cross-posted from: https://mastodon.online/users/hallenbeck/statuses/111293316231491706
I’ve made an updated graphic based on feedback. Thanks to @[email protected] for the critique. Updated graphic here:
https://thelastboyscout.uk/assets/img/son_xg_stats.webp
Is Son one of the best finishers EVER? Let’s look at some data. 👇
We use actual goals minus expected goals (xG) as a proxy for finishing skill. Players who consistently score more than their xG means they are scoring goals other players would miss. Generally, only the most elite goalscorers *consistently* outperform their xG.
And I can find no player who consistently beats Son. It’s astonishing.
Can you find anyone better at elite level?
Son has impressive overperformance of xG but that graphic has very suspect math. Just based off the raw Goals and xG numbers from the graphic:
- Son had a total of 92 goals from 17-18 to 23-24 with an xG of 65.8. That’s an overperformance of 40%.
- Kane had a total of 144 goals from 17-18 to 23-24 with xG of 119.2. That’s an overperformance of 21%.
- Messi had a total of 148 goals from 17-18 to 23-24 with xG of 120.5. That’s an overperformance of 23%.
- Mbappe had a total of 156 goals from 17-18 to 23-24 with xG of 136.7. That’s an overperformance of 14%.
This summing of xG from season to season and averaging of xG from season to season is total BS. Especially considering some of the injury plagued seasons by both players they played very few games, for instance Messi this year.
So yes, Son is a great finisher but please don’t bring this fraudulent math and claim he’s so head and shoulders above everyone else. Also, an argument can be made that scorers lose efficiency when they attempt higher volume, so I’d like to see him replicate that efficiency with a higher goal output.
Lastly, here’s a chart with Messi’s goals and xG so we can compare in the same age range (25-31).
From 12-13 to 18-19 Messi had a total of 250 goals with an xG of 185.02. That’s an overperformance of 35% with more than twice the output. Obviously it’s debatable but I think peak Messi is a better finisher than Son, and current Messi is not too far behind son despite being 5 years older.
A little heavy to say “fraudulent” and “total BS” - it’s a sketch I quickly knocked up from FBref data (which only goes back to '17) to help me understand Son’s low xG and relatively high actual goals. It’s meant to be indicative rather than precise. Critique and improvement always welcome, but no need to be so rude, mate. I’ll make some improvements based on your feedback.
I was a little surprised by these figures myself so went looking for any prior work in this area and found this from '21: https://www.independent.co.uk/sport/football/premier-league/tottenham/son-heung-min-xg-harry-kane-b1782968.html. The figures there broadly reflect what I’ve got here. The point being made remains the same. It also contains some theories as to why Son is such an outlier, but the conclusion seems to be Son is indeed unusual.
“Goodman agrees, though, that in a game of mostly uniform standards of finishing, Son’s numbers are startling.”
“Goodman” here is Mike L Goodman, the former managing editor of StatsBomb.
I agree it will be good to see how he fares with a) more shots; b) more goals; c) more play in the box. This season and beyond with him playing as a 9 should hopefully shed some light.
I also agree he’s probably not the one-of-a-kind genius Messi is. But to be even vaguely comparable is in itself quite a remarkable thing. And the bigger question for me is, is there anyone else, aside from Messi, that comes close?
Update: here’s an improved version: https://thelastboyscout.uk/assets/img/son_xg_stats.webp
I don’t think the claim that Son is one of the best finishers ever is BS or fraudulent.
I think that summing xG over performance by season is very very poor math. For instance, if Son scored 20 goals on 10 xG in one season followed by 0 goals on 1 xG, you would sum those to have an over performance of 0%. When in reality proper counting techniques should state 20 goals on a total of 11 xG for over performance of 82%.
The mathematical “technique” you chose grossly inflates his numbers compared to other finishers and grossly short changes Messi over the examined time period.
