clint van damme
Well-Known Member
MR and Adi looking at this thread
Because some on here bum XG (7 Viewers)
- Thread starter Saddlebrains
- Start date
Frostie
Well-Known Member
"The league table doesn't lie" is the biggest lie.
Evo1883
Well-Known Member
I'd be very surprised if the club doesn't massively use data
clint van damme
Well-Known Member
They'll use data far beyond XG, all clubs do now.
I'd be surprised if XG isn't quite rudimentary compared to some of the stuff they use.
clint van damme
Well-Known Member
It doesn't matter darling, you've had a lot on your plate recently, we'll try again soon, is the biggest lie
Frostie
Well-Known Member
clint van damme
Well-Known Member
Cruel!
shmmeee
Well-Known Member
Let's get this afternoon's game out of the way first, scckmee
Edit:- I hope you know Scala
I know enough Scala to know I don’t want to use it
Cally Fedora
Well-Known Member
For anyone furiously combing through the stats we won 2-1.
shmmeee
Well-Known Member
Nerd. Did everyone have fun? That’s what matters.
shmmeee
Well-Known Member
There’s a reasonable point here actually which is that all the models afaik are proprietary and closed source so we can’t critique them other than the final output.
I reckon with modern vision processing software you could feed a game in and get what you need to build an open xG model. Not sure how copyright would work though.
I don't think the copyright would be the issue per se, because you would develop your own algorithm/model so would own that.There’s a reasonable point here actually which is that all the models afaik are proprietary and closed source so we can’t critique them other than the final output.
I reckon with modern vision processing software you could feed a game in and get what you need to build an open xG model. Not sure how copyright would work though.
The interesting part is who owns the data, I.e. does it belong to the players, the club, the broadcaster etc.
I think there is even a legal case at the minute with a former player arguing that all his training data etc belongs to him and he should be paid for it.
Sent from my SM-S911B using Tapatalk
JAM See
Well-Known Member
Where would Simms's attempt that went out for a throw in sit in xG?There’s a reasonable point here actually which is that all the models afaik are proprietary and closed source so we can’t critique them other than the final output.
I reckon with modern vision processing software you could feed a game in and get what you need to build an open xG model. Not sure how copyright would work though.
It was at the other end obviously, but it looked a decent chance to me from over 100 metres away.
It's the algorithm/state machine that I'd like access to.
shmmeee
Well-Known Member
Used to be with sports stats that the person collecting the data owned it, not sure about now. But does your licence for watching iFollow include extracting data? What’s the difference between a machine vision algo and Statto with a notepad? I suspect it would go to court cos everyone wants a piece of the pie. Cant imagine legally getting access to every game would be cheap. Some kind of app you could run on your PC while watching a stream that collected data would be amazing.
Mucca Mad Boys
Well-Known Member
Likewise, Sheff Wednesday’s xG would be skewed by their goal. It was a goal mouth scramble where 2 close range shots were saved/blocked before the goal.
Aside from that, they didn’t really create much. Neither did we - it was a pretty scrappy away win.
robbiekeane
Well-Known Member
Chart gave me a boner to be fairFucking mental , imagine bumming that
stupot07
Well-Known Member
If you scroll through this it shows the xG For all the shots
Our highest was Ben Sheafs second, 0.29 xG, our next highest was the Godden shot in the 76 minute which was 0.10 xG
For them they had
Liam Palmer 52 minute 0.18 xG
Ugbo 68 minute 0.11 xG
Windass 68 minute 0.23 xG
Musaba 68 minute o.56 xG
Cadamateri 72 minute 0.11 xG
Wilks 89 minutes 0.29 xG
Coventry City vs Sheffield Wednesday - live score, predicted lineups and H2H stats
Coventry City vs Sheffield Wednesday on Sat, Oct 5, 2024, 14:00 UTC. Check live results, H2H, match stats, lineups, player ratings, insights, team forms, shotmap, and highlights.
www.fotmob.com
Sent from my Pixel 7 using Tapatalk
Frostie
Well-Known Member
Yeah it's great to look at but the thing to remember with this is it's just presenting you the raw data for individual shots (and not particularly reliable). To do a proper analysis on it you would need to apply some context.If you scroll through this it shows the xG For all the shots
Our highest was Ben Sheafs second, 0.29 xG, our next highest was the Godden shot in the 76 minute which was 0.10 xG
For them they had
Liam Palmer 52 minute 0.18 xG
Ugbo 68 minute 0.11 xG
Windass 68 minute 0.23 xG
Musaba 68 minute o.56 xG
Cadamateri 72 minute 0.11 xG
Wilks 89 minutes 0.29 xG
Coventry City vs Sheffield Wednesday - live score, predicted lineups and H2H stats
Coventry City vs Sheffield Wednesday on Sat, Oct 5, 2024, 14:00 UTC. Check live results, H2H, match stats, lineups, player ratings, insights, team forms, shotmap, and highlights.
