| Apparently, we have low xG 09:37 - Sep 3 with 8081 views | dmm |
I'm a bit ambivalent about xG but this is actually quite an interesting data analysis of our season so far: https://eflanalysis.com/analysis/efl-championship-stats-qpr-goals
It seems to conclude that we're pretty efficient outfit so far, but I must say I did chuckle at this statement: "One of the main reasons for their success is the number of goals scored by the team." Yeah, that's how football kind of works! |  | | |  |
| Apparently, we have low xG on 09:45 - Sep 3 with 5565 views | loftboy |
Lucky then it’s only goals in the onion bag that actually count, Far too much I’ve analyse of stats which is the main reason half the teams play boring tippy tappy passing with no penetration just so they can say “ we lost but we had 69% possessions so didn’t deserve to” |  |
| |
| Apparently, we have low xG on 09:47 - Sep 3 with 5550 views | DWQPR |
Total bow locks if you ask me. I just put our start down to a good set of players, good coaching and a bit of luck. Someone is making a lot of money out of this xG thing. And to see teams like Sheff Utd and Hull ranking highly in some areas and Fulham lowly. Total cobblers. |  |
| |
| Apparently, we have low xG on 09:52 - Sep 3 with 5520 views | BklynRanger |
It wouldn't hurt for the article to include a 2 line explanation/reminder of what an xG is and what it does?
I'm assuming it's some form of super-mojo forcefield that wraps round a player when the power-up symbol above their head reaches 100.
The offensive battles won stat is interesting - think Chair and Willock deserve a lot of credit for thir constant work, as well as the rest of the team obviously. |  | | |
| Apparently, we have low xG on 10:02 - Sep 3 with 5481 views | Wegerles_Stairs |
We are staring relegation from the Justice League in the face. | | | |
| Apparently, we have low xG on 10:02 - Sep 3 with 5472 views | CliveWilsonSaid |
| Apparently, we have low xG on 09:47 - Sep 3 by DWQPR |
Total bow locks if you ask me. I just put our start down to a good set of players, good coaching and a bit of luck. Someone is making a lot of money out of this xG thing. And to see teams like Sheff Utd and Hull ranking highly in some areas and Fulham lowly. Total cobblers. |
Another thing to take into account is the continuity of staff. The managers and coaching staff have been here a while now, as well as a lot of the players. Also many of the new signings are familiar with the manager, coaches or the club. These factors probably won’t show up in stats. I’d imagine that familiarity of staff and systems would bring greater efficiency, from my totally unprofessional viewpoint of course! [Post edited 3 Sep 2021 10:08]
|  |
| |
| Apparently, we have low xG on 10:17 - Sep 3 with 5385 views | RBlock |
| Apparently, we have low xG on 09:52 - Sep 3 by BklynRanger |
It wouldn't hurt for the article to include a 2 line explanation/reminder of what an xG is and what it does?
I'm assuming it's some form of super-mojo forcefield that wraps round a player when the power-up symbol above their head reaches 100.
The offensive battles won stat is interesting - think Chair and Willock deserve a lot of credit for thir constant work, as well as the rest of the team obviously. |
Basically Brooklyn, xG is a number assigned to each shot and represents the likelihood of that shot going in, calculated using previous data around similar shots. So they take into account things like position of the shot, body part, type of build up (rebound, cross, through ball) etc. The idea is it goes beyond the simple 'shots on target' metric, and tries to provide an idea of the quality of the chances a team creates.
