Andre grey

[Rjk]

Is in good company on this list.
IMG_20180522_131231.jpg

Dwight Gayle is decent too, but I can't see Newcastle wanting to let him go. And all of the others are well out of our reach.

Of the players who have between 500 and 1000 minutes solanke is top and Iheanacho is right up there.

[delmbox]

Expected is all very well and good but didn't he *actually* score 5 last season though

[Rjk]

5 goals and 3 assists from the equivalent of 16.5 games isn't a bad return, and only slightly below the xg.
He could be above it next season

[delmbox]

5 goals and 3 assists from the equivalent of 16.5 games isn't a bad return, and only slightly below the xg.
He could be above it next season

I'll be honest, and I'm not proud of this, I've never known what xg is

[Harry Monk]

I'll be honest, and I'm not proud of this, I've never known what xg is

Nor me. Every time I've seen it on MOTD I've always whispered under my breath "What the fcuk is xg?"

[Rjk]

I'll be honest, and I'm not proud of this, I've never known what xg is

It's basically a measure of how many goals you would expect to score from the shots you have taken.
If someone has hit 3 shots from 40 yards in a game but no other shots and all 3 of them have gone in, his xg is going to be tiny, but the chances of him repeating that are also tiny. The xg is a better prediction of what is likely to happen next time than the actual goals scored.
Likewise if a striker has 15 shots in the 6 yard box and none go in in one game, he will have a pretty high xg for that game and the chances are if they played it again he'd get some of them in.

The difficulty comes in that there are lots of different ways people use to calculate xg, and they're all called xg, but they differ wildly in sophistication.

On a simple level you just look at where the shot was hit from, find out what proportion of goals that were hit from there out of all the millions of goals on record, actually went in then that's the xg of that shot.

Other models look beyond this, to which foot the striker used or if it's a header, to where the pass before it has come.from, to whether it was behind or Infront of them, to.whether there was "defensive pressure".
The more sophisticated ones use the player tracking data to.take account the positions of the defenders and the goalkeeper as well, although until very recently this was only available for the top leagues.

It's only interested in shots though, other metrics exist for other aspects of the game

[StraightOuttaCanton]

It's basically a measure of how many goals you would expect to score from the shots you have taken.
If someone has hit 3 shots from 40 yards in a game but no other shots and all 3 of them have gone in, his xg is going to be tiny, but the chances of him repeating that are also tiny. The xg is a better prediction of what is likely to happen next time than the actual goals scored.
Likewise if a striker has 15 shots in the 6 yard box and none go in in one game, he will have a pretty high xg for that game and the chances are if they played it again he'd get some of them in.

The difficulty comes in that there are lots of different ways people use to calculate xg, and they're all called xg, but they differ wildly in sophistication.

On a simple level you just look at where the shot was hit from, find out what proportion of goals that were hit from there out of all the millions of goals on record, actually went in then that's the xg of that shot.

Other models look beyond this, to which foot the striker used or if it's a header, to where the pass before it has come.from, to whether it was behind or Infront of them, to.whether there was "defensive pressure".
The more sophisticated ones use the player tracking data to.take account the positions of the defenders and the goalkeeper as well, although until very recently this was only available for the top leagues.

It's only interested in shots though, other metrics exist for other aspects of the game

Strikes me as a load of rubbish. Position of goalkeeper and defenders may be a variable but how does it take account of one goalkeeper being De Gea and the other being a 5ft 10 midfielder who has had to go in goal because of sendings off/injuries?

Maybe the idea is that it evens out over time but it doesn’t sound any more complicated than you can’t score if you don’t shoot

[Rjk]

Strikes me as a load of rubbish. Position of goalkeeper and defenders may be a variable but how does it take account of one goalkeeper being De Gea and the other being a 5ft 10 midfielder who has had to go in goal because of sendings off/injuries?

Maybe the idea is that it evens out over time but it doesn’t sound any more complicated than you can’t score if you don’t shoot

There is an element of " it evens out over time" and an excellent keeper might make a difference, but at the top level there's only a small % difference between any of the keepers or strikers.

