xG And Analytics Under Amorim

[bosnian_red]

Yeah it's that annoying thread from EtH but we know it'll be discussed a lot under Amorim for everyone looking into every detail. But saw this on Twitter and thought it was worth a post.


Obviously the defensive side was horrendous for us, but was surprised to see we were at the 20th percentile for getting shots off on counters which is extremely anti-United and the opposite of what you'd expect our players should actually be good at. Of course shit set pieces too. Two key areas to focus on for Amorim going in possession.

 

[justsomebloke]

Very interesting. Hoping to see more on this.

 

[Malone_Post]

Sky just said that was..

Ipswich highest XG of the season
United’s 3rd lowest XG of the season
The 3rd lowest United have run all season

Not a great start from the lads. I can’t imagine Ruben is particularly impressed. Especially the running stat.

 

[the hea]

It would be interesting to see the distance covered stats for the two teams today.

 

[red woppit]

Sky just said that was..

Ipswich highest XG of the season
United’s 3rd lowest XG of the season
The 3rd lowest United have run all season

Not a great start from the lads. I can’t imagine Ruben is particularly impressed. Especially the running stat.

Not surprised with the running stat, we had Rashford up front, and Casemiro and Eriksen in midfield, with Evans in the back four, none of them would have moved a tremendous amount, possibly only Rashford, to get on the end of a cross perhaps.
Ipswich's passes were, for the most part, crisp and accurate, our passes very often laboured, with our players waiting for the ball to arrive.

 

[Chipper]

Sky just said that was..

Ipswich highest XG of the season
United’s 3rd lowest XG of the season
The 3rd lowest United have run all season

Not a great start from the lads. I can’t imagine Ruben is particularly impressed. Especially the running stat.

Haha, yeah, was just pulling some of these stats together myself.

xG = Ipswich 1.75 - 0.90 United (Opta)

Ipswich's highest xG of the season so far, narrowly eclipsing the 1.74 they got vs Spurs last time out. Only Aston Villa (0.82) have accumulated less xG against them in the 12 league matches they've played.

0.90 was United's 3rd lowest after Villa (0.52) and Leicester (0.70). It's the 5th most we conceded out of 12 so kind of average for us in that respect. Our precise average is now 1.57.

 

[Chipper]

Posession = Ipswich 40.3% - United 59.7%

That was the 2nd most of the ball we've had in any game this season. Versus Palace away we had 66.8%. We've averaged 52.2% prior to today.

It largely falls in line with Ipswich's typical posession stats. They have averaged 40.8% prior to today.

 

[Borys]

Sky just said that was..

Ipswich highest XG of the season
United’s 3rd lowest XG of the season
The 3rd lowest United have run all season

Not a great start from the lads. I can’t imagine Ruben is particularly impressed. Especially the running stat.

Interesting. I can't imagine we could've run less than today, unless it was a comfortable win.

 

[Chipper]

So I guess in summation - we had a fair bit of the ball by our own standards, but nothing out of the ordinary at all compared to how much ball teams usually have v Ipswich. We weren't much of a goal threat compared to what Ipswich usually give up, or what we usually create.

Passing Network shows that a lot of the ball we had was between the back 3/5 and much of it was deep, rather than passes between players in advanced positions so it's no surprise we didn't create much in the way of chances. We lost the ball a fair bit in the "danger zone", and seeing as we were passing it around there quite a lot that's unsurprsing too, but it didn't really directly lead to many shots/chances for them.

Mazraoui had a very busy game, on the ball a lot for us. Our goal coming from a move that he was involved in at the start.

Morsy the most involved player for them, able to find space in a semi-advanced CM role, slightly skewed off to their right. Passes between him Hutchnison/Tuanzebe/ Burns a common theme as they did most of their attacking down that flank. Their goal coming from a chain of Morsy > Hutchinson > Burns > Hutchinson. Both the passing network graphic and following heatmap (Ipswich on the left) reflect these trends.

Could have garnered all that from watching it but it's good to have it all down.

