| Accuracy of Xg, Xp estimates 10:40 - May 11 with 5708 views | xrayspecs |
So for L1 this season, this approach:
- failed to place the champions in the automatics or top 6
- identified 50% of teams getting automatic promotion
- identified 50% of team making play-off (plus Wednesday who they placed in autos)
It was better at predicting the bottom four.
So a little bit like the forecasts that gave SW 99% chance of top 2, these stats are not that reliable or useful What may be missing in the formula is resilience/mental strength, as Plymouth found a way to win despite not playing well on a regular basis.
|  | | |  |
| Accuracy of Xg, Xp estimates on 11:02 - May 12 with 1602 views | jayessess |
| Accuracy of Xg, Xp estimates on 10:35 - May 12 by TalkingBlues |
Nail on head, the major obstacle for most of us to get over is the what the jargon actually means. The "expected" bit, which in most minds (including my own) suggests some sort of expectation of a future outcome, but that's not the case here and in fact, the "G" shouldn't be referred to as "Goals" at all, because it doesn't relate to goals, but chances, nutty, but true.
All "xG" actually measures, is the quality of a chance, it in no way is intended to predict the number of goals a team will score, or predict an outcome. The team with the higher "xG" prior to kick off, is purely "expected" to make better quality chances in the game based off previous data, but in no way does that suggest an outcome, it in no way is intended to infer they will actually capitalise on those chances and score goals. The "expected" bit, is derived from the mathematical concept of "expected value" to measure the likelihood of an outcome occurring (which in itself, in my mind, suggests some sort or prediction, but apparently that's not the intention at all, !?). The whole concept of this stat appears to be that over time (by the end of the season) there will be a reversion to mean, so you can (and do) get wildly below, or above "xG" from one game to the next, but by the end of the season, this should have smoothed out, which begs the question itself, what's the point of the metric at all during the season?
What's the upshot of all this? The metric would be far better labelled completely differently, as its name implies in our minds, something that the metric is in no way designed to deliver on, neither "expected" or "goals", but merely expected "quality of chances" based on previous data, that has no bearing whatsoever on future outcome.
Basically, it's a nonsense stat that is widely misinterpreted and misused and is utterly pointless IMO, other than possibly to chuck in a spreadsheet to your boss at the end of the season to make it look like you know your job, if you work in data analytics at a football club that is. [Post edited 12 May 2023 10:51]
|
You don't think it's useful to measure the quality of chances a team creates?
As a fan, I think it's interesting to see whether your subjective evaluation of a game is actually born out. You always walk away from football making judgements that go beyond the scoreline - "we deserved to win that", "we had the better chances", "thought 3-0 flattered us a bit", "bit of a smash and grab that", "lucky to get a draw there I thought". We do it for longer periods too - "we've been unlucky of late, deserved to take more points", "I know we're top, but we just aren't playing that well, it's not going to last". xG's a useful metric to measure your hunch against!
I'm sure managers want to measure their hunches in the same way. Is the way we play generating good chances for us to score? Is the way we defend preventing the opposition from making good chances? Of course you can do these things from the eye test and video analysis, but I'm sure having something that measures that at a glance is useful. |  |
| |
| Accuracy of Xg, Xp estimates on 11:08 - May 12 with 1581 views | TalkingBlues |
| Accuracy of Xg, Xp estimates on 11:02 - May 12 by jayessess |
You don't think it's useful to measure the quality of chances a team creates?
As a fan, I think it's interesting to see whether your subjective evaluation of a game is actually born out. You always walk away from football making judgements that go beyond the scoreline - "we deserved to win that", "we had the better chances", "thought 3-0 flattered us a bit", "bit of a smash and grab that", "lucky to get a draw there I thought". We do it for longer periods too - "we've been unlucky of late, deserved to take more points", "I know we're top, but we just aren't playing that well, it's not going to last". xG's a useful metric to measure your hunch against!
