski. Believe it or not, they aren't outperforming their xG. What makes them, and the likes of Haaland, so special is that they accumulate insane numbers of xG with their movement and understanding of the game.
I actually deleted my post because I realised the person was agreeing with my sentiment.

You’ve replied with what seems like the opposite view which I don’t agree with. It’s still an absurd way of looking at it and it’s probably the reason no one uses it based on how you are suggesting.

Adjusting xG based on the ability of the striker has no benefit other than for the individual to see if they are improving or regressing. But the xG that is used by opta which bases it on the situation is more useful.
 
I actually deleted my post because I realised the person was agreeing with my sentiment.

You’ve replied with what seems like the opposite view which I don’t agree with. It’s still an absurd way of looking at it and it’s probably the reason no one uses it based on how you are suggesting.

Adjusting xG based on the ability of the striker has no benefit other than for the individual to see if they are improving or regressing. But the xG that is used by opta which bases it on the situation is more useful.

Most models are using the same philosophy. Even Opta. What anybody thinks isn't important in the grand scheme of things. It's not absurd because, even in the cases of the greatest finishers of our era (Ronaldo, Lewandowski, Kane, Suarez), they get it right. Their xG matches their final output. A lot of people get it wrong. It's not about outperforming it (nobody does, except for Messi), it's about getting high numbers. This is what they are basically recording.




Aside from the goals and Casemiro misjudging the open goal, did we have any other chances we should have scored yesterday? It didn’t seem as though we were missing chance after chance.

Bruno's shot in the 59th minute had an xG 0.32. Best chance of the game outside the goals for both sides. It counts as a big chance for Opta since it goes in 1/3 times.
 
Most models are using the same philosophy. Even Opta. What anybody thinks isn't important in the grand scheme of things. It's not absurd because, even in the cases of the greatest finishers of our era (Ronaldo, Lewandowski, Kane, Suarez), they get it right. Their xG matches their final output. A lot of people get it wrong. It's not about outperforming it (nobody does, except for Messi), it's about getting high numbers. This is what they are basically recording.
I don’t know if you are disagreeing with me or agreeing.

My point is if a particular shot/scenario has an xG of 0.3, it is the same xG for Haaland as it would be for Weghorst. It shouldn’t be any different. That’s how opta calculate it and that’s how it should be done. To calculate it with different values depending on who the player is would just be weird.
 
Aside from the goals and Casemiro misjudging the open goal, did we have any other chances we should have scored yesterday? It didn’t seem as though we were missing chance after chance.
Thought the same, we were getting into great areas but didn't test the keeper that much other than Antony and Bruno a couple of times from range.
Martial was basically non existent. You would barely notice him playing at all. It's like we're playing with 10 men. We created that much chance without a #9 we'd do much better when we have proper one I reckon. I know that ship had sailed but I really wanted us to sign Kane. He'd be perfect for ETH. What a shame.
 
I don’t know if you are disagreeing with me or agreeing.

My point is if a particular shot/scenario has an xG of 0.3, it is the same xG for Haaland as it would be for Weghorst. It shouldn’t be any different. That’s how opta calculate it and that’s how it should be done. To calculate it with different values depending on who the player is would just be weird.

Yes, i'm underlining the point by saying that even sublime strikers, who a lot of people think they have to be massively outperforming their xG stats, always score the number of goals the models suggest they should have scored. Ronaldo, for example, is the greatest goalscorer of all time, not (so much) because he's scoring chances others can't, but because his intelligence and skill allow him to "see" and "find" finishes that most forwards aren't able to. Thinking that it is the norm for top strikers to outperform their xG is where most misunderstandings occur with xG and you get people wondering how can Weghorst and Haaland get the same xG for the same shot.
 
Last edited:
I note that we currently have the second highest xG/90 in the Premier League (2.40, behind only Brighton at 2.57). Surely that is a promising sign, especially considering that we haven't exactly been playing that well overall. And also considering that last season's was 1.78.

xGA/90 not as positive - 11th, at 1.73. Significantly worse than last season's 1.33.