So we've scored goals like Chris Willock v Hull (Ingram mistake) and Dickie v Millwall, where shots from that distance tend to have a low probability of going in. So the xG assigned to those shots is low, despite the fact that they both count as a goal. It also means that other teams will rack up xG despite not scoring, when we ride our luck and rely on a super Seny save or Barbet block. Yes, we beat Coventry and Hull, and drew against Barnsley, but there were periods in both of those games where we were under the cosh and did well to stand up to the pressure. So that is why Coventry has 1.69 xG and we only had 1.56 xG, despite us winning 2-0. The important thing, of course, is that we won. That is what matters. |  | | |
| Apparently, we have low xG on 10:24 - Sep 3 with 5349 views | Northernr |
Repeating a question I asked on another thread, does it assign the same value to a 25 yard shot whether it's Rob Dickie or Jordy De Wijs taking it? One's good from 25 yards, one isn't, so one should have a greater value no? Does a 12 yard shot from John Spencer carry the same weighting as a 12 yard shot from Brett Angell? |  | | |
| Apparently, we have low xG on 11:15 - Sep 3 with 5199 views | RBlock |
| Apparently, we have low xG on 10:24 - Sep 3 by Northernr |
Repeating a question I asked on another thread, does it assign the same value to a 25 yard shot whether it's Rob Dickie or Jordy De Wijs taking it? One's good from 25 yards, one isn't, so one should have a greater value no? Does a 12 yard shot from John Spencer carry the same weighting as a 12 yard shot from Brett Angell? |
No, doesn't matter who is taking the shot. Obviously a Rob Dickie 30 yarder and Jordy De Wjis 30 yarder are different prospects altogether. Absolutely get your point, but from a practical standpoint you simply wouldn't have enough data to calculate Dickie's & De Wjis's individual xG from a certain area. I don't think I've ever seen De Wjis take a shot from outside the box. The data set isn't large enough.
What that does mean though, is you can use it to measure a players finishing. Mason Greenwood for instance has consistently outperformed his xG, scoring from very acute angles or difficult chances. So it backs up what you can tell from watching the player; he's a quality finisher. | | | |
Login to get fewer ads
| Apparently, we have low xG on 11:17 - Sep 3 with 5190 views | slmrstid |
xG is for people who have never attended a football game in their lives to get excited about.
For the rest of us who actually attend games we see it for what it is - an absolute load of nonsense not worth paying any attention to. |  | | |
| Apparently, we have low xG on 11:23 - Sep 3 with 5156 views | WadR |
| Apparently, we have low xG on 10:24 - Sep 3 by Northernr |
Repeating a question I asked on another thread, does it assign the same value to a 25 yard shot whether it's Rob Dickie or Jordy De Wijs taking it? One's good from 25 yards, one isn't, so one should have a greater value no? Does a 12 yard shot from John Spencer carry the same weighting as a 12 yard shot from Brett Angell? |
It does not. Which is a positive and a negative. If you look at long term trends for a player and see they're over performing on their xG across the course of a season or longer - it's a solid indicator that they're a better than average finisher.
It has its flaws as a system. For example, I remember seeing an xG graph for our game a couple of seasons ago where Hugill skied that shot over from 2 yards out with an open goal. And it had an xG rating of like 0.5 because the model in question didn't take into account the position of the goalkeeper.
But for long term trends it's generally pretty telling. I think when Bristol City had that great half season a few years ago with the league cup run and were in the top 2, the xG numbers had them in the bottom half. And then unsurprisingly, across the season, their xG performance more or less matched up with their performance as they regressed to the mean.
It's not perfect but it's a better way of measuring chances than shots on target or total shots for example. I think the fact that v successful clubs put so much faith in analytics teams that make signings based off stats like xG is probably a sign that it has merit. [Post edited 3 Sep 2021 11:25]
| | | |
| Apparently, we have low xG on 11:24 - Sep 3 with 5155 views | Northernr |
| Apparently, we have low xG on 11:15 - Sep 3 by RBlock |
No, doesn't matter who is taking the shot. Obviously a Rob Dickie 30 yarder and Jordy De Wjis 30 yarder are different prospects altogether. Absolutely get your point, but from a practical standpoint you simply wouldn't have enough data to calculate Dickie's & De Wjis's individual xG from a certain area. I don't think I've ever seen De Wjis take a shot from outside the box. The data set isn't large enough.
What that does mean though, is you can use it to measure a players finishing. Mason Greenwood for instance has consistently outperformed his xG, scoring from very acute angles or difficult chances. So it backs up what you can tell from watching the player; he's a quality finisher. |
Further harms it as a catch all measurement that some use it for though doesn't it? If it pays no attention to whether your centre forward is John Spencer or Brett Angell, or if your free kick taker is Harry Wilson or Yoann Barbet, that's quite a flaw no? | | | |
| Apparently, we have low xG on 11:25 - Sep 3 with 5143 views | Jeff |
| Apparently, we have low xG on 11:17 - Sep 3 by slmrstid |
xG is for people who have never attended a football game in their lives to get excited about.