I'm sure managers want to measure their hunches in the same way. Is the way we play generating good chances for us to score? Is the way we defend preventing the opposition from making good chances? Of course you can do these things from the eye test and video analysis, but I'm sure having something that measures that at a glance is useful. |
I think it's a massively over engineered and flawed (because it relies on what someone is telling it to look for, which is massively subjective) stat.
As you said, what you see with your eyes on the touchline and in the video analysis of the game, tells you all you need to know about what chances you had and what you could/should have done better. I don't need an algorithm to tell me whether a player should have scored, or not, or if they had a great quality chance to score. [Post edited 12 May 2023 11:10]
|  |
| |
| Accuracy of Xg, Xp estimates on 11:37 - May 12 with 1550 views | jayessess |
| Accuracy of Xg, Xp estimates on 11:08 - May 12 by TalkingBlues |
I think it's a massively over engineered and flawed (because it relies on what someone is telling it to look for, which is massively subjective) stat.
As you said, what you see with your eyes on the touchline and in the video analysis of the game, tells you all you need to know about what chances you had and what you could/should have done better. I don't need an algorithm to tell me whether a player should have scored, or not, or if they had a great quality chance to score. [Post edited 12 May 2023 11:10]
|
But the eye test and video analysis is also incredibly subjective, isn't it? We routinely get threads on here laughing at opposition managers' post-game assessments. | |
| |
| Accuracy of Xg, Xp estimates on 12:05 - May 12 with 1537 views | TalkingBlues |
| Accuracy of Xg, Xp estimates on 11:37 - May 12 by jayessess |
But the eye test and video analysis is also incredibly subjective, isn't it? We routinely get threads on here laughing at opposition managers' post-game assessments. |
Any sport is subjective, because a host of different people are biased one way or another, people are always looking at different aspects of a game to measure its success by. Some of us will think a player had a great game, where others will consider them to have had an average, or poor game, depending on what our bias was at the time, in the same way we likely subconsciously measure how good a ref was, by how many decisions went our way in the game, or weren't given against us.
Some managers are fair and give a very unbiased and honest verdict, but others (possibly thinking about fan backlash, upsetting their players, or getting sacked etc) are far less truthful in their appraisal, but I'd imagine that they are not unaware of the reality, they probably just choose not to air it. Equally, I'd imagine that some of the views ITFC fans have of opposition managers appraisals of games, are also biased. | |
| |
| Accuracy of Xg, Xp estimates on 12:25 - May 12 with 1528 views | NorrisHatter |
| Accuracy of Xg, Xp estimates on 11:08 - May 12 by TalkingBlues |
I think it's a massively over engineered and flawed (because it relies on what someone is telling it to look for, which is massively subjective) stat.
As you said, what you see with your eyes on the touchline and in the video analysis of the game, tells you all you need to know about what chances you had and what you could/should have done better. I don't need an algorithm to tell me whether a player should have scored, or not, or if they had a great quality chance to score. [Post edited 12 May 2023 11:10]
|
That's true to a considerable extent, which is why I think that the high profile that xG has obtained is because it enables every manager / scout to "watch" every other game being played on a given day, without having to attend the match, and then how that's been used by unusually successful clubs.
Its supposedly the technique (enhanced in the way that I mentioned by SmartOdds in a previous post) that has enabled Brentford to rise above the position one would expect of a club of their means by spotting players who are outperforming their stats in lower / foreign leagues and then snapping them up for a bargain price, benefitting from their superior performance and then selling on at a profit.
Of course, there's more to recruitment than that - the player has to fit your culture for instance - a star striker who is an objectionable git may detract more from the team than they add.
