Interestingly enough what does that make the likes of Jesus & Sterling? Terrible players?
Rashford has scored 62 goals from an XG of 60.79.
Raheem Sterling has scored 101 goals from an XG of 110.
Gabriel Jesus has scored 82 goals from an XG of 82.35
xG and finishing under ETH
- Thread starter Samid
- Start date
Rashford has scored 62 goals from an XG of 60.79.
Raheem Sterling has scored 101 goals from an XG of 110.
Gabriel Jesus has scored 82 goals from an XG of 82.35
Do you have stats to back up your last statement? How much do they out-perform?
We outperform our xG most matches. Finishing is not an issue.
Technically his last statement is right, yes the best finishers outperform xG because that is what finishing is. But it's a stat that doesn't say all that much and doesn't correlate to being the best goalscorer or even being a good player at all. The best goalscorers generate high xG, they can finish right on xG but it's the ones who generate high xG who are the top scorers year after year. Best finishers could be some random guy in the 3rd Division who isn't a good player but the few times he gets a good chance, can finish clinically.
Which would make sense.
I believe XG is just based on historical data which shows how likely a chance is likely to be scored - and funnily enough, the top players get more of these chances than the average players so it's already skewed by them.
If you were to build an XG model purely based on the current season in the PL - Erling Haaland has scored 15 goals from 38 shots (more goals + shots than any other player). So any XG model would be skewed more by Haaland more than any other player. Btw 38 shots is more than half the number Bournemouth as a club have taken so far this season.
I definitely agree with that.
But xG is something I actually find rather interesting and certainly not something I would just dismiss as modern nonsense.
And I'm speaking as a relatively old fart here (I do detest many aspects of modern football).
I am not against it as a whole, but too many people just ignore what they watched and just use numbers and stats for arguments. Numbers will NEVER tell a full story.
True that, Yoda.
However, I'd be much more worried if we didn't create chances.
We could've/should've scored at least three more last night (xG more or less bears this out, incidentally).
I believe we now have the best overall xG in the EL (skewed by last night, yes). Which is, well, what it is: not a bad thing at any rate.
Oh, yes - I agree with that.
I suspect that certain fans these days don't actually watch any games (beyond highlights), they just browse stats.
To me, what's interesting about stats is to compare them with the old eye test: sometimes that can be extremely interesting/useful. A player you reckoned mishit passes all over the place turns out to have hardly mishit a single pass (but you, pretty much expecting him to waste a lot of passes, registered the two or three he did mishit - and ignored the rest).
(On the other hand, sometimes those two or three were absolutely crucial ones - and it doesn't matter that the player's completion percentage was in the 90s.)
The problem with the eye test is the quality of the eyes watching it.
Scoring is obviously the most important part of the game. But when rebuilding and developing a new system, new coaching, new ethos it’s actually the last part that comes.
The most impressive thing so far is that we are improving on the foundation stuff eg Defence is good, midfield has improved, creativity is getting better, the pieces of the jigsaw are coming into place.
The most impressive thing so far is that we are improving on the foundation stuff eg Defence is good, midfield has improved, creativity is getting better, the pieces of the jigsaw are coming into place.
Of course.
And even high-quality eyes (as it were) can be influenced by bias(es) - even subconscious ones.
The modern obsession with stats can actually be a positive in that regard (or so I would argue): if you don't dismiss stats altogether (as modern bollocks), you can use them to moderate/balance out what your eyes tell you there and then.
It has to be a combination, though - that's the very point: stats in isolation are almost always potentially misleading (in one way or another). And relying only on your eyes (or more precisely, your memory of what you've seen) is almost always a fallacy in itself.
So is there any measure to define whether better strikers get more opportunities? Obviously they play for better sides so will get more chances by default but maybe that could be equalised?
If they aren’t better at scoring from any % chance then surely they must simply get more chances through their movement or anticipation.
That's exactly what it is. The best strikers find themselves in more high quality chances to accumulate higher xg. They take more shots, are able to create high quality chances for themselves from low quality positions, and are consistent in finding these spaces where they get big chances. That's why you never really worry about a striker if he goes on a run of missing chances. You worry once they stop getting the shots off, you worry once they stop finding themselves on big chances, once they're unable to turn small moments into big chances, etc. It's why I think Ronaldo is done at this level, because for the most part, it's not like he's missing chances (though that was the case vs Omonia last week), he just isn't involved or looking like a danger at all. Can't turn half decent positions into good shots. Rashford on the other hand, just an off finishing night but on another day he'd score 3.
Edit: also why it's likely Darwin Nunez will probably go on a scoring run. He gets loads of chances. I remember for a while people felt Cavani was a donkey because of how many he'd miss... But he just kept finding himself in those positions so he'd score 30+ regularly.