For the rest of us who actually attend games we see it for what it is - an absolute load of nonsense not worth paying any attention to. |
But this point of view is almost a form of 'inverse snobbery' about xG - i'm a home and away season ticket holder (as is RBlock above) and neither of us think it's nonsense.
xG on it's own, in isolation, is like a lamppost for a drunk - more leaning on than illumination. and no-one, not me, not RBlock, nor anyone else is suggesting that it's the be all and end all, or that it's more important than the actual 'business of winning football matches' and the end result
but xG can, alongside other metrics, can certainly give indicators about how, and why things happen. |  |
| |
| Apparently, we have low xG on 11:29 - Sep 3 with 5122 views | dannyblue |
The bit in that article that interested me was this hint that we have accurate passing in the final third. I'd like to see more on that. But it chimes with my subjective opinion. Chair and Willock in particualr are great at showing for the ball and offloading it accurately, keeping posession in attack without playing safe. |  | | |
| Apparently, we have low xG on 11:34 - Sep 3 with 5104 views | BazzeR |
https://dataviz.theanalyst.com/season-reviews/2021/?competition_id=10&season
This site has more info and explains what each section means by clicking on the pink box with the explanation mark.
The problem I have with these sites how accurate are they?
Are these the type of sites that clubs like Brentford base their transfer targets on also. [Post edited 3 Sep 2021 11:51]
| | | |
| Apparently, we have low xG on 11:38 - Sep 3 with 5088 views | slmrstid |
| Apparently, we have low xG on 11:25 - Sep 3 by Jeff |
But this point of view is almost a form of 'inverse snobbery' about xG - i'm a home and away season ticket holder (as is RBlock above) and neither of us think it's nonsense.
xG on it's own, in isolation, is like a lamppost for a drunk - more leaning on than illumination. and no-one, not me, not RBlock, nor anyone else is suggesting that it's the be all and end all, or that it's more important than the actual 'business of winning football matches' and the end result
but xG can, alongside other metrics, can certainly give indicators about how, and why things happen. |
To be fair you're probably right its an inverse snobbery but its really something I have no interest in and I dont really care what our xG is - all that matters is the score in games and where you are in the table.
The rest is just noise to me.
And also things like xG are pointless at this stage in the season when so few games have been played. I saw something posted on here a couple of weeks ago about how Championship average attendances were comparing pre-Covid - teams had only played 1 or 2 home games at that point! Compared to 17-18 games pre-Covid. Whats the point in making a big fuss of a graphic like that when you're comparing such daft stats? | | | |
| Apparently, we have low xG on 11:41 - Sep 3 with 5082 views | RBlock |
| Apparently, we have low xG on 11:24 - Sep 3 by Northernr |
Further harms it as a catch all measurement that some use it for though doesn't it? If it pays no attention to whether your centre forward is John Spencer or Brett Angell, or if your free kick taker is Harry Wilson or Yoann Barbet, that's quite a flaw no? |
You know I'm a few years younger than you and have drunk away the functioning parts of my brain, so I'm completely missing the John Spencer and Brett Angell references here, FYI.
No statistic in isolation is perfect. We've seen that with possession stats for a Brendan Rodgers Swansea team as well, where they have 65% possession, spent pissballing about at the back without creating any quality chances (another bugbear of yours, I know). It's not a catch-all, but just a more complete method of measuring the quality of chances. I don't live and die by any statistics, they are simply an imperfect tools to try and reflect a game of football. | | | |
| Apparently, we have low xG on 11:50 - Sep 3 with 5025 views | Northernr |
| Apparently, we have low xG on 11:41 - Sep 3 by RBlock |
You know I'm a few years younger than you and have drunk away the functioning parts of my brain, so I'm completely missing the John Spencer and Brett Angell references here, FYI.