That's why football stats (like KPIs in business) draw your attention to important "whats" and comparing them with targets that should reasonably be achieved. The skill of management / analysts is in working out the why and how they should change things (if necessary) as a result. | | | |
| Accuracy of Xg, Xp estimates on 12:34 - May 12 with 1521 views | jayessess |
| Accuracy of Xg, Xp estimates on 12:05 - May 12 by TalkingBlues |
Any sport is subjective, because a host of different people are biased one way or another, people are always looking at different aspects of a game to measure its success by. Some of us will think a player had a great game, where others will consider them to have had an average, or poor game, depending on what our bias was at the time, in the same way we likely subconsciously measure how good a ref was, by how many decisions went our way in the game, or weren't given against us.
Some managers are fair and give a very unbiased and honest verdict, but others (possibly thinking about fan backlash, upsetting their players, or getting sacked etc) are far less truthful in their appraisal, but I'd imagine that they are not unaware of the reality, they probably just choose not to air it. Equally, I'd imagine that some of the views ITFC fans have of opposition managers appraisals of games, are also biased. |
Every assessment an individual makes of a game comes from a particular standpoint, that goes for managers as much as fans. That's why people become interested in trying to measure these things, isn't it? | |
| |
| Accuracy of Xg, Xp estimates on 12:35 - May 12 with 1519 views | TalkingBlues |
| Accuracy of Xg, Xp estimates on 12:25 - May 12 by NorrisHatter |
That's true to a considerable extent, which is why I think that the high profile that xG has obtained is because it enables every manager / scout to "watch" every other game being played on a given day, without having to attend the match, and then how that's been used by unusually successful clubs.
Its supposedly the technique (enhanced in the way that I mentioned by SmartOdds in a previous post) that has enabled Brentford to rise above the position one would expect of a club of their means by spotting players who are outperforming their stats in lower / foreign leagues and then snapping them up for a bargain price, benefitting from their superior performance and then selling on at a profit.
Of course, there's more to recruitment than that - the player has to fit your culture for instance - a star striker who is an objectionable git may detract more from the team than they add.
That's why football stats (like KPIs in business) draw your attention to important "whats" and comparing them with targets that should reasonably be achieved. The skill of management / analysts is in working out the why and how they should change things (if necessary) as a result. |
In terms of targeting individual players, obviously you'd be looking at individual stats rather than the squads and looking for someone that returns unusually high numbers compared to the mean, but if it was that simple, every team in the Country using data analytics would be targeting the same players, foreign and domestic. As you say, there are a multitude of other factors, both data based and those intangibles, such as character, culture etc that need to combined to find the right fit and allow the individual to continue to outperform the metrics, those sort of assessments require human analysis and contact. | |
| |
| Accuracy of Xg, Xp estimates (n/t) on 12:45 - May 12 with 1513 views | TalkingBlues |
| Accuracy of Xg, Xp estimates on 12:34 - May 12 by jayessess |
Every assessment an individual makes of a game comes from a particular standpoint, that goes for managers as much as fans. That's why people become interested in trying to measure these things, isn't it? |
Don't know what happened there, screen went wonky, let's try again!
Different managers look for different qualities to fit their system, so they will focus on different aspects of data and use that in conjunction with what they see with their own eyes to make a list of recruitment targets. The intangibles will come next, interview, human contact etc to determine if they believe the player is the right fit.
The data is certainly very useful for identifying certain areas that a manager may be specifically interested in, like work rate, ground covered, pass completion etc etc, but it is only part of the process. [Post edited 12 May 2023 12:51]
| |
| |
Login to get fewer ads
| Accuracy of Xg, Xp estimates on 13:06 - May 12 with 1494 views | Herbivore |
| Accuracy of Xg, Xp estimates on 11:08 - May 12 by TalkingBlues |
I think it's a massively over engineered and flawed (because it relies on what someone is telling it to look for, which is massively subjective) stat.