Sent from my Pixel 7 using Tapatalk
For example, a proper xG model shouldn't count the Ugbo & Musaba shots. As they are in the same passage of play as the Windass goal it's not possible for all 3 to be scored.
Calista
Well-Known Member
Does xG take account of incidents such as the ball across the box that Godden missed by a whisker? He didn't get a toe to it so does it count for nothing? You could ask the same about all dangerous crosses / moves that don't quite result in a goal attempt. They might come far closer to scoring than a tame effort from an unlikely position that has an xG attached to it.
No, it doesn't, which is why it's now being surpassed - it's considered the more basic of the advanced metrics.
You can now look at expected threat etc which would involve things like that.
For example, Southampton dominated the ball for large parts against us but offered zero threat outside of their goal. While we played without the ball and arguably offered more goal threat with it - so you can look at metrics on this.
Sent from my SM-S911B using Tapatalk
Nuskyblue
Well-Known Member
field tilt, is that a thing? I'm sure I saw sometihng about this a few years ago.
To an extent, field tilt is generally the % share of the match you spend in the opposite half - but it's that kind of thing.
We can also look at things like xA (expected assist) as a measure of final ball quality or attacking inputous.
Sent from my SM-S911B using Tapatalk
Mucca Mad Boys
Well-Known Member
If you scroll through this it shows the xG For all the shots
Our highest was Ben Sheafs second, 0.29 xG, our next highest was the Godden shot in the 76 minute which was 0.10 xG
For them they had
Liam Palmer 52 minute 0.18 xG
Ugbo 68 minute 0.11 xG
Windass 68 minute 0.23 xG
Musaba 68 minute o.56 xG
Cadamateri 72 minute 0.11 xG
Wilks 89 minutes 0.29 xG
Coventry City vs Sheffield Wednesday - live score, predicted lineups and H2H stats
Coventry City vs Sheffield Wednesday on Sat, Oct 5, 2024, 14:00 UTC. Check live results, H2H, match stats, lineups, player ratings, insights, team forms, shotmap, and highlights.
Sent from my Pixel 7 using Tapatalk
A pretty even game than when all said and done. 0.9 xG for that one goal boosts that metric.
It is interesting that the xG shows we sat back after their goal.
xG doesn’t but there’s a metric called ‘expected threat’ xT which does measure this.
That does account for when teams are creating opportunities but don’t get shots off.
Last edited:
Incorrect.
All players with the exception of Messi have regressed to the mean or advanced to the mean across the top 5 leagues - meaning xG remains accurate and largely there's little difference in finishing quality across the board.
Sent from my SM-S911B using Tapatalk
Calista
Well-Known Member
? Are you saying that given a particular scoring opportunity, all players would be equally likely to stick it in the back of the net?
Torquay Sky Blue
Well-Known Member
I agree, I've recently been working with a couple of black guys and taken to calling everyone bro or bruv. It's corny as fuck and has to stop
Skyblueweeman
Well-Known Member
It's most definitely not.
Skybluedownunder
Well-Known Member
Seeming as people on here like making little singy songs…
Here’s one for the choir boys…
You can stick your fucking XG up your arse,
You can stick your fucking XG up your arse,
You can stick your fucking XG,
Stick your fucking XG,
Stick your fucking XG up your arse
Sent from my iPhone using Tapatalk
fernandopartridge
Well-Known Member
Philosoraptor
Well-Known Member
More or Less: Behind the Stats - Are there more possible games of chess than atoms in the universe? - BBC Sounds
We investigate how the vast possibilities in chess compare to the vastness of the universe
www.bbc.co.uk
Sick Boy
Super Moderator
I lived in Deptford over 20 years ago and it was pretty common even then. It’s spread out to the Midlands these days as well, brav.
Martin180
Well-Known Member
I hear you fam
shmmeee
Well-Known Member
More or Less: Behind the Stats - Are there more possible games of chess than atoms in the universe? - BBC Sounds
We investigate how the vast possibilities in chess compare to the vastness of the universe
Right, now multiple that by a million and square it and you’re still not close to the number of possible states in a football match.