It is what it is, of course - just three games, small sample. Still, so far we're creating many more scoring chances than we did last season, while not looking as solid defensively.
 
Last edited:
I note that we currently have the second highest xG/90 in the Premier League (2.40, behind only Brighton at 2.57). Surely that is a promising sign, especially considering that we haven't exactly been playing that well overall. And also considering that last season's was 1.78.

xGA/90 not as positive - 11th, at 1.73. Significantly worse than last season's 1.33.

It is what it is, of course - just three games, small sample. Still, so far we're creating many more scoring chances than we did last season, while not looking as solid defensively.
Yes it is very promising and I think it’s going under the radar a lot with fans and pundits. It means if we can start taking our chances hopefully with Hojlund too, we are doing the right things.

I’m still frustrated with the spurs game as we did enough to put the game beyond spurs in the first half.
 
Aside from the goals and Casemiro misjudging the open goal, did we have any other chances we should have scored yesterday? It didn’t seem as though we were missing chance after chance.
This is also my feeling of the season so far. I have a bit of a problem with how those half chances are calculated (we seem decent at creating half chances, just don't create nearly enough clear cut chances).
 
This is also my feeling of the season so far. I have a bit of a problem with how those half chances are calculated (we seem decent at creating half chances, just don't create nearly enough clear cut chances).

There's Antony's shot - not a big chance per se, but with a pretty high Post-Shot xG. Also, Bruno's flick outside in the 59th minute. Those two had a combined xG of 0.73 (if you use PSxG for Antony's).

With the np goals and Casemiro's header, that's 5 chances with an xG or PSxG of more than 0.20 - plus a penalty won. Not bad, although Forest's four were a bit on the high side too.

Against Spurs we had four such (of whom 3 came in the first half), same as Spurs.

Against Wolves, we had two (including Varane's goal), against Wolves' three.

In other words - we've produced more good scoring chances/finishes with each passing game, and what we had against Forest equals what we managed in the first two games combined, if we include the penalty.
 
Last edited:
Finishing not so much of an issue today but a lack of chances created.

Fine margins though as Jesus' goal was a good chance when we were chasing the game late and Garnacho was only just offside.

 
Oh wow, an xG Twitter post. Like Arsenal's xG was so high at 1-1.

Definite lack of chances. Arsenal sat back which is why good transitions were important. That's where all of our danger came from.

Our biggest mistake this game was the collective defending of their 1st goal.
 
Oh wow, an xG Twitter post. Like Arsenal's xG was so high at 1-1.

Definite lack of chances. Arsenal sat back which is why good transitions were important. That's where all of our danger came from.

Our biggest mistake this game was the collective defending of their 1st goal.

If I'm reading fbref correctly, Arsenal's xG up to Rice's shot that made it 2-1 was 1.5. The last two goals had a combined 0.81 xG, and there were no other shots in that period.

Remarkably, Rice's shot has a post-shot xG of only 0.12. In other words, that exact shot from that exact position fails to go in almost 9 times out of 10. The xG for it is even less (0.05). I guess that underlines the crucial impact of the blocking work they put in on Evans. It should be whoever it was wrestling him getting the headlines - not Rice being touted as The Man Of The Big Moments. He just fired off a low-percentage shot from a low-percentage position where he shouldn't normally be allowed to get a shot off on goal at all. Which translates as "luck".
 
Last edited:
https://understat.com/league/EPL

We're 11th on xPTS. 3 of 5 games have been at home. Faced two of the relegation candidates.

So much for the progress everyone was going about after the first two games.

What? Even if you call Forest a relegation candidate which they are not, we won that game. We lost 3 games to top 4 contenders. Making it look like we had an easy start and fumbled it is absurd. Of course we should have done better but non of our losses is a shock result.
 
What? Even if you call Forest a relegation candidate which they are not, we won that game. We lost 3 games to top 4 contenders. Making it look like we had an easy start and fumbled it is absurd. Of course we should have done better but non of our losses is a shock result.
Losing 3-1 at home to brighton isn't a shock?
 