As you said, what you see with your eyes on the touchline and in the video analysis of the game, tells you all you need to know about what chances you had and what you could/should have done better. I don't need an algorithm to tell me whether a player should have scored, or not, or if they had a great quality chance to score. [Post edited 12 May 2023 11:10]
|
Having read your posts on here, I think you could do with as much assistance as you can get to help you understand what you're watching. |  |
| |
| Accuracy of Xg, Xp estimates on 13:07 - May 12 with 1495 views | Vegtablue |
| Accuracy of Xg, Xp estimates on 08:14 - May 12 by Herbivore |
It's not a prediction. |
I'm of the same mind (I admit to not reading many posts so will likely be repeating people, apologies).
People may wish to extrapolate from the data. Alternative xG tables have been created, some using a simple summary of xG team performance, others through awarding conventional points on a match-by-match basis. From these further opinions and predictions will be made, but this is all separate from the stats themselves.
'Expected' means 'what should have happened', when one places their faith in the xG supercomputers to assign true goalscoring probabilities to every opportunity. It doesn't say what will or is likely to happen because it's rooted in past information.
I will add I believe xG to be an approximation with a number of shortcomings. It isn't unhelpful, but I wouldn't use it to predict the weather. [Post edited 12 May 2023 13:11]
| | | |
| Accuracy of Xg, Xp estimates on 13:12 - May 12 with 1481 views | TalkingBlues |
| Accuracy of Xg, Xp estimates on 13:06 - May 12 by Herbivore |
Having read your posts on here, I think you could do with as much assistance as you can get to help you understand what you're watching. |
4/10 | |
| |
| Accuracy of Xg, Xp estimates on 13:18 - May 12 with 1463 views | BlueandTruesince82 |
| Accuracy of Xg, Xp estimates on 18:03 - May 11 by clive_baker |
Take your point that they're probably directionally right, but it does beg the question how reliable or meaningful the absolute XG and XP measures are when 21 of the 24 sides have over delivered their XG, some by as many as 20-30 goals. CBA to sum the XG columns on both tables but the actuals must be at least 15% higher than the XG.
When you've got a sample size over the season of 1,100 team performances to pull from it doesn't feel right that there should be such a consistent discrepancy, and would suggest the XG is too stingy. |
Well very becaues it's expected so it analyisys game days and calculates how many goals a team would expect to score
In Plymouths case all those deflections means they massively outperformed their expectation of scoring |  |
| |
| Accuracy of Xg, Xp estimates on 13:44 - May 12 with 1439 views | C_HealyIsAPleasure |
Just wrote a long post on this thread which then disappeared into the ether…FFS
Shorter version - I am a big believer in xG generally as a measure, and it can be predictive to a point. Ipswich in recent years being an example of that, under Lambert we were top but xG suggested we were well above where we should have been (and I remember some posters on here pouring scorn on anyone pointing that out seeing as we were top of the ACTUAL table), and sure enough a regression came. Similarly this season the opposite happened
However no-one has mentioned the biggest flaw in that actual goals change games, and from an xG perspective missing chances that change the game situation can actually lead to more chances and xG accumulating. Hence it’s not as simple as just looking at the numbers and saying a certain team deserved to win |  |
| |
| Accuracy of Xg, Xp estimates on 13:49 - May 12 with 1433 views | xrayspecs |
| Accuracy of Xg, Xp estimates on 08:14 - May 12 by Herbivore |
It's not a prediction. |
Sorry, if I am being slow here, but predicted is a synonym for expected.
As per the source data, “expected points are calculated by simulating each game 10,000 times”.
They are using the data to simulate the result of every game in L1 last season from which they have constructed their own league table. And sadly for PA fans, they missed out on the play-offs again... | | | |
| Accuracy of Xg, Xp estimates on 13:57 - May 12 with 1424 views | jayessess |
| Accuracy of Xg, Xp estimates on 13:49 - May 12 by xrayspecs |
Sorry, if I am being slow here, but predicted is a synonym for expected.
As per the source data, “expected points are calculated by simulating each game 10,000 times”.
