Not Luck After All

[ucandomagic]

Infogol Early XG - Leeds 0.43 - Blades 0.72

Wrong again Mr Bielsa - you said analysis of the score would show that one team won but analysis of the data would say different.

Expected Magic as well as Real Magic!

Edit: XG was 1.18-0.98 at the end ! Still close.

 

[DronnieBlade]

XG stuff all very interesting.

The KEY differential yesterday (and Tuesday) was the clear superiority of our defensive shape, pattern, and sheer determination to keep yet another clean sheet.

My cup runneth over with admiration for the way in which the whole team has defended as a unit when required to do so. It’s an incredible springboard to success when defensive solidity can be relied upon, week in week out.

To think that in recent weeks we have gone away to ALL 3 of the realistic auto promotion candidates and come away undefeated with 7 points is a phenomenal effort at the business end of the season.

UTB

 

[WorkingtonBlade]

I think XG is an interesting and relevant statistic, but for this game I really struggle to match the XG figures being quoted with the match I just watched.

By my reckoning Leeds had 5 clear cut chances against 2 for us, and they had more half-chances too. I would have expected Leeds to score a lot more goals from their chances than I’d expect us to score from ours.

 

[Houston_Blade]

I think XG is an interesting and relevant statistic, but for this game I really struggle to match the XG figures being quoted with the match I just watched.

By my reckoning Leeds had 5 clear cut chances against 2 for us, and they had more half-chances too. I would have expected Leeds to score a lot more goals from their chances than I’d expect us to score from ours.

I'd say in terms of 'clear cut chances' it was 2 a piece. Ours being the 2 Basham chances and theirs being the one that hit the post and the one that was volleyed over the bar from 8 yards.

Apart from that they had more half chances but we had the one where the keeper was sent off and McGoldrick's header. XG doesn't really tell the story of yesterday though I agree on that.

 

[Chali 2na]

I think XG is an interesting and relevant statistic, but for this game I really struggle to match the XG figures being quoted with the match I just watched.

By my reckoning Leeds had 5 clear cut chances against 2 for us, and they had more half-chances too. I would have expected Leeds to score a lot more goals from their chances than I’d expect us to score from ours.

In my mind, we had 3: Basham twice and McGoldrick header from a corner.

They had 2: Jansson's chance at the end and Taylor hitting the post.

The difference is that they had other massive oppurtunities to create a clear cut chance. There were pull backs which fell to Alioski and Hernandez where the wrong decision was made.

 

[HodgysBrokenThumb]

For me it was all to do with spygate. Bielsa had Utd watched all week, and Basham never scored from more than 2 yards out, so they left him to it. Masterful double-bluff, or just Bash in training?

 

[Beans]

In my mind, we had 3: Basham twice and McGoldrick header from a corner.

They had 2: Jansson's chance at the end and Taylor hitting the post.

The difference is that they had other massive oppurtunities to create a clear cut chance. There were pull backs which fell to Alioski and Hernandez where the wrong decision was made.

The Tyler Roberts chance that hit the post was actually really low quality - Infogol reckons only 4% of shots of that type from that location result in a goal (so 0.04xG).

Jansson’s one also limited by the way the ball came to him (didn’t have time to control) and the amount of defenders around him. Might also have been on his weaker foot.

Agreed Leeds had more bits of dangerous play that they’ll feel they should have turned into proper chances, but as you say we were in on goal twice and had a close range header from Mcgoldrick.

Edit - 3 times in on goal if you include the red card.

 

[Bath Blade]

The Tyler Roberts chance that hit the post was actually really low quality - Infogol reckons only 4% of shots of that type from that location result in a goal (so 0.04xG).

Jansson’s one also limited by the way the ball came to him (didn’t have time to control) and the amount of defenders around him. Might also have been on his weaker foot.

Agreed Leeds had more bits of dangerous play that they’ll feel they should have turned into proper chances, but as you say we were in on goal twice and had a close range header from Mcgoldrick.

Edit - 3 times in on goal if you include the red card.

Appreciate this is a question without a simple response (apologies if it has been covered in the pod), but who decides on the XG of a particular chance, as a lot of it seems subjective, yet there is clearly somebody able to assess that a particular chance may have an XG of 0.27, for example.

 

[HodgysBrokenThumb]

The Tyler Roberts chance that hit the post was actually really low quality - Infogol reckons only 4% of shots of that type from that location result in a goal (so 0.04xG).

Jansson’s one also limited by the way the ball came to him (didn’t have time to control) and the amount of defenders around him. Might also have been on his weaker foot.

Agreed Leeds had more bits of dangerous play that they’ll feel they should have turned into proper chances, but as you say we were in on goal twice and had a close range header from Mcgoldrick.

Edit - 3 times in on goal if you include the red card.

I agree with the analysis of the Roberts chance. Given where Egan (I think) was standing, and Henderson’s position, there was little space for that shot to go in. Good move and shot, though.

 

[Beans]

Appreciate this is a question without a simple response (apologies if it has been covered in the pod), but who decides on the XG of a particular chance, as a lot of it seems subjective, yet there is clearly somebody able to assess that a particular chance may have an XG of 0.27, for example.

The xG models compile data of thousands of shots and basically work out how often each shot is scored from each location and each type (header / volley / controlled pass) and amount of defensive pressure etc. So if a chance has an xG of 0.27 in your example, that means that historically 27% of other shots taken from that spot have been scored.

Our own Blades Analytics wrote an explainer here:

https://www.uptheblades.com/latest-articles/2018/8/22/explaining-expected-goals

 

[georgebernardshaw]

Can someone please explain what the fuck this thread’s all about?

 

[PokerBlade]

Can someone please explain what the fuck this thread’s all about?

Sheffield United in being objectively better team shocker.

 

[Chali 2na]

The Tyler Roberts chance that hit the post was actually really low quality - Infogol reckons only 4% of shots of that type from that location result in a goal (so 0.04xG).

That's astonishing, 4% seems incredibly low!

 

[Houston_Blade]

The Tyler Roberts chance that hit the post was actually really low quality - Infogol reckons only 4% of shots of that type from that location result in a goal (so 0.04xG).

Jansson’s one also limited by the way the ball came to him (didn’t have time to control) and the amount of defenders around him. Might also have been on his weaker foot.

Agreed Leeds had more bits of dangerous play that they’ll feel they should have turned into proper chances, but as you say we were in on goal twice and had a close range header from Mcgoldrick.

Edit - 3 times in on goal if you include the red card.

Brentford and Leeds have both done pretty much the same from the same position in the last week so it can't be that easy as you say.

 

[georgebernardshaw]

Sheffield United in being objectively better team shocker.

Objectively?? This is outrageous, and I’m outraged in a mature citizen sort of outrage.....

 

[Houston_Blade]

Appreciate this is a question without a simple response (apologies if it has been covered in the pod), but who decides on the XG of a particular chance, as a lot of it seems subjective, yet there is clearly somebody able to assess that a particular chance may have an XG of 0.27, for example.

Beans and Blades Analytics sit down after the game over a steaming hot Bovril and make this shit up so it gives them something to talk about on their podcast.

 

[Pinchy]

XG stuff all very interesting.

The KEY differential yesterday (and Tuesday) was the clear superiority of our defensive shape, pattern, and sheer determination to keep yet another clean sheet.

My cup runneth over with admiration for the way in which the whole team has defended as a unit when required to do so. It’s an incredible springboard to success when defensive solidity can be relied upon, week in week out.

To think that in recent weeks we have gone away to ALL 3 of the realistic auto promotion candidates and come away undefeated with 7 points is a phenomenal effort at the business end of the season.

UTB

We don’t get results against the top six you know. Never have; Never will.

 

[Pinchy]

The xG models compile data of thousands of shots and basically work out how often each shot is scored from each location and each type (header / volley / controlled pass) and amount of defensive pressure etc. So if a chance has an xG of 0.27 in your example, that means that historically 27% of other shots taken from that spot have been scored.

Our own Blades Analytics wrote an explainer here:

https://www.uptheblades.com/latest-articles/2018/8/22/explaining-expected-goals

Stop loitering on here and get on with the fucking podcast!

I’m not a statto but this feels like the meanest defensive run ever. The nearest thing to it that I can recall was the brief Vonk/Ablett period under Kendall when we conceded about 4 goals in 13 games or something similar.

7 clean sheets in a row is amazing, heroic, soul-satisfying defending.

 

[SUFC Guy]

Beans and Blades Analytics sit down after the game over a steaming hot Bovril and make this shit up so it gives them something to talk about on their podcast.

That may be very true and I appreciate that your post may be tongue in cheek but I do love listening to those two fellas. Always a good sensible and fair opinion with just a small handful of stats chucked in there for good measure

Keep it up fellas

UTMB!