Generally I have an opinion based on watching games and then I use stats to confirm or disprove that view. Sometimes your eyes can be extremely deceiving and that's generally the case for tackles. An example that I find amusing is the difference of perception between Pogba and McTominay in 20-21 the difference between them was essentially one more tackle every 3 games for McTominay and yet the perception between both players was massive.
I would suggest non-penalty xG accrued per 90 minutes. The best players are usually above 0.6. The below link is for the top 5 leagues last season, this season hasn’t gone on long enough to draw firm conclusions. Sort by npxG and disregard anyone with less than 20 90s played. Top three were Lewandowski, Schick and Moussa Dembele.
https://fbref.com/en/comps/Big5/2021-2022/shooting/players/2021-2022-Big-5-European-Leagues-Stats
Edit: forgot the link!
Last edited:
No doubt that it's already been said but a 1-0 against those stats is a freak result and nothing to worry about. It may be embarrassing on a basic level but that's it.
I'm also a bit worried about the quality of chances we create. Although we do a decent job of creating half-chances we do a fairly poor job of creating great chances. That's all pursuant to my eyeball test and not whatever xG might claim. None of the several chances that Rashford had last night were anything close to glorious chances, but they were chances nevertheless and he couldn't convert on any of them. McTominay made the most of out of a fairly tricky chance -- first to control the ball, then take a second touch before the shot -- even though he was just outside the six-yard box. It wouldn't be fair to call it luck, as McTominay worked out a difficult ball for the goal, but there was an element of surprise that one of our clumsier players with the ball at his feet was the one who was able to score on a half-chance, one of about a dozen half-chances in the match.
One of the constants in the quality of the very top strikers over the decades is the ability to consistently score on half-chances. Quite a few of Ruud's and Rooney's goals came from chances that you wonder how the fukk could they score on that angle or through that many defenders? It's not fair to compare anyone to Ronaldo and Messi, but I'll just mention that so many of their goals over the goals came out of nothing, situations where the ball was nowhere close to being on a platter.
What we really need next season is a striker who can make the most of a half chance and consistently score on great chances. Martial doesn't give us that, and Ronaldo is in the dying embers of his career. Casemiro solved our CDM problem for the next 2-3 seasons, but now we need a 9 to round out what is a promising squad under ETH.
Oh, I hope so. I really, really hope so (it would be fantastic to have an actual, consistent top class player in that position again).
As for the second part - yes, undoubtedly.
Relying on Martial is madness. And no - we don't have anyone else (Rashford is never going to do it). We absolutely need to make an astute purchase (hopefully in January - if not, then definitely next summer).
Is Jesus a top player? He's had a good start at Arsenal but was pretty hit and miss at City.
I can understand some people getting annoyed with xG being brought up in threads where it isn't the topic of conversation, but it's equally annoying when people come into a thread where the first word of the title is "xG" only to moan at the use of it. I bet when they started showing shots on goal on the telly the same people were like "what's that supposed to be good for? The final score's right there at the top that'll tell you who was better!"
yeah we're discussing slightly different things here
me what xG is measuring
you the relevance/importance of those statistics
why would I need stats to back it up? they wouldn't be the best finishers if they weren't outperforming their xG
but look up Salah's record in the PL for Liverpool, he's been consistently overachieving for 5 years (maybe not this season as he's been off form)
I had a perfect view of that from behind the goal. Everyone around me was cracked up laughing when the referee awarded a corner, for exactly the reason you stated.
Why is everyone obsessed about players outperforming their xG? Getting in the right position, so getting a high xG, is more important at the top level than outperforming the xG.
Looking at Understar I'm not really seeing it.
17/18 was pretty crazy, of course, but for a lot of the best players it seems like it's more about increasing their xG rather than outperforming it by a lot. His Liverpool career excluding this season is 118 goals, 114.2 xG.
I mean he's tailed off but he's outperforming his xG, clearly, and he's doing that because he's an above average finisherLooking at Understar I'm not really seeing it.
17/18 was pretty crazy, of course, but for a lot of the best players it seems like it's more about increasing their xG rather than outperforming it by a lot. His Liverpool career excluding this season is 118 goals, 114.2 xG.
I'll admit I thought his stats were better, though
The best finishers almost by definition outperform their xG.
The point is that outperforming your xG is generally less valuable than the ability to actually accumulate/generate high xG in the first place. In other words the best goalscorers tend to be the best chances-getters more than they are the best finishers.
The point is that outperforming your xG is generally less valuable than the ability to actually accumulate/generate high xG in the first place. In other words the best goalscorers tend to be the best chances-getters more than they are the best finishers.
yeah I agree with this completely
Exactly this. The best goalscorer is someone who can get himself into four 0.5xG scoring positions every game and score two goals every game rather than someone who will continually shoot from range and outperform the xG (eg taking five 0.10xG shots and scoring one goal every game).