This was our worst xG loss at home in the xG era. Literally rock bottom.

Previous worst was the hammering against Spurs during lockdown when we got a penalty early then collapsed after a red card.

**Times Manchester United conceded >2xG at Old Trafford since 2017/18 season:**
*(**3 further games conceded >2xG but also accumulated >2xG themselves)*

MUN 0.9 - 4.0 MCI on 29-10-2023 https://fbref.com/en/matches/f6bfec...anchester-City-October-29-2023-Premier-League

MUN 0.9 - 3.4 TOT on 04-10-2020 https://fbref.com/en/matches/cded7e...ttenham-Hotspur-October-4-2020-Premier-League

MUN 1.6 - 3.2 LIV on 24-10-2021 https://fbref.com/en/matches/b886be...ited-Liverpool-October-24-2021-Premier-League]

MUN 0.4 - 2.5 PSG on 12-02-2019 https://fbref.com/en/matches/59a63a...int-Germain-February-12-2019-Champions-League]

MUN 1.1 - 2.3 VIL on 29-09-2021 https://fbref.com/en/matches/8ad62c...Villarreal-September-29-2021-Champions-League

Credit: JC on FPL Analytics Discord
 
xG differential (our xG minus our xGA) in the PL this season:

1. Luton 1.7
2. Forest 1.6
3. Brentford 0.5
4. Crystal Palace 0.4
Spurs 0.4
6. Woves 0
7. Burnley -0.1
8. Sheffield United -0,3
Fulham -0,3
10. Brighton -0.7
11. Arsenal -1.4
12. City -3.1

Shows, as did the eye test, that the Luton game was really only the second time this season that we dominated chance creation in a PL game. As such, that's a real positive even though the result was no better than against Fulham or Burnley.
 
This seems accurate enough. According to Opta, we average 1.41 xG and 1.43 xGA after 12 games. Both numbers are worse than last season's averages, which weren't impressive to begin with. One could argue that there's a small paradox here in the sense that, although United threaten as much as they're being threatened, we haven't drawn a game yet. We have seven wins in the league, all of them by a single goal. This is reflected in the stats: You can't have a safe cushion when you don't create more and better chances than your opponent. Other than that, there's zero balance. If the goal was to turn up the volume in the attack with high pressure and aggression, this isn't reflected on the pitch. On the contrary, an alarmingly above 1 xgA last season (despite the clean sheets), has become even worse. The numbers look mediocre because the team is plain bad. The ceiling for a good performance is a narrow win, and it just takes a good effort from the opponent, an individual error or a 50/50 call going the other way for all to go straight to hell. Most of our games are similar to a coin toss. Since the coin can't land on its edge, we don't know what we're going to get.

A few other observations for comparison (for anyone who might be interested):

City are still the best team on the island, but this isn't really news. Their xG difference remains the best in the PL, but at 0.96 it's the first time we see it below 1 during Guardiola's era. Their 1.80 xG is also the worse recorded under Pep. Do they seem a bit more mortal this season? Time will tell. But it looks like KdB's long absence is having an effect, after all, even in such a well-oiled machine like City's. People have also correctly pointed out that they missed Rodri dearly in the games they lost thus far. I'd add Stones, too. Thing is, 0.84 xGA is also the worst they have produced under a manager who is obsessed with control. Last season, Haaland's introduction saw Guardiola using 5 defenders in the build-up to accommodate a forward who rarely contributes in ball circulation. This season, he has developed a good understanding of spaces and off the ball movement with Alvarez, who isn't exactly the prototypical Pep player in the attacking half, either. Haaland/Alvarez, while trying to integrate a new midfielder (Kovacic instead of Gundogan) seems to be a bit too much when it comes to overall balance. Doku is also a player who doesn't offer control and possession with his dribbling, but immediate impact. It all looks a bit un-Guardiola-like at the moment. Even with Rodri on the pitch, the 4-4 draw at The Bridge is exactly the type of game Guardiola wants to avoid.