Most models even out over about 9 games

[Rjk]

It's slightly more complicated than just counting up shots, last season Brentford had massively more shots than any other team, more than 100 more than anyone else but they weren't the highest in terms of xg because they didn't lot of shots from bad areas.
We had a lot of shots, and they were mostly in good areas

[Pedro de la Rosa]

Is in good company on this list.
IMG_20180522_131231.jpg

Dwight Gayle is decent too, but I can't see Newcastle wanting to let him go. And all of the others are well out of our reach.

Of the players who have between 500 and 1000 minutes solanke is top and Iheanacho is right up there.

How in the hell is Solanke top when he's scored 1 goal? Doesn't that mean he couldn't finish his tea?

[Rjk]

How in the hell is Solanke top when he's scored 1 goal? Doesn't that mean he couldn't finish his tea?

Possibly yes, but with small numbers of games you would expect a much higher variance from the predicted value. He's only played about 6 X 90 minutes. He's scored 1 and his xg thinks it should be 4 or 5. You'd probably see the 2 values get closer together with more game time

[xsnaggle]

It all sounds a load of bollox to me. A striker can either score goals or he can't!! If he can't he aint a striker. Simples!

[ken smith]

I'll be honest, and I'm not proud of this, I've never known what xg is

I'm proud I don't know what it means. Means I'm not a new age, sky generation, championship manager playing Fulham fan.
That or old.

[Badly Ironed Shirt]

I'm proud I don't know what it means. Means I'm not a new age, sky generation, championship manager playing Fulham fan.
That or old.

I'm with Ken on this one. The only stat that I look at is goals per game ratio - then what division did they get those goals in. All this xg stuff sounds a little like p-hacking to me.

[Rjk]

I'm proud I don't know what it means. Means I'm not a new age, sky generation, championship manager playing Fulham fan.
That or old.

I don't think there's any link between understanding x and supporting Fulham, or indeed any of the other things you mentioned.
I'm pretty sure we had a higher xG than fulham this season

[Harry Monk]

It's basically a measure of how many goals you would expect to score from the shots you have taken.
If someone has hit 3 shots from 40 yards in a game but no other shots and all 3 of them have gone in, his xg is going to be tiny, but the chances of him repeating that are also tiny. The xg is a better prediction of what is likely to happen next time than the actual goals scored.
Likewise if a striker has 15 shots in the 6 yard box and none go in in one game, he will have a pretty high xg for that game and the chances are if they played it again he'd get some of them in.

The difficulty comes in that there are lots of different ways people use to calculate xg, and they're all called xg, but they differ wildly in sophistication.

On a simple level you just look at where the shot was hit from, find out what proportion of goals that were hit from there out of all the millions of goals on record, actually went in then that's the xg of that shot.

Other models look beyond this, to which foot the striker used or if it's a header, to where the pass before it has come.from, to whether it was behind or Infront of them, to.whether there was "defensive pressure".
The more sophisticated ones use the player tracking data to.take account the positions of the defenders and the goalkeeper as well, although until very recently this was only available for the top leagues.

It's only interested in shots though, other metrics exist for other aspects of the game

Ah right. I think.

[delmbox]

It's basically a measure of how many goals you would expect to score from the shots you have taken.
If someone has hit 3 shots from 40 yards in a game but no other shots and all 3 of them have gone in, his xg is going to be tiny, but the chances of him repeating that are also tiny. The xg is a better prediction of what is likely to happen next time than the actual goals scored.
Likewise if a striker has 15 shots in the 6 yard box and none go in in one game, he will have a pretty high xg for that game and the chances are if they played it again he'd get some of them in.

The difficulty comes in that there are lots of different ways people use to calculate xg, and they're all called xg, but they differ wildly in sophistication.

On a simple level you just look at where the shot was hit from, find out what proportion of goals that were hit from there out of all the millions of goals on record, actually went in then that's the xg of that shot.

Other models look beyond this, to which foot the striker used or if it's a header, to where the pass before it has come.from, to whether it was behind or Infront of them, to.whether there was "defensive pressure".
The more sophisticated ones use the player tracking data to.take account the positions of the defenders and the goalkeeper as well, although until very recently this was only available for the top leagues.

It's only interested in shots though, other metrics exist for other aspects of the game

Thankyou. I still don't understand, but thankyou