They are using the data to simulate the result of every game in L1 last season from which they have constructed their own league table. And sadly for PA fans, they missed out on the play-offs again... |
It's a rating given to a past event though, so it can't be a prediction.
If Luke Woolfenden misses a shot from 30 yards and I say "I wouldn't expect him to score from there", I haven't made a prediction, I've evaluated the difficulty of the opportunity. | |
| |
| Accuracy of Xg, Xp estimates on 14:00 - May 12 with 1420 views | xrayspecs |
| Accuracy of Xg, Xp estimates on 13:44 - May 12 by C_HealyIsAPleasure |
Just wrote a long post on this thread which then disappeared into the ether…FFS
Shorter version - I am a big believer in xG generally as a measure, and it can be predictive to a point. Ipswich in recent years being an example of that, under Lambert we were top but xG suggested we were well above where we should have been (and I remember some posters on here pouring scorn on anyone pointing that out seeing as we were top of the ACTUAL table), and sure enough a regression came. Similarly this season the opposite happened
However no-one has mentioned the biggest flaw in that actual goals change games, and from an xG perspective missing chances that change the game situation can actually lead to more chances and xG accumulating. Hence it’s not as simple as just looking at the numbers and saying a certain team deserved to win |
Thanks. I think we have similar views, I can see the logic but based on the results it generated, there is clearly something not quite right with the algorithm.
On your point about missing chances leading to more chances, could a goalmouth scramble with three or four blocked chances create an Xg for that period of play of greater than one, when only one goal can be scored? | | | |
| Accuracy of Xg, Xp estimates on 14:05 - May 12 with 1408 views | xrayspecs |
| Accuracy of Xg, Xp estimates on 13:57 - May 12 by jayessess |
It's a rating given to a past event though, so it can't be a prediction.
If Luke Woolfenden misses a shot from 30 yards and I say "I wouldn't expect him to score from there", I haven't made a prediction, I've evaluated the difficulty of the opportunity. |
Yes, I get that.
My point is that they have used the data to retrospectively simulate last seasons results and create a league table, which has a few glaring errors.
Back in the things I am happy not to understand file.
Cheers. | | | |
| Accuracy of Xg, Xp estimates on 14:50 - May 12 with 1365 views | TalkingBlues |
| Accuracy of Xg, Xp estimates on 13:18 - May 12 by BlueandTruesince82 |
Well very becaues it's expected so it analyisys game days and calculates how many goals a team would expect to score
In Plymouths case all those deflections means they massively outperformed their expectation of scoring |
This is the problem with the way they’ve labelled this stat, it’s misleading, it’s actually got nothing to do with goals, or how many goals would be expected to be scored, it measures (in a flawed manner) only the quality of chances to score, not goals. | |
| |
| Accuracy of Xg, Xp estimates on 14:51 - May 12 with 1367 views | C_HealyIsAPleasure |
| Accuracy of Xg, Xp estimates on 14:00 - May 12 by xrayspecs |
Thanks. I think we have similar views, I can see the logic but based on the results it generated, there is clearly something not quite right with the algorithm.
