xg is not a Liverpool fan creation btw, some here seem to be under that impression.
xG Goal Expectancy Rate (a mentalist RAWK statistical spinoff thread) ft Babu of RAWK fame
- Thread starter Harold_Giles
- Start date
It's not devoid of meaning but doesn't seem useful beyond «quantifying» the bleeding obvious.
You certainly can't use it to argue that the team «should have» won. If you keep creating clear cut chances without capitalizing accordingly, the absence of a consistently good (enough) finisher is the obvious explanation - unless you're going to blame it on bad luck or phenomenal goal keeping (which gets absurd fast). If you lack good (enough) finishing, you've no business claiming that you »should have» won.
Similarly, if you keep shipping soft goals, you probably need better, more focused defenders and/or a better keeper - unless that too is just bad luck (not bloody likely if it's a pattern).
You certainly can't use it to argue that the team «should have» won. If you keep creating clear cut chances without capitalizing accordingly, the absence of a consistently good (enough) finisher is the obvious explanation - unless you're going to blame it on bad luck or phenomenal goal keeping (which gets absurd fast). If you lack good (enough) finishing, you've no business claiming that you »should have» won.
Similarly, if you keep shipping soft goals, you probably need better, more focused defenders and/or a better keeper - unless that too is just bad luck (not bloody likely if it's a pattern).
I do see the merit of this type of decomposition, but I don't understand how you score the chances. Penalty is easy enough, but I reckon different analysts score less clear situations differently. Which makes it problematic to analyze, since part of the variation could be from measurement.
There needs to be a very clear system in place to ensure each equal situation across games is scored equal.
There needs to be a very clear system in place to ensure each equal situation across games is scored equal.
Are you serious with the thread title? xG is not a RAWK invention, you're just making yourself look stupid.
Expected goals is a method for estimating the quality of chances that a football team creates or concedes in a match. It may take a lot of data crunching to create specific xG values, but the underlying idea makes football sense. How many good chances did a team create? How many half-chances? Just how "good" were they? How many good chances did they concede, and so on? These are intuitive football questions.
When you're following a match, you're watching for the creation of chances, getting excited when it appears a scoring chance might be conjured for your team or getting worried when the other team is building one. We all watch for something like "expected goals." Managers and players create tactics aimed at creating good chances and preventing them for their opponents.
The current method includes factors such as distance to goal, angle to goal, the type of pass that assisted the shot (cross / throughball / etc), the type of shot (header / foot / other body part), the type of play that leads to the shot (set play / counterattack / established possession) as well as a variety of other factors.
Here's some literature to help you come up to speed:
- http://www.optasportspro.com/about/optapro-blog/posts/2017/blog-expected-goals-explained/
- https://deepxg.com/2015/10/21/anatomy-of-a-shot/
- http://11tegen11.net/2014/02/10/what-is-expg/
- http://www.telegraph.co.uk/football...does-show-man-city-should-win-premier-league/
- https://www.pinnacle.com/en/betting...ted-goals-for-soccer-matches/JESJH3RKXCMUF9TY
- https://www.mlssoccer.com/glossary/expected-goals
- https://analyticsfc.co.uk/2016/09/02/on-the-lexicon-of-expected-goals/
It's an estimate for the number of goals a team would have scored in a game if their finishing and opponent's goalkeeping was average. A team with goals scored consistently below xG probably has issues with their finishing, while a team with goals conceded consistently below opponent's xG probably has an above average keeper. Pickford and Hart are the worst faring keepers at the moment by this metric, with De Gea among the best. Only 7 games though so sample size etc.
It shouldn't really reassure fans if their team is 'deserving' of more goals than you're scoring, because it means you're shit in front of goal. Like United were for much of last year.
xG, despite the daft name, is a decent stat. Most stats, like shots on target or tackles made, don’t factor in context.
Lots of shots on target for instance can mean you hammered a team, or it can mean you got desperate and resorted to pot shots from distance that the keeper caught without effort. At least xG tries to differentiate between the two.
xG is just a way of counting chances, while also considering how good those chances were. The name sure is sucky though.
Lots of shots on target for instance can mean you hammered a team, or it can mean you got desperate and resorted to pot shots from distance that the keeper caught without effort. At least xG tries to differentiate between the two.
xG is just a way of counting chances, while also considering how good those chances were. The name sure is sucky though.
I have no real problem with using 'expected goals' as a metric to say maybe you deserve more from a game than you got - however I can see one big problem with it and that is that it doesn't factor in the player the chance fell to
For example, statistically a player may have a 30% chance of scoring a goal but if that player is Gary Neville then surely that drops to nearer 5-10% whereas if it's RvN it must surely jump up to around 50%?
