Reply With QuoteI wonder what the two extra games we "should" have won are? I'm guessing Bristol City & Milwall maybe?
| + Visit Cardiff FC for Latest News, Transfer Gossip, Fixtures and Match Results |
https://experimental361.com/2020/12/...-8-9-dec-2020/
Just a bit of fun, usual disclaimers about using xG etc etc, but that ins't bad viewing is it?
We have one of the lowest xG against and one of the highest xG for
I wonder what the two extra games we "should" have won are? I'm guessing Bristol City & Milwall maybe?
I've had a look, you are correct.Originally Posted by City123
I'd have thought Bristol and Middlesbrough perhaps
I take it we should have lost at Stoke?
they had it as 1.2-0.8 so possibly a draw in their system (just)
about 0.8 of their 1.2 was the pen as well
WTF is an "expected goal" anyway.
The only ones that matter are those in the score at FT.
I hate stats for stats sake.
And I've got A Level Statistics.
It's a guess at saying who had the better chances. It is a bit of rubbish but it's a better indication of some stats how games have gone.
Yes. A team could play absolutely dreadfully for 89 minutes but prevent the opposition from getting a decent chance. They could have 20 shots on goal, none on target, all missed by a mile, all from improbable positions, but there is no doubt they've been the better side. The opposition go forward and are presented with an open goal after a mistake. xG for the better team struggles to get anywhere near 0.4, xG for the team outplayed is 1.
Isn't that a contradiction? Burnley FC have stayed in the premier league, and qualified for Europe, by allowing teams to shoot from rubbish positions then being clinical with their one or two chances. Might be rubbish to watch and a more limited footballing strategy but definitely high level of skill involved.
Yes, there are various things that it is useful for, it "works" better over a run of several games, rather than individual games, as the little contributing factors that lead to chances being incorrectly scored tend to even themselves out.
Also as mentioned before, there is no agreed standard way of calculating it, the most basic type (which you usually see) is just based on shot location (i.e. from all the other shots we have in our database from that exact location, how many went in).
From there some people try to refine that model, by adding in other factors like :
where the pass has come from
if the attacker is being pressurised
game state
and some even track the position of the keeper and all the defenders (those aren't typically freely available)
There are others that try to take into account non-shot metrics - things like, how often the ball has been passed into the box, or touches in the box etc.
I'm not as keen on those as I think they surely tend to favour "proper" footballing sides, over someone playing a direct, counter attacking style.
At the start of the season, Reading won a ridiculous number of games in a row, and looked like they could romp away with the division. However, their xg for and against was just a bit better than average, so looking at that you'd say that they were overachieving their performances, and eventually their results would come more into like with the xG - and they have.
Another example in Mason Greenwood. He burst onto the scene last season and scored 10 premier league goals. However, his xG from those games was a little over 3.
Some elite finishers - the likes of Messi etc, can consistently outperform their xG (by basically being better than anyone else in the database who has ever taken a shot from there). However even then it's usually only by a little bit - not 3 x as much. So Greenwood's scoring rate was always likely to slow down.
Other things it is useful for is monitoring how a side is progressing with time - i.e. is a manager gradually improving the team or making them worse, as sometimes a fortunate or unfortunate run of results can be deceiving.
i.e. a team is posting decent xG for and against, but lose a run of games and the manager is sacked. Then a new manager comes in and posting the same xG they win a few more games - is that down to the manager, or would it have happened anyway?
If as time goes on you're consistently posting lower and lower xG then it is probably time to change the manager.
And Burnley's xG usually reflects that. They restrict shots to poor locations i.e. the opposition will have low xG, and they create a low number of very high xG chances (usually by being 1 on 1 with the keeper).
Before Liverpool signed Virgil Van Dyke, their defence was already giving up about the fewest number of shots per game in the division. the problem was they were almost always massive chances, with very high xG. When Van Dyke came in it didn't really effect the number of chances they were giving up, but they were generally a lot harder to take.
Posting Permissions
