Experimental 361 | Page 3 | Vital Football

Experimental 361

gijsbert_bos said:
I think the problem some people are having is not understanding these models and what they are supposed to show. If we get 13 corners then Bostwick will have a pretty decent XG. Same I imagine for Shackell had he been playing and perhaps Wharton if he continues to score from corners.
Click to expand...
I don't think that's right. As I understand it, it counts only shots/chances had in that match.
 
Luke Imp said:
Exactly, and that's the point. McCartan's was a more difficult goal, and as such has a lower XG rating.

Your XG will be higher if you create 4no Boswick chances than 4no McCartan chances because the former were easier finishes that you'd expect to result in goals 99% of the time.

Of course, they'll be the odd anomaly where you dominate and end up scoring when a player pings one into the top bin and you win 1-0 with a low XG but by and large, the better quality of the chances, the more likely you are to win a game and that's what it's showing.
Click to expand...

and using knowledge of football, it is logical to assume that a fella capable of scoring a more difficult goal - as mccartan did - is even more capable of tapping the ball in from one yard. so his individual xd should go up.

if the team create 4 'bostwick chances' i.e. corners... mccartan is also in the box for every corner, just as bostwick is - so has equal chance of scoring the bostwick-type tap in at every corner.

the chart confuses/conflates team chances created, with individual player's chance of scoring.
 
gijsbert_bos said:
I think the problem some people are having is not understanding these models and what they are supposed to show. If we get 13 corners then Bostwick will have a pretty decent XG. Same I imagine for Shackell had he been playing and perhaps Wharton if he continues to score from corners.
Click to expand...
'some people' not understanding... might even include the fella running this show.

from his site:
Ranking the top three players
One thing I was keen to do was make the graphics larger so that they could look more crisp and make better use of modern phone screens, but I quickly realised there was room to include additional data alongside the main timeline. I’ve plumped for including a small colour-coded bar graph showing the three players who racked up the highest expected goals totals in the game and therefore posed the greatest threat.

haha. no it doesn't show that.

he gives the highest ranking to the easiest chances, so one tap-in opportunity for bostwick [even if he misses it!] apparently means he posed a greater threat in a game, than say mccartan who, for example, has just scored a hat-trick of top corner worldies.

this is where it goes off the rails imo. so a fella misses three penalties, his ranking would be 2.4xd this in no way shows that he posed the greatest threat in any game.

and to be clear, i can see what he is doing with the cumulative team Xd. he is just over-egging it with the individual player Xd.

do betting companies really use this kind of model to predict which individual has the greatest chance of scoring the next goal? or to predict an individual’s chance of scoring in a future game?
 
nlondonimp said:
and using knowledge of football, it is logical to assume that a fella capable of scoring a more difficult goal - as mccartan did - is even more capable of tapping the ball in from one yard. so his individual xd should go up.

if the team create 4 'bostwick chances' i.e. corners... mccartan is also in the box for every corner, just as bostwick is - so has equal chance of scoring the bostwick-type tap in at every corner.

the chart confuses/conflates team chances created, with individual player's chance of scoring.
Click to expand...
Yes, but it's about the quality of the chance, not the player who takes the chance.
 
nlondonimp said:
'some people' not understanding... might even include the fella running this show.
Click to expand...

I like a statistical analysis as much as the next geek, and I also think the late Charlie Hughes was the most misunderstood coach in English history, but these XG stats make no sense.

If as (I think) you already pointed out, they make no distinction between player abilities, then Messi is going to look like he's always over performing as he scores an impossible number of impossible goals.

And is it just shots? Because crosses that are missed could be just as good an indicator of a possible goal than a mis-hit effort from 25 yards. And do they count shots that reached the goal line or blocked shots too? Do goalmouth scrambles count as 1, 2 or 3 attempts?

Would it be the case that the own goal yesterday has a zero rating despite the fact that the ball was in a very dangerous area before it was nodded in?
 
To be honest I think that's part of the point. Comparing XG with actual is just another layer of information to help analyse team / individual performances.
 
Brendan Bradley said:
I like a statistical analysis as much as the next geek, and I also think the late Charlie Hughes was the most misunderstood coach in English history, but these XG stats make no sense.

If as (I think) you already pointed out, they make no distinction between player abilities, then Messi is going to look like he's always over performing as he scores an impossible number of impossible goals.

And is it just shots? Because crosses that are missed could be just as good an indicator of a possible goal than a mis-hit effort from 25 yards. And do they count shots that reached the goal line or blocked shots too? Do goalmouth scrambles count as 1, 2 or 3 attempts?

Would it be the case that the own goal yesterday has a zero rating despite the fact that the ball was in a very dangerous area before it was nodded in?
Click to expand...

it's usually used as a predictor of an outcome of a match rather than the scoreline
 
If I'm reading it correctly it looks like we've scored twice as many goals as XG suggests 30 vs 15. Whilst our goals conceded is in line with XG.
 
I'm reading it as:

XG is 27.62 and scored 30.
XGA is 15 and conceded 13.

So we're scoring slightly more and conceding slightly less than those figures show.
 
You must log in or register to reply here.