On your point about missing chances leading to more chances, could a goalmouth scramble with three or four blocked chances create an Xg for that period of play of greater than one, when only one goal can be scored? |
We don’t have similar views, I think it’s generally a good indicator whereas it’s clear from your posts in this thread you think it’s a load of rubbish and are desperate to poke any hole you can it
For the example you give below the answer is no, scrambles or that nature are aggregated in the score, as I understand it anyway. That isn’t the case for more loosely connected events (eg. A chance which leads to a corner that generates another chance) which is where some of the flaws lie but then over time those things even up to a degree, same as anything else (see also the league table after 1 game) | |
| |
| Accuracy of Xg, Xp estimates on 15:17 - May 12 with 1336 views | Cheltenham_Blue |
Xg and Xp are literally guesses pointless to take them as a direct indication of how well a team will do. |  |
| |
| Accuracy of Xg, Xp estimates on 15:19 - May 12 with 1334 views | Sawtrich |
Yes, I think people are taking the word 'expected' too literally. If a team has an xG of 1.46 that isn't literally expected to happen. It's just a measure of chances created and how good they were. It's just a bit more advanced than saying you'd 'expect' to score once for every nine attempts you have. (I made up nine, no idea what the real figure is). It's just a rough guide to how a game went. Interesting that the two big outliers, Plymouth and Oxford, are teams with a significantly different goal difference to the teams around them in the actual table. There's a lot of things it can't measure, but it's a useful tool. To dismiss it completely is as wrong headed as relying on it completely imo. |  | | |
| Accuracy of Xg, Xp estimates on 16:03 - May 12 with 1301 views | xrayspecs |
| Accuracy of Xg, Xp estimates on 14:51 - May 12 by C_HealyIsAPleasure |
We don’t have similar views, I think it’s generally a good indicator whereas it’s clear from your posts in this thread you think it’s a load of rubbish and are desperate to poke any hole you can it
For the example you give below the answer is no, scrambles or that nature are aggregated in the score, as I understand it anyway. That isn’t the case for more loosely connected events (eg. A chance which leads to a corner that generates another chance) which is where some of the flaws lie but then over time those things even up to a degree, same as anything else (see also the league table after 1 game) |
Neither desperate nor poking holes in it.
And my question was based on your comment about it being flawed.
Ok, I will replace similar views with generally good versus can see the logic but feel it does not capture all of the relevant data. | | | |
| Accuracy of Xg, Xp estimates on 17:25 - May 12 with 1265 views | Vegtablue |
| Accuracy of Xg, Xp estimates on 14:50 - May 12 by TalkingBlues |
This is the problem with the way they’ve labelled this stat, it’s misleading, it’s actually got nothing to do with goals, or how many goals would be expected to be scored, it measures (in a flawed manner) only the quality of chances to score, not goals. |
I think your reasoning's gone wonky!
They are clear it's a metric designed to measure the probability of a shot resulting in a goal. Shots are data points, which are assigned their probabilities of resulting in a goal (defined as 1.0), and a team's xG total is the addition of all these converted data points. If one shot is assigned a probability of 0.8 and another 0.2, the team's xG will be 1.0: the expectation being 1 goal will have resulted.
Distinction between quality of a chance to score and probability of there being a goal may be argued with debatable outcome, but it can't serve to redefine what the analysts have set out to achieve (even if one feels they've failed in their endeavours). | | | |
| Accuracy of Xg, Xp estimates on 19:17 - May 12 with 1235 views | TalkingBlues |
| Accuracy of Xg, Xp estimates on 17:25 - May 12 by Vegtablue |
I think your reasoning's gone wonky!
They are clear it's a metric designed to measure the probability of a shot resulting in a goal. Shots are data points, which are assigned their probabilities of resulting in a goal (defined as 1.0), and a team's xG total is the addition of all these converted data points. If one shot is assigned a probability of 0.8 and another 0.2, the team's xG will be 1.0: the expectation being 1 goal will have resulted.
Distinction between quality of a chance to score and probability of there being a goal may be argued with debatable outcome, but it can't serve to redefine what the analysts have set out to achieve (even if one feels they've failed in their endeavours). |
From what I can see, they claim that there are 2 areas the stat covers, the quality of the chance and thus the likelihood of a goal being scored from that chance. I think this is where they go wrong with the terminology, because the quality of a chance, is for me, everything leading up to the point just prior to the player taking the shot/header/free kick etc and if the stat were used solely for that purpose, I don't have an issue, though I still think there needs to be vastly more metrics taken into consideration even for that.
