League 2 Top Scorer 2024/25

[applepie2]

Presumably xG doesn't take into account the brilliance or ineptitude of the opposing goalkeeper (or defender who makes an astonishing back heeled clearance on the line)? Over a series of matches that should typically even out, but I bet Jatta's miss when he didn't quite put enough pace into his attempted pass into the net explains why his goals scored is a bit less than his xG. I didn't blame him for that "miss".

[ncfcog]

Presumably xG doesn't take into account the brilliance or ineptitude of the opposing goalkeeper (or defender who makes an astonishing back heeled clearance on the line)? Over a series of matches that should typically even out, but I bet Jatta's miss when he didn't quite put enough pace into his attempted pass into the net explains why his goals scored is a bit less than his xG. I didn't blame him for that "miss".

Yes, that's a good observation. It's important to have a reasonable sample size though to be able to ensure any extreme scenarios like the one with Jatta are not skewing the rest of the data.

[Mud Pie]

Unfortunately it's not really as simple as that. Yes, if you compare Abbott's goal at Peterborough to Didzy's second goal at Stanley there is an obvious difference in the xG value. So Abbott's goal was circa 0.02, whilst Didzy's second goal at Stanley was 0.82. Now interestingly I'd have had Didzy's first goal up in the 0.80's as he was literally on the line when he chested it in, but Wyscout have it at just 0.37. Now there is nothing 'just' about that value, it would still represent a big chance but the players position is only one part of the equation.

There are many factors involved. Header, foot shot, strong or weak foot, the assist, the second assist (in this case Abbott's cross), the players movement before making contact with the ball, defender and keeper positioning are all assessed. Additionally, there is a considerable amount of historical data available now so this is also taken into consideration by comparing to similar chances already in the data.

Then, of course you have the ludicrously clever algorithms that work it all out in a matter of seconds.

The thing to remember though is that xG is not an exact science (yet). It's a great barometer for measuring performance etc but it can differ between different data companies. Wyscout and Opta are probably the most accurate and the ones used mostly by professional clubs but FotMob for example have their own version and whilst it presents the shot xG live 'in game' you tend to find it re-evaluates it shortly after the game has finished.

I love working with xG data because despite the fact there can be discrepancies they're generally within a workable tolerance. Used correctly they help to identify trends relating to both team and player performance. The recent visualisations I shared regarding Notts managers is an example of that. You get a snapshot of something that presents an element of a player or teams game that needs some closer attention.

Notts' xG isn't the best in the league but what is good right now is that the comparison between xG and goals scored is very close. If your xG is considerably lower than goals scored it would intimate that you are heavily reliant on scoring worldies or you have had more than your fair share of luck. The question in this case is just how sustainable that is. Historically it has been proven in most cases it isn't sustainable and if those metrics start to level off you are likely to see a drop in results or goals scored.

The opposite of that would be underperforming your xG value. This would suggest you are wasteful in front of goal so you may want to dig a bit deeper into that by analysing your missed chances etc.

Finally (honest), we now also have xG on target data. This is a slightly different metric that focusses on shots on target. Basically, the likelihood of a goal from an attempt on target. I'll delve into that one another time if that's ok!

thanks a ton for explaining all that, og! So much clearer with examples we all remember, but it's a complex old game, innit!

[Boomerangpie]

A lot more complex than many think it is that?s for sure