When something happens over and over again you can't blame bad luck you have to analyse further. My theory would be that Liverpool have flooded their attack with players you would likely class more as wingers or AMCs than central strikers and so obviously it stands to reason non-specialists will require more chances
Also, it would be interesting to see Data pulled for just the last 10yrs of PL football because I listened to an interesting theory recently about GKs and the change in their physique since the 90s. The average GK is a few inches taller and much more athletic now than his counterpart in the 90s. It seems logical that if keepers have improved then it must be slightly harder to score goals?
For example, statistically a player may have a 30% chance of scoring a goal but if that player is Gary Neville then surely that drops to nearer 5-10% whereas if it's RvN it must surely jump up to around 50%?
When something happens over and over again you can't blame bad luck you have to analyse further. My theory would be that Liverpool have flooded their attack with players you would likely class more as wingers or AMCs than central strikers and so obviously it stands to reason non-specialists will require more chances
Also, it would be interesting to see Data pulled for just the last 10yrs of PL football because I listened to an interesting theory recently about GKs and the change in their physique since the 90s. The average GK is a few inches taller and much more athletic now than his counterpart in the 90s. It seems logical that if keepers have improved then it must be slightly harder to score goals?
The predictions in this article are certainly playing out so far. Although most of us reached the same conclusions from watching games without any scientific data study. It predicts that Chelsea should struggle without Morata but again most would have predicted that anyway.
http://www.telegraph.co.uk/football...does-show-man-city-should-win-premier-league/
Is this where the Premier League gets the 'big chances missed' stat from, players squandering chances with high xG values?
Wait, people are unaware of the usage of xG? 
That's a very good point.
You can conclude lot of things using stats but end of the day finishing and having composure infront of goal is also biggest skill in the game. So if a team is repeatedly not scoring as many expected then probably they aren't creating clear chances or lacks composure.
The stats seem to say that Klopp's system is like Peps in that it works but needs players that are exceptional or all peaking at the same time. The chances will be created in abundance but that leads to the back line being exposed.
Again, these are conclusions that you would reach by watching the side without any study.
In a specific type of situation, yes - that stands to reason.
Overall, however, it might be that better (taller and more athletic) keepers go hand in hand with defenders who - if not worse per se - tend to defend differently, partly because of general trends, partly because officials are stricter (less physicality, less contact allowed, etc.).
And if «differently» equals «better conditions for attackers», the keeper thing may make no practical difference.
I think it is an interesting measure. I do expect big top strikers to perform way over the expected goals for other players. Also some players will score more of a certain types of goals. For a penalty expert the chances of scoring will be higher compared to a worse player. The goalkeeper you play against need to be factored in as well. Although individual stats would have too high variance to really use to give a more accurate view of things.
Penalties is one area you could use it I guess. Say if player X have scored 13/20 penalties. Then you could say his expected goal for the next one would be 65% based on his own stats.
Penalties is one area you could use it I guess. Say if player X have scored 13/20 penalties. Then you could say his expected goal for the next one would be 65% based on his own stats.
Do Opta ever show what the standard deviation of their xG scores? Given that +/-1SD events are actually quite common, it doesn't really mean much if for example a team gets 2 xG, if the standard deviation is close to 1?
I agree. These stats are interesting but don't think we can draw many conclusions as stats leave some important factors.
I can fully see the uses but as said by many much of the analysis seems to conclude similarly to what any halfway perceptive fan could tell you anyway. Last season our xG was much higher than it 'should' have been given the actual goals we scored, which means we were missing too many chances, which any non-xG geek could have told you.
That said it definitely has its uses so, barring the laughs you can have at Rawkites trying to pretend a high xG without actual goals means they somehow deserved to win a match, there's not a lot of madness here.
That said it definitely has its uses so, barring the laughs you can have at Rawkites trying to pretend a high xG without actual goals means they somehow deserved to win a match, there's not a lot of madness here.
If anyone is interested then The Game podcast this week had Duncan Alexander from Opts explaining all this plus what goes on behind the scenes at a place like Opta with regards to collection and interpretation of all the football stats.
Maybe it is Klopp...
Baby Yogi hurts my head 
All the stats in football now are preponderous. Why can't we just watch a game and say he's shite, he's good like we used to.
All the stats in football now are preponderous. Why can't we just watch a game and say he's shite, he's good like we used to.
It seems to me they have miscounted the ox-G ratio
coat
coat
A brief video explanation:
I'm sorry but this thread is pretty embarassing. xG, as some already pointed out is a widely accepted statistical model and if you can put aside your dislike for Liverpool the guy quoted in the OP actually makes some interesting points. At least compared to insights like "he's fecking shit"...
Have a link?
Pretty much, I've read some of his stuff and don't get me wrong he's a condescending prick in the vein of PoP but it does give you food for thought.
is that anything to do with Stevie xG?
xG allows for a more quantitative and accurate description of what's going on than a bunch of drunk fans going, "yeah their attack is shit".