Liverpool with 1.93 xG are the best team in the final third. This shouldn't come as a surprise, since Klopp's way to fix things is, more often than not, "more attack" (Liverpool have often achieved an xG above 2). The addition of Szoboszlai in the midfield and that designed box in the midfield, with TAA tucking inside, shows his intentions. Still, Liverpool are a team in the process of rebuilding. Mac Allister is a lot of things, but a designated defensive midfielder who will help maintain a semblance of balance in a side that's set primarily to attack, he is not. Klopp could have used Fabinho/Henderson's experience for just one season. Up front, they look great, with Salah being his usually magnificent self and even Nunez (despite his shortcomings), being the agent of chaos that, unlike Pep, Klopp is far more comfortable with. But having the 4th (Szoboszlai) and the 8th (Mac Allister) most dribbled past players in the PL in your midfield is a problem. Especially when you're chasing City.

Arteta's men seem to be the most "changed" side compared to last season. Their 1.47 xG is significantly worse than last season's 1.83 xG. Their 0.72 xGA is what allows them to stay near the top of the table. They have played a lot of difficult games at home, so things may change in the foreseeable, but nevertheless, it is an impressive statistic at this level of football. And with Timber out with injury. One could argue that Rice's impact is more than obvious. This is what he brings to the table, first and foremost. On the other hand, Arsenal do seem to struggle with Havertz and his role on the pitch. I don't know why we haven't seen him as a false #9 more often. Arteta is still trying to figure out what his best midfield is going forward.

It will be interesting to see which path, Klopp's or Arteta's, will get closer to City. United, on the other hand, don't seem to have a point of reference for our season. Something to start from and then see in which areas we can focus on to see incremental improvements.

On a side note, because the love-in for one or the other component of McFred is still a thing, and it also affects the discussions about what we need in our midfield, i feel it's worth pointing out that the best sides defensively (xGA), the ones that are being threatened less than anybody else, are the two sides (among the big boys) who base their defending around possession and control through passing and movement. Arsenal and City.
 
Last edited:
On a side note, because the love-in for one or the other component of McFred is still a thing, and it also affects the discussions about what we need in our midfield, i feel it's worth pointing out that the best sides defensively (xGA), the ones that are being threatened less than anybody else, are the two sides (among the big boys) who base their defending around possession and control through passing and movement. Arsenal and City.
Uh...thats a nice burn and might bring you the rightful anger of some of our scottish fans.

Apart from that, thanks for the rundown. Very interesting read.
 
This seems accurate enough. According to Opta, we average 1.41 xG and 1.43 xGA after 12 games. Both numbers are worse than last season's averages, which weren't impressive to begin with. One could argue that there's a small paradox here in the sense that, although United threaten as much as they're being threatened, we haven't drawn a game yet. We have seven wins in the league, all of them by a single goal. This is reflected in the stats: You can't have a safe cushion when you don't create more and better chances than your opponent. Other than that, there's zero balance. If the goal was to turn up the volume in the attack with high pressure and aggression, this isn't reflected on the pitch. On the contrary, an alarmingly above 1 xgA last season (despite the clean sheets), has become even worse. The numbers look mediocre because the team is plain bad. The ceiling for a good performance is a narrow win, and it just takes a good effort from the opponent, an individual error or a 50/50 call going the other way for all to go straight to hell. Most of our games are similar to a coin toss. Since the coin can't land on its edge, we don't know what we're going to get.