Once you reach the point where the player strikes the ball, I believe that is a completely different event and you must consider a multitude of different metrics to try and establish the likelihood of scoring, none of which are taken into consideration. i.e. everything is based on an "average player" but what's an average player? they should be using previous data specific to that player in similar scenarios (not data from 1000's of shots from a variety of players, playing in different different positions in that league) establishing if the opportunity fell to their strong, or weak foot, minutes played in the game prior to the opportunity, the speed they were running at when the ball was struck, whether they had they changed direction just before striking the ball, etc etc etc etc, all key metrics, there are many more. I don't believe you can simply apply the same set of metrics to both areas and hope for the best, which is exactly what they do.
To assess the quality of a chance, it's better than nothing, but still woefully short on data, but as a tool to establish the likelihood of a goal being scored, it couldn't be more unsuitable if they tried IMO. [Post edited 12 May 2023 19:22]
| |
| |
| Accuracy of Xg, Xp estimates on 20:55 - May 12 with 1204 views | Vegtablue |
| Accuracy of Xg, Xp estimates on 19:17 - May 12 by TalkingBlues |
From what I can see, they claim that there are 2 areas the stat covers, the quality of the chance and thus the likelihood of a goal being scored from that chance. I think this is where they go wrong with the terminology, because the quality of a chance, is for me, everything leading up to the point just prior to the player taking the shot/header/free kick etc and if the stat were used solely for that purpose, I don't have an issue, though I still think there needs to be vastly more metrics taken into consideration even for that.
Once you reach the point where the player strikes the ball, I believe that is a completely different event and you must consider a multitude of different metrics to try and establish the likelihood of scoring, none of which are taken into consideration. i.e. everything is based on an "average player" but what's an average player? they should be using previous data specific to that player in similar scenarios (not data from 1000's of shots from a variety of players, playing in different different positions in that league) establishing if the opportunity fell to their strong, or weak foot, minutes played in the game prior to the opportunity, the speed they were running at when the ball was struck, whether they had they changed direction just before striking the ball, etc etc etc etc, all key metrics, there are many more. I don't believe you can simply apply the same set of metrics to both areas and hope for the best, which is exactly what they do.
To assess the quality of a chance, it's better than nothing, but still woefully short on data, but as a tool to establish the likelihood of a goal being scored, it couldn't be more unsuitable if they tried IMO. [Post edited 12 May 2023 19:22]
|
Incorporating a lot of your suggestions would be practically impossible and would mostly result in insufficient data sizes to determine xG to a greater level of accuracy than using the "average player", but this doesn't make your criticisms less valid! I believe they've created an algorithm that provides a rough approximation, and there will be plenty of occasions when it's quite far from reflecting what it purports to do.
In respect to the first paragraph though, I still think your terminology argument fails to land. In the event you identify a variable that affects the probability of a goal occurring, I would suggest you seriously consider factoring it into the quality of the goalscoring opportunity. We've established the broad-brush-stroke implementation of xG will always be a challenge to identifying true probabilities, but this in turn then applies to the true quality of the chance. The two are intrinsically linked.
For example, the ball is played to John in the six-yard area, goalkeeper behind him, ball at his feet. High quality chance, high probability goal. You then learn John has been here many times before and averages one goal in three from similar positions. Lower probability goal, still high quality chance for our John and the attacking team? Even worse, he's already walked into the post and has forgotten if he's defending or shooting. But the quality of the chance holds firm? Nein. Both increase and decrease proportionally. Disconnecting the quality of the opportunity from the player in front of it, then reattaching it for the probability of the outcome, serves to only disguise the true quality of the opportunity to score. See Ward-Prowse shaping up for a 25 yard free kick vs. Cristiano Ronaldo / Cameron Burgess if still unconvinced.
So xG doesn't drill down deeply enough to render Match of the Day a waste of time, but it can be very helpful in highlighting trends / outliers / weaknesses / strengths that could otherwise be missed. | | | |
| |
By continuing to use the site, you agree to our use of cookies and to abide by our Terms and Conditions.
