FBref have it 2.1 - 1.1. xG philosophy use a cruder model.
xG under Ralf
- Thread starter Pogue Mahone
- Start date
we have played against low and mid range teams and couldnt win the game. what will happen against AM ?
Seriously? Is it just me that cringed massively when I thought he quoted xG stats? Fair play if it is, maybe I'm not in tune with the majority.
Maybe you could look at it behind the scenes and see if it's a useful stat (think Big Sam and ProZone), but surely there's only one important stat you'd mention with a camera rolling, it's the one you get 0, 1 or 3 points for. You surely explain the result and performance, if you had better chances and missed them then you just say that, given football can never be objectively reviewed. Find it so bizarre.
Oblak is hopeless at penalties, so there's that
Isn’t it just a modern way of saying ‘we deserved to win the game’?
Instead of feelings though it has a statistical weight to it.
Unfortunately our keepers don’t save them either!
I just personally don't think statistics add much weight at all really, feels all very American. Feelings about whether you deserved to win when you didn't are always going to be subjective, and the stats themselves have subjectivity in their basis. Both are pretty meaningless if you're in the business of winning things. Can't see anything gained by quoting xG over just saying "we've been punished for not taking our chances, etc", reminds me of Liverpool when they couldn't stop hitting the woodwork.
Is xg based on the finishing of an average premier league team?
The eye test (or old man shakes fist at cloud) crowd don't like it.
Maybe, just maybe, they have done that and determined it is a useful stat.
Why would you cringe, surely there's no one left who doesn't get the value of xG and generally advanced metrics?
We currently have a problem with finishing, that is very evident, whereas previously our strikers (for all the scorn we put on them) generally have outperformed their xG. Ragnick is not the solution to our long term aim of winning trophies but he's very clearly putting some kind of basic template in place and, more importantly, is our sacrificial lamb to get through this season and oversee some huge names leaving which will give the new manager a much better start.
If managers had to adhere to that, post match interviews would be short: "Ralf, what do you make of tonight's performance and do you think the result is a fair reflection of the game?" - "We drew the game" - "Thank you for your time"
Not that I'd mind.
I can understand the slight cringe at the mention of xG, but if you can get past that, saying "we underperformed in relation to our xG" is just the same as saying "we had better chances and missed them", no reason why you'd have to consider it as a worse thing to say unless you consider anti-intellectualism an important trait in the public persona of football people.
I'm less worried at the moment. The trend under Ole was our stats were poor but we were still getting results. Once it started to average out we were in trouble.
It's the opposite now. The stats say we're doing things right, but we aren't getting the luck. It will average out again.
It's the opposite now. The stats say we're doing things right, but we aren't getting the luck. It will average out again.
It may not be normal (yet) to actually quote xG - but Ralf is very open when dealing with the media, he seems to communicate with journos in a manner comparable to how he holds lectures at coaching conventions. Which is - well, I'm sure we can debate how wise that is.
However, the point he actually makes isn't anything managers/coaches haven't been making for a long time (only, they obviously didn't use xG terminology back in the day): look at the chances we created, not the goals we scored. This really isn't anything new - a certain type of manager (not all managers, but the more "cerebral" types) has been making similar remarks for decades.
However, the point he actually makes isn't anything managers/coaches haven't been making for a long time (only, they obviously didn't use xG terminology back in the day): look at the chances we created, not the goals we scored. This really isn't anything new - a certain type of manager (not all managers, but the more "cerebral" types) has been making similar remarks for decades.
These xG stats are like all statistics based on averages and expected results. What they don't take into account is anomalies like braindead defenders and strikers who couldn't hit a barn door. Until we fix those problems we are not going to 'return to the norm'.
Under Ole the averages and expected results were poor, but we had individuals pulling rabbits out of the hat. We signed some players (a player?) who upset that and we returned to the norm. What can Ralf do to remove the anomalies from the current team?
Under Ole the averages and expected results were poor, but we had individuals pulling rabbits out of the hat. We signed some players (a player?) who upset that and we returned to the norm. What can Ralf do to remove the anomalies from the current team?
Going by recent trends, our players are actually quite good at scoring in the shootout. Its the other end which is the problem.
Exactly how I feel about this, and why I still have us as favorites for the fourth position.
That's not really true, is it? Good xG models also consider the position of the keeper and of other players. Also if it was controlled possession or a rebound, with which body part you took the shot, etc.
https://fbref.com/en/expected-goals-model-explained/ said:
I wouldn't go that far to all us favourites. We have improvements in our play but there's generally not a huge difference between us, Spurs or Arsenal. And we've got some pretty big issues from now until the end of the season, notably with the striker position. It'll be hard to get enough consistency in the league when our strikers are looking as ancient and useless as Ronaldo and Cavani.