A few other observations for comparison (for anyone who might be interested):

City are still the best team on the island, but this isn't really news. Their xG difference remains the best in the PL, but at 0.96 it's the first time we see it below 1 during Guardiola's era. Their 1.80 xG is also the worse recorded under Pep. Do they seem a bit more mortal this season? Time will tell. But it looks like KdB's long absence is having an effect, after all, even in such a well-oiled machine like City's. People have also correctly pointed out that they missed Rodri dearly in the games they lost thus far. I'd add Stones, too. Thing is, 0.84 xGA is also the worst they have produced under a manager who is obsessed with control. Last season, Haaland's introduction saw Guardiola using 5 defenders in the build-up to accommodate a forward who rarely contributes in ball circulation. This season, he has developed a good understanding of spaces and off the ball movement with Alvarez, who isn't exactly the prototypical Pep player in the attacking half, either. Haaland/Alvarez, while trying to integrate a new midfielder (Kovacic instead of Gundogan) seems to be a bit too much when it comes to overall balance. Doku is also a player who doesn't offer control and possession with his dribbling, but immediate impact. It all looks a bit un-Guardiola-like at the moment. Even with Rodri on the pitch, the 4-4 draw at The Bridge is exactly the type of game Guardiola wants to avoid.

Liverpool with 1.93 xG are the best team in the final third. This shouldn't come as a surprise, since Klopp's way to fix things is, more often than not, "more attack" (Liverpool have often achieved an xG above 2). The addition of Szoboszlai in the midfield and that designed box in the midfield, with TAA tucking inside, shows his intentions. Still, Liverpool are a team in the process of rebuilding. Mac Allister is a lot of things, but a designated defensive midfielder who will help maintain a semblance of balance in a side that's set primarily to attack, he is not. Klopp could have used Fabinho/Henderson's experience for just one season. Up front, they look great, with Salah being his usually magnificent self and even Nunez (despite his shortcomings), being the agent of chaos that, unlike Pep, Klopp is far more comfortable with. But having the 4th (Szoboszlai) and the 8th (Mac Allister) most dribbled past players in the PL in your midfield is a problem. Especially when you're chasing City.

Arteta's men seem to be the most "changed" side compared to last season. Their 1.47 xG is significantly worse than last season's 1.83 xG. Their 0.72 xGA is what allows them to stay near the top of the table. They have played a lot of difficult games at home, so things may change in the foreseeable, but nevertheless, it is an impressive statistic at this level of football. And with Timber out with injury. One could argue that Rice's impact is more than obvious. This is what he brings to the table, first and foremost. On the other hand, Arsenal do seem to struggle with Havertz and his role on the pitch. I don't know why we haven't seen him as a false #9 more often. Arteta is still trying to figure out what his best midfield is going forward.

It will be interesting to see which path, Klopp's or Arteta's, will get closer to City. United, on the other hand, don't seem to have a point of reference for our season. Something to start from and then see in which areas we can focus on to see incremental improvements.

On a side note, because the love-in for one or the other component of McFred is still a thing, and it also affects the discussions about what we need in our midfield, i feel it's worth pointing out that the best sides defensively (xGA), the ones that are being threatened less than anybody else, are the two sides (among the big boys) who base their defending around possession and control through passing and movement. Arsenal and City.

Thoroughly interesting read, and good points all.

"The numbers look mediocre because the team is plain bad." There it is, and it's been like that all season. Very disconcerting.
 
On a side note, because the love-in for one or the other component of McFred is still a thing, and it also affects the discussions about what we need in our midfield, i feel it's worth pointing out that the best sides defensively (xGA), the ones that are being threatened less than anybody else, are the two sides (among the big boys) who base their defending around possession and control through passing and movement. Arsenal and City.
The love-in for McTominay is confined to his attacking contributions. He can just about do a job as a DM in the absence of anyone else, but only an idiot would believe that he's actually any good there.

There was a fair bit of rejoicing when we got rid of Fred, but our ability to control games appears to have got worse rather than better since he left.
 
The love-in for McTominay is confined to his attacking contributions. He can just about do a job as a DM in the absence of anyone else, but only an idiot would believe that he's actually any good there.

There was a fair bit of rejoicing when we got rid of Fred, but our ability to control games appears to have got worse rather than better since he left.