What a load of garbage. Klopp's system is just trash.
Good stuff, genius.
Read the thread title yourself, before you start calling other people stupid.
It's a spinoff by some mod, if that clears things up for you.
Here's my problem, this stuff has absolutely zero predictive power.
The chances a team creates in a match (xG, or whatever) tell you only about that match. Not the match to come, or the match before it. Also, it's all interpretive.
I think people often over complicate football.
You're suggesting that nothing a player or team does in the past has any connection with how they perform in the future?
We need a Liverpool specific formula for working out the probability of there being an hilarious Liverpool result.
It will need to include high loading factors such as
1. Wind strength
2. Percentage of the population of town or city of opposition team defending in oppositions box
3. Recency and frequency of negative things written about 'the boys' as a function of how many of 'the boys' read them
4. Number of offside goals scored by Manchester United
5. Angle of Klopp's Jaw
6. Velocity and Angle of Klopp's glasses as they fly from his head during a rant
7. Volume of saliva collected from the surface of the fourth official's face after Klopp has screamed at it from a distance perpendicular to the depth of the rut in the pitch caused by Henderson's gormless jaw dragging along the floor as he runs.
8. The quantum defensive paradox constant Mg + Kla + Lov = LOL
It will need to include high loading factors such as
1. Wind strength
2. Percentage of the population of town or city of opposition team defending in oppositions box
3. Recency and frequency of negative things written about 'the boys' as a function of how many of 'the boys' read them
4. Number of offside goals scored by Manchester United
5. Angle of Klopp's Jaw
6. Velocity and Angle of Klopp's glasses as they fly from his head during a rant
7. Volume of saliva collected from the surface of the fourth official's face after Klopp has screamed at it from a distance perpendicular to the depth of the rut in the pitch caused by Henderson's gormless jaw dragging along the floor as he runs.
8. The quantum defensive paradox constant Mg + Kla + Lov = LOL
No. I'm saying that a team having loads of chances in one match doesn't mean that they'll have loads of chances in the next.
A win = 3 points.
Draw = 1 point.
Loss = 0 points.
The only statistics that actually matter (goal difference too, and tangible things which have an actual impact on the table). I think this is a lot of wasted brain power.
So when City beat Watford 6-0, you really think that gives you no clue whether its City or Watford who will make more chances in the next game?
That's called form.
And no, it doesn't. I've seen teams come off the back of great form and play shit and lose without creating anything. That can happen in any game despite xG predictive models. Which means the whole thing is empirically suspect.
Also, those are actual goals. Not expected goals.
That applies to any statistic though, and it's not a claimed application. A team with xG say 3.0 average with a 95% interval of 2-4 have a higher likelihood of creating more chances next game compared to a team with 2 average and 1-3 CI, given a sufficient sample of games.
That's not zero predictive power.
The obvious reason for using statistics in football is firstly the amount of information available increase, secondly you dont have to watch the game. The latter can be of big importance, e.g. when scouting. Although not that relevant for a fan
A better comparison would be a game where Liverpool lose 3-0 but have an insanely high xG number. Meaning that they actually outplayed the other team, or that they weren't as bad as the scoreline suggests. Yet that is also irrelevant as it has no bearing on whether or not they'll score any goals in their next match against different opposition.
If Liverpool lose 4-0 but create chances which might have put them 4-0 ahead on another day, would that mean that they have a better chance of winning their next game than the team that just beat them? I seriously doubt it. Football doesn't work like that. There are elements of luck involved and factors that aren't quantifiable (mood, spirit, confidence, etc).
People can do all the simulations they like but it doesn't work as a predictive model. It can be interesting alongside actual data, but is metaphysical happenstance/irrelevancy by itself.
If Liverpool lose 4-0 but create chances which might have put them 4-0 ahead on another day, would that mean that they have a better chance of winning their next game than the team that just beat them? I seriously doubt it. Football doesn't work like that. There are elements of luck involved and factors that aren't quantifiable (mood, spirit, confidence, etc).
People can do all the simulations they like but it doesn't work as a predictive model. It can be interesting alongside actual data, but is metaphysical happenstance/irrelevancy by itself.
Of course, but what really counts is goals for and goals against. Not potential goals for and goals against that statistically were flukes. We enter the realm of the insane when we start making judgements about actual things that happened and whether or not they're a true representation of statistic normality. The fact is that you either score or don't. If you score lots of goals, you have a better chance of scoring lots of goals in the immediate future. If you almost score a goal, that has no bearing on anything. It's just a team that can't finish but might "click" if some non-tangible elements align.
It smacks of amateur poker players using HUD statistics and applying them to scenarios that aren't as simple to quantify.