I'm optimistic long term even if Ralf stays (ideally Ten Hag comes in), but I don't think we get top 4 next season. Not the end of the world though.
This. Not saying we'll end up with the same final product but what we are doing now is similar to what Klopp and Liverpool was doing when he was first brought in. Team trying to press without having the full fitness levels to do it, dominating games for periods but then fading, and not getting the results that most of the performances deserved.
I'm much more encouraged by this process compared to us scraping 3 points each week under Ole with moments of magic and sketchy performances with no real desired way to play.
That's why I still like Real Madrid model. We will beat you because we are better than you both individually and as a team. Rest of all is just BS.
It looks at thousands of similar chances across whatever group they look at (big leagues? All of Europe? Any and all leagues?) And take the percentage that it ends in a goal. So basically yeah, the average players finishing ability when presented a chance. Over time it's been found that there isn't much to be found from xG overperformance other than standard variance, considering there is barely a difference in xG overperformance between greats like Lewandowski and Ronaldo compared to average random strikers like Chris Wood or Callum Wilson. What matters more is the frequency of players getting in chances/frequency of team creating chances, as that stuff is more consistent compared to xG overperformance which inevitably comes down.
It's just analysis. If they ask him to analyse the game, that's what he'll do. Any analysis that looks at goals as the be all and end all is literally useless.
Think that's kind of narrow minded. Humans aren't machines, there's always a margin of error and as long as there is, the outcome of games will be reliant on luck.
You only have to look at penalties. Theoretically, a professional player should be able to shoot every single one in a manner that the keeper can't get at the end of it. It's reproducable and trainable. Still even the very best takers miss one in ten. The same goes for chance conversion. Sometimes you lose or draw games that you should never draw because you were on the wrong side of the standard distribution while your opponent landed some lucky punches. That's how it is. Football is about results but the best you can do is maximizing probabilities to win.
And because it is a low scoring game it is much more likely to get a surprise win in football then for example in basketball.
Maybe that is part of the reason why statistics aren't that popular among football fans, due to the higher probability that a match ends in a different way than some stats predict they appear to be less valuable compared to other sports.
I'd say looking at goals is the gold standard in terms of evaluating success. But to limit yourself only to goals will leave a lot of potential on the table. So why not making use of everything that is possible? Goals do not cover all aspects of the game (that can be measured) even if they cover the the 2nd most important ones. The most important ones being the result in itself. Looking at advanced metrics as tools to identify strenghts and weaknesses or to prioritize what to engage first. Obviously in the wrong hands they can be used to screw you just like the famous eye test. Blame the user not the tool.
Correct me if I’m wrong but are you saying that there’s hardly any difference from Ronaldo burying the same chance as Chris wood?
Exactly
In basket ball, a buzzer beater makes up a tiny fraction of the points tallye. In football, a last minute goal can turn the complete match on its head. IMO that's also why all the "clutch" and "choking" mindset transitioning from American sports to football is misplaced. Football doesn't work that way.Statistically, yep.
Since 2014/15 in the big 5 leagues:
- Ronaldo - 221.21 xG, 223 goals - 101%
- Lewandowski - 228.25 xG, 229 goals - 100%
- Chris Wood - 53.6 xG, 50 goals - 93%
- Callum Wilson - 59.91 xG, 59 goals - 98%
- Olivier Giroud - 61.62 xG, 70 goals - 114%
- Lacazette - 113.74 xG, 129 xG - 113%
- Raul Jimenez (not counting stats in Portugal) - 41.1 xG, 40 goals - 97%
- Danny Ings , 52.55 xG, 59 goals - 112%
- Martial - 57.63 xG, 65 goals - 113%
- Rashford - 58.15 xG, 59 goals - 101%
- Ibrahimović - 97.12 xG, 107 goals - 110%
- Messi - 197.74 xG, 233 goals - 118%
- Werner - 99.68 xG, 94 goals - 94%
- Michail Antonio - 60.69 xG, 54 goals - 89%
- Morata - 84.43 xG, 80 goals - 95%
I’m already a fan of xG but just wanted to thank you for posting this. Is there anything similar for teams that shows xG over/under
This should make it obvious for everyone that it’s better in the long run (in a rebuilding phase) to consistently create more xG regardless of current performance than consistently outperforming it. Ie, our situation right now is so much better to build from than what we had under Ole, especially since our ability to create chances is a direct effect of tactical changes implemented by the manager.