I’ve redone it with a little more care (see above) and agree the summing of xG over performance wasn’t a good idea. It was a vestigial artefact where I’d been mucking about and I wish I’d removed it, but it was late and I couldn’t be bothered as it didn’t really affect the point I was making and it’s not intended to be rigorous maths.
The results in the updated version aren’t significantly different apart from Messi. The reason Messi’s over performance has shifted so much in the update is because I removed the three PSG and MLS years, which I think are an unfair reflection on him.
I don’t agree averaging the xG overperformance “grossly” inflated the numbers (unless you consider 6% gross). I do agree including the PSG and MLS years grossly short changes Messi, but I noted something to that effect in the original.
You don’t need to be quite so hostile to people sharing their stuff here btw - save that for Twitter or Reddit if hostilities float your boat. I take it you’re not and have never been an educator? But no harm done - thx for the feedback - it was useful and improved the analysis :)
You’re right, I am sorry about being hostile. I should have been more light hearted about it. I just thought that it’s another hyperbolic post contriving to exxagerate how good a player is, just like a lot of Twitter posts.
However, I disagree that cherry picking data would improve your analysis. Include all the years, but weight them properly (as the overall G/xG would).
Keep up the OC. Sorry again and cheers!
Do you know where I can get xG data before 2017? That’s where FBref stops.
I just googled it for Messi. I assume it’s available for a smattering of top forwards.
deleted by creator
Would you say he’s THE most elite finisher, though? He’s definitely up there when comparing the data of his peers. That’s what I mean but underrated. He takes fewer shots than the others, so that could be skewing things. But then again, if you look at other players who take a similar number of shots, they don’t overperform like Son consistently does. I don’t know, he seems to be a bit of a outlier. I can’t find any other player like him. Will be very interesting to see what his numbers are like at the end of this season with him playing as a 9 and without Kane. Will he still overperform?
Isn’t a player’s xG a product of team performance and individual performance?
Not saying this just for the memes, but surely spurs players take an xG hit for playing for Tottenham simply because the team itself has not looked particularly threatening for the past few seasons.
No, it takes into account the events and circumstances of the play, not the individual skill of the players involved. At least, that’s how Opta’s xG model works.
xG does not take into account the quality of player(s) involved in a particular play. It is an estimate of how the average player or team would perform in a similar situation.
More here: https://fbref.com/en/expected-goals-model-explained/
Oh I think I see what you’re saying: a poor team will create poorer chances in general, leading to lower xG shots. There’s truth in that, but is it the case at Spurs? The club has had an average league position of 4.6 over the last 10 seasons, which suggests it has had above average players, regardless of (lack of) trophies, so you’d have thought, on balance, those players would in general be capable of “good play” resulting in high xG chances. All that is besides the point, though. This analysis is about how Son has an uncanny ability to score low xG chances.
Handy hint: you can eyeball low xG goals by looking at a player’s Goal Log on FBref then sorting on the xG column, low to high. Always fun to check out the really low ones. The PSxG stat (called xGOT on Sofascore) gives us an indication of how savable the shot is - the closer to 1, the less likely to save. It’s a useful indicator of the “quality” of the shot to go along with the “difficulty” represented by xG.
E.g. https://fbref.com/en/players/92e7e919/goallogs/dom_lg/Son-Heung-min-Goal-Log
deleted by creator
I’m afraid I don’t think it works like that, although the specifics of xG algorithms that stats companies use are generally closely guarded secrets, they often talk about how it works in loose terms.
xG is usually produced by a huge predictive model which takes in a combo basically every possible data point you can imagine to do with a football match.
Opta talks about their model being fed by things like historic data for the ball position, limb positions, the quality of passing building up to a shot (some models talk about every pass from when the ball was out of play), the quality of the keeper in front and all sorts of other stuff they don’t let on about.
Given all that I’m not sure it’s possible to meaningfully normalise a player’s xG because it’s a product of other players on a pitch. So I think it’s fair to conclude that a poorly performing team will very likely be negatively impacting an individual’s xG.
deleted by creator