With that said, we never talk about long-shot merchants as good finishers anyway because long-shots are regarded as a different skillset. We always talk about the ability to finish high-scoring chances or 1v1s when we refer to a good finisher, so even if you’re slightly below average at finishing from the position that Antony scored against Arsenal from, continually getting in that position is going to get you a lot of goals.
xG is one of the worst things introduced to football...
Excellent posts, couldn't agree more.
Interesting point about Ronaldo, it is indeed a huge worry. The only hope is he goes on some kind of killing spree, finishing chance after chance. It doesn't look likely, in fact he's the last of our forwards that I can imagine this could happen to. He just doesn't look like a threat.
I like that we're slowly improving our game to accumulate more xG in a game, this is the right direction because that's the part of the game ETH can control. Finishing is not.
It's actually hilarious that those people are willing to waste their time on a topic they're completely disinterested in.
I thought this was a great video.
Apparently one of the comments notes no team ever has come close to xG of 5. Is that true? Fascinating if true.
Apparently one of the comments notes no team ever has come close to xG of 5. Is that true? Fascinating if true.
You have games like that where the keeper grows in confidence as the game goes on, the crowd grows more nervous, the players get affected by it etc. It all just amps up the longer it goes on. If we score after 10 minutes we probably go on to win 4-0. Just a freak game really, I certainly wouldn't be worried about a trend or anything.
The thing with xG is that it’s inherently flawed, yet is used as some kind of golden bullet. xG says this so it must be the case, which is nonsense. For instance look at our game the other night, our two best opportunities to score in the game both resulted in the same xG, 0.0. There was a cross early on from (I think) Bruno on the left where nobody ran in for a simple tap in, and there’s when Ronaldo ignored the ball bouncing at his feet on the 6 yard line with an off balance keeper to claim a penalty (which probably had a higher xG than a penalty). Stats have a massive place in the “improvement” of the game albeit for me I prefer a bit of magic than how the likes of City play, but that’s a personal gripe of it becoming too sanitised with no risks, but your layperson demanding to be taken seriously because the xG says and the xG is the be all is only turning people off of it.
That’s aside from how football as a flowing game isn’t very easy to read from stats, as no two incidents will ever be the same aside from maybe a penalty or a direct free kick. Open play will always be that any moment will be from a series of events. I love cricket for incident, and data is changing that massively but it’s a lot easier to, because every ball is a new event.
In essence I guess I’m trying to say that the problem isn’t xG, it’s in the people that use it without understanding the massive limitations that it has. I can promise you that any data driven club will only use xG as the thread to be picked at rather than the shirt itself.
There's different xG models out there but it happens. Found these just looking through Barcelona, Bayern, Liverpool, Man City and Real Madrid results since 14/15 (domestic league only).
A quick look at PSG's league results say they have hit 5+ xG too but I can't be bothered going through anymore/typing them out. Then there's all the leagues and competitions out there so safe to say it does indeed occur.
21/22 Bayern 6.88-1.23 Bochum (7-0)
19/20 Man City 6.83-0.34 Watford (8-0)
15/16 Barcelona 6.13-0.62 Gijon (6-0)
18/19 Barcelona 6.04-1.64 Huesca (8-2)
19/20 Bayern 5.84-0.54 Bremen (6-1)
15/16 Deportivo 2.20-5.76 Barcelona (0-8)
17/18 Man City 5.73-0.61 Palace (5-0)
15/16 Real Madrid 5.72-0.93 Rayo Vallecano (10-2)
17/18 Barcelona 5.55-0.75 Espanyol (5-0)
18/19 Bayern 5.36-0.87 Hannover (3-1)
14/15 Man City 5.35-0.87 QPR (6-0)
21/22 Liverpool 5.33-0.14 Leeds (6-0)
17/18 Real Madrid 5.31-0.98 Real Sociedad (5-1)
21/22 Hertha Berlin 0.85-5.19 Bayern (1-4)
Last edited:
At the women‘s Euros this summer England racked up 6.54 when they beat Norway 8-0. You expect that quite a few games in women‘s international football might have such high xG games given the stark disparities in levels between some nations but both sides were FIFA top 10 ranked in this game. It would be interesting to see if there’s a higher xG average in top level women‘s games compared to their male counterparts.
Tweet
— Twitter API (@user) date
xG would be considerably less useful if there were too much correlation between player quality and shot conversion.
Yeah the whole premise of xG is that we assume the majority of players finish roughly at xG over a long period of time.
The issue I have with xG is just the people wanting it to be more powerful than it is.
People want to take an xG value, round it to the nearest integer, call that "the goals you should have scored," and then claim that goals above that are "overperformance" and goals below that are "underperformance." That's not really a sound use of the stat. It's only marginally better than a pundit saying "he should have scored that."
People want to take an xG value, round it to the nearest integer, call that "the goals you should have scored," and then claim that goals above that are "overperformance" and goals below that are "underperformance." That's not really a sound use of the stat. It's only marginally better than a pundit saying "he should have scored that."
Last edited: