Miller

[danumdon]

Either the system changes or he takes some time out for me. All we seem to be achieving is destroying the fellas confidence. It's always the case that when a player's confidence is shot they cannot achieve even the most basic of commands such as shoot, regardless of which foot its on, he's also started to really struggle to receive and control any ball sent through to him, this has had the result of killing stone dead any attacking intent we attempt to achieve with him playing the front hold up role.

If we don't aim to change our playing strategy then i'd say Lavary should get the nod for the next few games to show what he has, looks the more likely prospect at the minute, no one else on the horizon who can improve us.

[DonnyOsmond]

Decent article here which I think answers my questions.

https://jobsinfootball.com/blog/what-is-expected-goals-xg/#:~:text=xG%20is%20calculated%20using%20a,scale%20between%200%20and%201.

Key sections:
One thing that's important to note is that different xG models can be used by different organisations and competitions. Each model has its own characteristics, although they generally all rely on the same major factors: distance to goal, angle to goal, body part with which the shot is taken, plus the type of assist or prior action (eg. cross, through-ball, set-piece, short pass, dribble etc.) Models use all the information they have on shots with similar characteristics to come up with a mathematical value relating to how much a player would be expected to score the relevant chance.

...

There are limitations in terms of the data available. For instance, there's a lack of information on the exact state of play when a shot is taken. As the years go on and the data continues to get better, these limitations will gradually be removed or at least reduced.



That's kind of what I thought. So Miller's miss against Rochdale will have been graded as "20 yards out, shot, player running with ball". For which I'd guess you'd expect a goal maybe once in ten? Whereas in fact it was an absolutely open net and if taken with the left foot, you'd expect a goal 19 times out of 20.

Similarly with the Sutton chance. Shot from cross in front of goal 4 yards out might be xG=0.3. But factor in the goalkeeper being 6 yards off to the side, no defender either in front to block or making a challenge and it would be xG=0.95.

I guess the idea is that over a season, these factors even out. I'm not sure that's the case with Miller though, where he has frequently created or got on the end of really good chances and fluffed them because of the left foot thing. I'd say with the chances he's made, I'd expect a decent finisher to be scoring every other game (xG = 0.5).

It depends on the site, StatsBomb are probably the market leaders, their xG takes into account defenders/goalkeepers positioning, others will too. StatsBomb also offer post shot xG figures such as shot velocity which can show how good a player is at finishing.

[BillyStubbsTears]

Decent article here which I think answers my questions.

https://jobsinfootball.com/blog/what-is-expected-goals-xg/#:~:text=xG%20is%20calculated%20using%20a,scale%20between%200%20and%201.

Key sections:
One thing that's important to note is that different xG models can be used by different organisations and competitions. Each model has its own characteristics, although they generally all rely on the same major factors: distance to goal, angle to goal, body part with which the shot is taken, plus the type of assist or prior action (eg. cross, through-ball, set-piece, short pass, dribble etc.) Models use all the information they have on shots with similar characteristics to come up with a mathematical value relating to how much a player would be expected to score the relevant chance.

...

There are limitations in terms of the data available. For instance, there's a lack of information on the exact state of play when a shot is taken. As the years go on and the data continues to get better, these limitations will gradually be removed or at least reduced.



That's kind of what I thought. So Miller's miss against Rochdale will have been graded as "20 yards out, shot, player running with ball". For which I'd guess you'd expect a goal maybe once in ten? Whereas in fact it was an absolutely open net and if taken with the left foot, you'd expect a goal 19 times out of 20.

Similarly with the Sutton chance. Shot from cross in front of goal 4 yards out might be xG=0.3. But factor in the goalkeeper being 6 yards off to the side, no defender either in front to block or making a challenge and it would be xG=0.95.

I guess the idea is that over a season, these factors even out. I'm not sure that's the case with Miller though, where he has frequently created or got on the end of really good chances and fluffed them because of the left foot thing. I'd say with the chances he's made, I'd expect a decent finisher to be scoring every other game (xG = 0.5).

It depends on the site, StatsBomb are probably the market leaders, their xG takes into account defenders/goalkeepers positioning. They also offer post shot xG figures such as shot velocity which can show how good a player is at finishing.

I'd be very interested to see what they had Miller's xG figures to be for the Sutton and Rochdale away games then.

[DonnyOsmond]

Decent article here which I think answers my questions.

https://jobsinfootball.com/blog/what-is-expected-goals-xg/#:~:text=xG%20is%20calculated%20using%20a,scale%20between%200%20and%201.

Key sections:
One thing that's important to note is that different xG models can be used by different organisations and competitions. Each model has its own characteristics, although they generally all rely on the same major factors: distance to goal, angle to goal, body part with which the shot is taken, plus the type of assist or prior action (eg. cross, through-ball, set-piece, short pass, dribble etc.) Models use all the information they have on shots with similar characteristics to come up with a mathematical value relating to how much a player would be expected to score the relevant chance.

...

There are limitations in terms of the data available. For instance, there's a lack of information on the exact state of play when a shot is taken. As the years go on and the data continues to get better, these limitations will gradually be removed or at least reduced.



That's kind of what I thought. So Miller's miss against Rochdale will have been graded as "20 yards out, shot, player running with ball". For which I'd guess you'd expect a goal maybe once in ten? Whereas in fact it was an absolutely open net and if taken with the left foot, you'd expect a goal 19 times out of 20.

Similarly with the Sutton chance. Shot from cross in front of goal 4 yards out might be xG=0.3. But factor in the goalkeeper being 6 yards off to the side, no defender either in front to block or making a challenge and it would be xG=0.95.

I guess the idea is that over a season, these factors even out. I'm not sure that's the case with Miller though, where he has frequently created or got on the end of really good chances and fluffed them because of the left foot thing. I'd say with the chances he's made, I'd expect a decent finisher to be scoring every other game (xG = 0.5).

It depends on the site, StatsBomb are probably the market leaders, their xG takes into account defenders/goalkeepers positioning. They also offer post shot xG figures such as shot velocity which can show how good a player is at finishing.

I'd be very interested to see what they had Miller's xG figures to be for the Sutton and Rochdale away games then.

I have WyScout and if it's the chance at the beginning of the second half that's 0.7, as a team we got 1.3 for the game. I believe Gaz has Instat so his figure may vary from that slightly.

[BillyStubbsTears]

Ah now. Assuming you're referring to the Sutton game, the question then is, is that 0.7 calculated from the chance that it became, when he took the ball with his right foot and brought the defender into play? Or the chance it should have been if he'd simply stepped into the ball with his left and eliminated the defender from the issue?

Seems to me it must be the former. Because the latter really is as easy a chance as a striker could ever hope for. I simply don't believe that an average striker would miss that 3 times in 10.

[DonnyOsmond]

Ah now. Assuming you're referring to the Sutton game, the question then is, is that 0.7 calculated from the chance that it became, when he took the ball with his right foot and brought the defender into play? Or the chance it should have been if he'd simply stepped into the ball with his left and eliminated the defender from the issue?

Seems to me it must be the former. Because the latter really is as easy a chance as a striker could ever hope for. I simply don't believe that an average striker would miss that 3 times in 10.

Miller did miss it, so it's clearly easier to miss then you assume. 70% of the time that chance is scored is a very good percentage. Other than a penalty I doubt you'll get many higher percentage chances.

[BillyStubbsTears]

He missed it because he didn't trust his own left foot. That's my point. He made it into a far harder chance. Just as he did at Rochdale, with that chance that I would have genuinely backed myself at 56 and with a shit left foot to have scored.

Having taken up that position and being prepared to use a left foot, that chance was nigh on impossible to miss.

Not actually impossible of course, because bizarre things happen. But I simply do not believe that a decent Tier 4 striker with a left foot he has any confidence in whatsoever would fail to score 30% of the time in that position.

[drfchound]

Some really amazing players available if ever the forum puts out a football team.
Including a penalty taker who would never miss the target

[Bentley Bullet]

We're missing a trick here. Why don't the scouts look closer to home and give BST a run-out? Either that or see what Daniel Day-Lewis is doing these days.

[BillyStubbsTears]

BB.

I think you missed the bit where I gave praise to Miller's creation of chances. Why doesn't that surprise me?