No statistic in isolation is perfect. We've seen that with possession stats for a Brendan Rodgers Swansea team as well, where they have 65% possession, spent pissballing about at the back without creating any quality chances (another bugbear of yours, I know). It's not a catch-all, but just a more complete method of measuring the quality of chances. I don't live and die by any statistics, they are simply an imperfect tools to try and reflect a game of football. |
Rub it in there 
John Spencer's probably the sharpest finisher I've seen at QPR, and Brett Angell probably the worst. If I've got it right, xG places all its value in where abouts the shot is taken from and what sort of a shot it is, and pays absolutely no attention to who's taking it. So a shot from 15 yards by Charlie Austin carries the same xG rating as a shot from 15 yards by Seny Dieng.
If that's true, that's a megaflaw. Am I being stupid here? | | | |
| Apparently, we have low xG on 11:52 - Sep 3 with 5010 views | eastside_r |
And another thing, it should be EG not xG. Can’t these people spell? |  | | |
| Apparently, we have low xG on 11:57 - Sep 3 with 4983 views | PlanetHonneywood |
| Apparently, we have low xG on 11:50 - Sep 3 by Northernr |
Rub it in there
John Spencer's probably the sharpest finisher I've seen at QPR, and Brett Angell probably the worst. If I've got it right, xG places all its value in where abouts the shot is taken from and what sort of a shot it is, and pays absolutely no attention to who's taking it. So a shot from 15 yards by Charlie Austin carries the same xG rating as a shot from 15 yards by Seny Dieng.
If that's true, that's a megaflaw. Am I being stupid here? |
If Seny Dieng is notching from 15 yards, we either in trouble or just taking the phish! |  |
| |
| Apparently, we have low xG on 12:05 - Sep 3 with 4950 views | RBlock |
| Apparently, we have low xG on 11:52 - Sep 3 by eastside_r |
And another thing, it should be EG not xG. Can’t these people spell? |
The Romans got there first with EG, so the football data lot had to find something else. | | | |
| Apparently, we have low xG on 12:08 - Sep 3 with 4936 views | HanwellHoopster |
| Apparently, we have low xG on 11:50 - Sep 3 by Northernr |
Rub it in there
John Spencer's probably the sharpest finisher I've seen at QPR, and Brett Angell probably the worst. If I've got it right, xG places all its value in where abouts the shot is taken from and what sort of a shot it is, and pays absolutely no attention to who's taking it. So a shot from 15 yards by Charlie Austin carries the same xG rating as a shot from 15 yards by Seny Dieng.
If that's true, that's a megaflaw. Am I being stupid here? |
It's a megaflaw if you're using xG to compare two isolated chances, one of which was from your main striker, and another from a GK.
I guess the best use of xG is across a whole season across multiple chances. So if a whole team is outperforming it's xG across the first half of a season (where, on average, most chances will have fallen to attacking players because the sample size is big enough, with some natural outliers), you can start to draw conclusions. One conclusion might be that the team has very clinical finishers - another that they've been a bit lucky and the team will regress to the mean in the second half of the season. That's where you need to also use your eyes to judge how to best interpret that data.
As others have highlighted, across a whole season you can judge a player's ability to finish by if they outperform or underperform the xG of their chances - assuming the sample size is big enough, which it should be for Austin/Dykes/Willock et al but might not be for JDW.
Edit - I think it's quite helpful in judging players who are over performing in poorly performing sides. Someone getting 10 goals in a team with a very low xG might be a better striker than someone getting 20 goals in a team with triple the xG. I try to use it to dominate my fantasy football league and it has served me fairly well to date. [Post edited 3 Sep 2021 12:10]
| | | |
| Apparently, we have low xG on 12:13 - Sep 3 with 4903 views | RBlock |
| Apparently, we have low xG on 11:50 - Sep 3 by Northernr |
Rub it in there
John Spencer's probably the sharpest finisher I've seen at QPR, and Brett Angell probably the worst. If I've got it right, xG places all its value in where abouts the shot is taken from and what sort of a shot it is, and pays absolutely no attention to who's taking it. So a shot from 15 yards by Charlie Austin carries the same xG rating as a shot from 15 yards by Seny Dieng.