Also, the accuracy with which xG/G performance has been able to predict team trajectories time and time again should prove to people that it’s useful (Chelsea dropping out of title race this season, Leicester dropping down the table last season, Brighton doing much better this season) if interpreted correctly. The caveat obviously being that you don’t know exactly when the regression to the mean is coming, just that it is coming, so obviously you can’t predict next week’s results with 100% accuracy.
Just took the data from understat.com, they have every season since 14/15. They aren't necessarily as good of an xG model as some others (not sure whose data they use), but i like how they collect it so it works anyway. Can play around with the site easily.
Well except that you can predict future success by analysing the stats.
So as an example; the best team in the league has the best xG difference generally. So if our xG is trending in the right direction it supports the idea that we are doing the right things but getting unlucky and we just need to stay the course.
What you wouldn’t want to do is make a knee jerk reaction and change formations or personnel based on results when the underlying performance has improved or whatnot.
Like anything though stats have value but should never be used in isolation when you also have subjective information to use. I wouldn’t want to have any manager who relies on one over the other when you can simply use all the tools at your disposal and that includes your own eyes.
P.s - one of the best stats I’ve learnt in recent years is that the title winners often score the most set piece goals in a season. Obviously it’s part of having lots of possession etc but I can’t help but wonder how 5-6 goals from corners this season could have turned draws into wins and losses into draws. If we had got a second goal from a corner or free kick in recent games we’d likely end up winning 3-0 instead of drawling 1-1.
Ralf & Co. will be replaced by the end of the season. If we don’t get top 4, it will be a disaster financially and from a squad building perspective.
I don’t think that reading of Martial is complete. He scores a lot of lower probability goals but he has little of the other attributes of a strikers game. He’s the opposite of a fox in the box that scores big chances. This is why he’s so patchy because scoring lower probability goals is not sustainable over time (Unless you’re peak Messi).
Harry Kane is at 112% so up there, although his numbers this season are bad.
I think that what you're saying is probably what the board is thinking, while the reality is that getting a manager like Pochettino over ETH (and going with RR's advice on future planning) will mean that we'll be doing the Ole strategy of buying big names (Kane, Rice) rather than unearthing gems who would fit into a system and likely cost less individually. The lattes is more what Liverpool did in their rebuilding phase (Salah, Mané, Robertson, Firmino all cost less than a Donny van de Beek IIRC) before getting to the point where they could splurge on individual positions because the rest of the team was set.
There is more than one way to build a club. Real Madrid, Bayern, PSG, Juventus, Man City have typically bought big name players and reloaded rather than rebuilt. They’ve won a lot of titles over the last decade… I would argue that the outcomes are probably better that way vs the Liverpool style of rebuild.
I'd rather keep Rangnick than bring in Pochettino for what it's worth. Ten Hag or Ralf are the 2 options for me. I don't think it'd be a disaster from squad building, while financially, it's the Glazer's that get hit so whatever. They can figure it out.
Squad building wise though... we aren't getting Haaland anyway. Its not all that smart to spend 100m on Rice. Tchouameni is a pipe dream with the cost it would require and competition. We have shown 0 interest in Kessie so aren't likely to go for him. Those are really the only bigger name players on the potential move this summer that would be slightly realistic and make sense and require CL football. We can still get guys like Kamar, Brozović, Bennacer if we show interest, all of the strikers apart from probably Lautaro Martinez that would be useful options... We can still get Ten Hag without it.
Squad building wise, it just rules out guys that were more or less unrealistic anyway, and helps us move on some older guys on massive wages that ultimately don't fit and shouldn't be here next season like Cavani, Ronaldo or Pogba. Really don't think it changes much in the grand scheme of things for us, beyond whatever long term financial implications there are from 1year of Europa League vs CL football (not sure at what point of success in Europa League is equivalent to what level of CL football anyway).
That's two different things I think. All those clubs were at the top (at least nationally) and operated that way to stay there. Liverpool were far off the top, so they had to completely build the backbone of a new squad. A rebuild is what needed when you have only ruins left of previous glory (so the Liverpool way looks to be the way United should follow at the moment, as United is in the same spot as they were).
Liverpool spent 600m putting together that squad. They also had a manager that had a crystal clear idea of how he wanted to play. There seems to be this narrative that Liverpool built that side on the cheap. No. They had a lot of 40-50m transfers. I think our squad is much better than theirs was when Klopp started. Benteke was his no.9 ffs!
Oh I did not mean to imply that they build their squad on the cheap. But they build a squad of players that can all play a valuable role for a title winning squad, and then added the really expensive signings like VVD on top to iron out the last issues. Compared to clubs that just add one world class player every summer to refresh their squad and stay on the top that's a different approach.