The comment had to do more with the state of the midfield as part of a machine that, sadly, seems utterly directionless at the moment. As far as the bigger picture is concerned, McTominay's goal contributions mean very little. The way he approaches the midfield role renders him a problem more than an answer to our deeper issues. Similarly, Fred never was and never will be the solution to anything at this level of football. The fact that we miss a player like him (his energy and tenacity), but can't seem to find room for players who can actually pass the ball, says a lot more about how bad we are and not so much about how useful these two are.

Perhaps i should have included others, so that it doesn't seem like i'm going after the easy targets with my criticism. The cost, both in terms of style and tactics, of getting Bruno's high volume end-product should also be put under a bigger microscope at some point. Or the fact that Casemiro, in a season that he was deemed a crucial signing that helped us massively finish 3rd, was the most dribbled past player in the PL. And that's the same season in which a worse for wear Fabinho was trying to salvage whatever he could from the shower of crap that were Liverpool's defensive transitions.

It was interesting reading Michael Cox's latest opinions nearly a third into the new season. He wrote that in most competitive and high quality matches, where we see an array of modern tactical ideas on display against each other, you can notice three things: Firstly, the defenders and the goalkeepers have become the playmakers burdened with playing the ball out of the back. Secondly, the midfielders' primary function is to receive, protect and progress the ball in very congested areas and under enormous pressure due to the opposition's press. Finally, we seem to have two types of attackers: The finishers and the ones who can make something happen off the dribble. I believe that when you look at it that way (and Cox isn't far from the truth here), you start worrying about the bigger names and the roles. There can only be a shake of the head for run-of-the-mill players like McFred, Maguire, Dalot etc.
 
The comment had to do more with the state of the midfield as part of a machine that, sadly, seems utterly directionless at the moment. As far as the bigger picture is concerned, McTominay's goal contributions mean very little. The way he approaches the midfield role renders him a problem more than an answer to our deeper issues. Similarly, Fred never was and never will be the solution to anything at this level of football. The fact that we miss a player like him (his energy and tenacity), but can't seem to find room for players who can actually pass the ball, says a lot more about how bad we are and not so much about how useful these two are.

Perhaps i should have included others, so that it doesn't seem like i'm going after the easy targets with my criticism. The cost, both in terms of style and tactics, of getting Bruno's high volume end-product should also be put under a bigger microscope at some point. Or the fact that Casemiro, in a season that he was deemed a crucial signing that helped us massively finish 3rd, was the most dribbled past player in the PL. And that's the same season in which a worse for wear Fabinho was trying to salvage whatever he could from the shower of crap that were Liverpool's defensive transitions.

It was interesting reading Michael Cox's latest opinions nearly a third into the new season. He wrote that in most competitive and high quality matches, where we see an array of modern tactical ideas on display against each other, you can notice three things: Firstly, the defenders and the goalkeepers have become the playmakers burdened with playing the ball out of the back. Secondly, the midfielders' primary function is to receive, protect and progress the ball in very congested areas and under enormous pressure due to the opposition's press. Finally, we seem to have two types of attackers: The finishers and the ones who can make something happen off the dribble. I believe that when you look at it that way (and Cox isn't far from the truth here), you start worrying about the bigger names and the roles. There can only be a shake of the head for run-of-the-mill players like McFred, Maguire, Dalot etc.
Indeed. How we’ve managed to assemble a group of players with so little technical ability is a remarkable achievement. But it’s our lack of “proper” midfielders (and I don’t count either Bruno or Mount) that can do what you describe above that’s really been killing us; as much under Ole as now.

It’s all very well moaning about our defenders’ tedious sideways passing, but they need players in front of them that actually want the ball. And then not immediately give it away…
 
https://understat.com/player/8260

According to this Haaland has outperformed his XG for his entire career (apart from the start of this season). Examples like this are why I don't subscribe to the idea that xG is the best method of how good a team is because being clinical (like Haaland) is an actual skill that some teams (like Chelsea) lack. Teams often end the season with points totals very different from what their xG says they should be so having a good xG doesn't mean anything.

It's just one stat. It's no better a metric than possession or shots on target or, most importantly, the score.