If that's true, that's a megaflaw. Am I being stupid here? |
It is a flaw if you want to look at xG in isolation and base your assessment entirely on that, yes. But as WadR puts it above, it can also be a positive as it allows us to compare a forwards goals compared to his xG, and infer that a forward regularly overperforming their xG is an above average finisher.
It's a tool, and its utility depends on how you use it. My younger brother (a professional gambler, who watches more football than anyone I know, regardless of league, level, country) places a lot of stock in it, when choosing who to back, when to trade out a position, or when to re-back, to find his edge on the bookies. It's not perfect, but it's a much more accurate metric than using possession stats or straight shots on target. | | | |
| Apparently, we have low xG on 12:17 - Sep 3 with 4886 views | Match82 |
| Apparently, we have low xG on 10:24 - Sep 3 by Northernr |
Repeating a question I asked on another thread, does it assign the same value to a 25 yard shot whether it's Rob Dickie or Jordy De Wijs taking it? One's good from 25 yards, one isn't, so one should have a greater value no? Does a 12 yard shot from John Spencer carry the same weighting as a 12 yard shot from Brett Angell? |
I THINK and could be wrong, that xG is meant to represent the degree of success teams have getting into positions which give them a chance to score. It doesn't actually calculate how likely they are to score from there.
So in some ways it's actually more of a metric to see how well your players / team are doing vs expectations rather than whether you are "lucky" and will regress to the mean. If so, then the flaw here is the take of those using the data to draw those conclusions such as someone saying "qpr can't keep this up". A better conclusion would be "qpr have players which are exceeding their xG and as therefore have a team of John Spencers rather than a team of Brett Angells.
Not saying that we are the best team in the division because the xG itself will tell you that we haven't been as good at creating some of those chances.
Warning: all of the above could be bllocks |  | | |
| Apparently, we have low xG on 12:20 - Sep 3 with 4880 views | mobileDeWijs |
"A lot of the football world will have at least heard of xG by now. Put simply, xG is a way to measure the likelihood of a shot becoming a goal.
Not all shots are equal in their quality – one shot might be a speculative 40-yarder and another might be a two-yard tap-in. Therefore, xG measures the quality of each shot before the player shoots, taking into account many factors, including:
The shot angle
The distance from goal
Whether it was with the head, or with the weaker/stronger foot
Whether it was from a cross, through ball, short pass etc
Whether there were multiple defenders in the way
It is important to note that different data providers have slight differences in the factors they consider to go into their xG model. Nevertheless, the xG value is always presented as a number between zero (no chance of a goal) and one (a certain goal)."
From Tom Worville on the athletic. Another poster mentioned we have had a bit of luck - thats exactly it:thaat luck has meant chances which might not have been converted last season (Dykes drought, Bonne, etc) have been this year.
xG only puts a numerical value on something people have always used to describe football - chances. I look forward to seeing this forum's reaction to when xT becomes more commonly used | | | |
| Apparently, we have low xG on 12:25 - Sep 3 with 4853 views | eastside_r |
| Apparently, we have low xG on 12:20 - Sep 3 by mobileDeWijs |
"A lot of the football world will have at least heard of xG by now. Put simply, xG is a way to measure the likelihood of a shot becoming a goal.
Not all shots are equal in their quality – one shot might be a speculative 40-yarder and another might be a two-yard tap-in. Therefore, xG measures the quality of each shot before the player shoots, taking into account many factors, including:
The shot angle
The distance from goal
Whether it was with the head, or with the weaker/stronger foot
Whether it was from a cross, through ball, short pass etc
Whether there were multiple defenders in the way
It is important to note that different data providers have slight differences in the factors they consider to go into their xG model. Nevertheless, the xG value is always presented as a number between zero (no chance of a goal) and one (a certain goal)."
From Tom Worville on the athletic. Another poster mentioned we have had a bit of luck - thats exactly it:thaat luck has meant chances which might not have been converted last season (Dykes drought, Bonne, etc) have been this year.
xG only puts a numerical value on something people have always used to describe football - chances. I look forward to seeing this forum's reaction to when xT becomes more commonly used |
I look forward to xTC - really enjoy his posts. | | | |
| |
By continuing to use the site, you agree to our use of cookies and to abide by our Terms and Conditions.