 

[Broomhall Blade]

In my mind, we had 3: Basham twice and McGoldrick header from a corner.

They had 2: Jansson's chance at the end and Taylor hitting the post.

The difference is that they had other massive oppurtunities to create a clear cut chance. There were pull backs which fell to Alioski and Hernandez where the wrong decision was made.

Doesn't Sharp's count as a clear cut chance ? If he had not been rugby tackled by their keeper(still can't believe he had the nerve to protest) it was an open goal.
It was far more clear cut than McGoldricks header and Basham's chance for me anyway.

 

[blade too long]

I think XG is an interesting and relevant statistic, but for this game I really struggle to match the XG figures being quoted with the match I just watched.

By my reckoning Leeds had 5 clear cut chances against 2 for us, and they had more half-chances too. I would have expected Leeds to score a lot more goals from their chances than I’d expect us to score from ours.

I counted 4 for us . A goalline scramble 1st half and basham had a similar opportunity end of the first half as well as his goal and Sharp getting denied a certain goal by the red card

A) the stats you can analyse fall short of Leeds o sufc 1 . The dominant staf

 

[Beans]

That's astonishing, 4% seems incredibly low!

Yeah I thought that too! Turns out scoring goals is really hard

Looking at it again though, it's a first-time shot (makes it harder as doesn't have the ball under control), the angle isn't ideal (quite wide rather than central) and you've got defenders quite close to him. I think if he'd scored we'd be saying it was a fantastic finish.

 

[Chali 2na]

Doesn't Sharp's count as a clear cut chance ? If he had not been rugby tackled by their keeper(still can't believe he had the nerve to protest) it was an open goal.
It was far more clear cut than McGoldricks header and Basham's chance for me anyway.

Yes it probably does actually, I'd forgotten about that

 

[Chali 2na]

Yeah I thought that too! Turns out scoring goals is really hard

Looking at it again though, it's a first-time shot (makes it harder as doesn't have the ball under control), the angle isn't ideal (quite wide rather than central) and you've got defenders quite close to him. I think if he'd scored we'd be saying it was a fantastic finish.

View attachment 50973

Having seen that picture, 4% makes more sense now. I think my mind was playing tricks on me.

 

[Maidenhead]

I'd say in terms of 'clear cut chances' it was 2 a piece. Ours being the 2 Basham chances and theirs being the one that hit the post and the one that was volleyed over the bar from 8 yards.

Apart from that they had more half chances but we had the one where the keeper was sent off and McGoldrick's header. XG doesn't really tell the story of yesterday though I agree on that.

And the free header from about 8yds in the first half that flashed past the post?

 

[Tom Bott]

I’m not a statto but this feels like the meanest defensive run ever

If United beat Bristol City and keep a clean sheet on the 30th March, it will beat (equalled last night) a 34-year-old second-tier record of seven straight home league wins without conceding a goal set by Portsmouth in the 1985-86 season. This run started on 20th August 1985 and ended on 23rd November 1985 with a 0-3 home defeat by...Sheffield United.

Apologies for forgetting who posted this earlier

So yes our defence is magnificent.

UTB.

 

[Houston_Blade]

And the free header from about 8yds in the first half that flashed past the post?

Still quite a hard chance I'd say. It was a half chance in my estimation, hard to get the ball on target with pace from that position.

 

[BushBlade]

Having seen that picture, 4% makes more sense now. I think my mind was playing tricks on me.

Not just you. All of us. That’s why I like XG. We all have a perception of how easy a chance is and we have a tendency to overestimate it. Our brain tells us ‘should have finished that’ whilst ignoring how many times we’ve seen similar chances missed.

I remember reading a basic analysis years ago that only one in three one on ones with the keeper resulted in a goal. I found that staggering- I was brought up watching Keith Edwards, I think I saw him miss one. If he got one on one, he scored. I just assumed that was normal. So I started paying more attention when watching other games and realised that apart from players like Edwards and Lineker, strikers did miss one on ones. Quite often. But still pundits would always say ‘he should’ve finished that’.

 

[BushBlade]

Still quite a hard chance I'd say. It was a half chance in my estimation, hard to get the ball on target with pace from that position.

All the stats indicate that headed goals are much harder than they look.

 

[Houston_Blade]

All the stats indicate that headed goals are much harder than they look.

I just use the benchmark "would I have scored that?". In 99.9% of occasions the answer is "yes" but I was like a mix between Maradona, Messi and Pele so my bar is set high