Graphics from @markrstats / @markstatsbot on 'X' / Twitter (puts out some really cool stuff!) and whoscored.com.

Now let the progress begin!

 

[TMDaines]

So I guess in summation - we had a fair bit of the ball by our own standards, but nothing out of the ordinary at all compared to how much ball teams usually have v Ipswich. We weren't much of a goal threat compared to what Ipswich usually give up, or what we usually create.

Passing Network shows that a lot of the ball we had was between the back 3/5 and much of it was deep, rather than passes between players in advanced positions so it's no surprise we didn't create much in the way of chances. We lost the ball a fair bit in the "danger zone", and seeing as we were passing it around there quite a lot that's unsurprsing too, but it didn't really directly lead to many shots/chances for them.

Mazraoui had a very busy game, on the ball a lot for us. Our goal coming from a move that he was involved in at the start.

Morsy the most involved player for them, able to find space in a semi-advanced CM role, slightly skewed off to their right. Passes between him Hutchnison/Tuanzebe/ Burns a common theme as they did most of their attacking down that flank. Their goal coming from a chain of Morsy > Hutchinson > Burns > Hutchinson. Both the passing network graphic and following heatmap (Ipswich on the left) reflect these trends.

Could have garnered all that from watching it but it's good to have it all down.

Graphics from @markrstats / @markstatsbot on 'X' / Twitter (puts out some really cool stuff!) and whoscored.com.

Now let the progress begin!

Bruno is going to have be a lot more disciplined in keeping his position in the structure going forward.

 

[Chipper]

xG - United 2.93 - 0.83 Bodo Glimt (Opta model)

Posession = 72.4% - 27.6%

Passing network

Obviously that 'umbrella' of the back 3/5 is going to have a lot of the ball in this shape and will be a recurring theme throughout, often passing the ball between themselves. We were able to be a lot higher up the pitch tonight compared to vs. Ipswich in terms of where these passes were being made. They didn't press as much as Ipswich did/we pinned them back a lot more/a bit of both.

Mazraoui again on the ball more than anyone else and we absolutely dominated posession.

"Fields gained" needs explanation on the progression stats, it's a bit convoluted. If you're 50 yards from your opponents goal and pass it 20 yards forward you get 0.4 of a field gained because 20 is 40% of 50. If you were 30 yards from goal and passed it 10 yards forward it's 0.33 of a field gained because 10 is 33.33% of 30. Any forward pass to someone standing on the goal line would be 1 if he's not offside. Add all the passes up at the end and that's the total. A more complicated "progressive pass" stat basically. Think it also subracts for backwards passes. Same formula used on progression via carry and progression received.

Anyway, Maz the highest there with his 9.33 on the passing side of things.

 

[Chipper]

A ton of a stuff to dig into here:



Good win. Quite a lot of positives. Don't want to dwell on it too much, one negative is that "danger zone" losses of the ball. 31 is high, it's about what the worst teams in the league averages for that (Brentford 30.50, Leicester 30.42, Southampton 29.50) and was the same story vs. Ipswich.

Having said that, Everton are actually one of the best teams in the league for winning the ball back themselves. Surprising for a Dyche team to me but it's statistically true.

 

[mav_9me]

A ton of a stuff to dig into here:



Good win. Quite a lot of positives. Don't want to dwell on it too much, one negative is that "danger zone" losses of the ball. 31 is high, it's about what the worst teams in the league averages for that (Brentford 30.50, Leicester 30.42, Southampton 29.50) and was the same story vs. Ipswich.

Having said that, Everton are actually one of the best teams in the league for winning the ball back themselves. Surprising for a Dyche team to me as I'd picture them sitting off and being more passive, but it's statistically true.

Interesting about Everton winning the ball that much.

Surprised the xG being so low, felt like we created quality chances obviously.

Also interesting to see our defensive action height being low like that, compared to how high we were vs Bodo. Might be a reflection of how little Bodo pressed vs Everton. The next progession should be keeping a high line despite being pressed meaning we control the game in the opposition half.

 

[Bright_Eyes]

Yeah it's that annoying thread from EtH but we know it'll be discussed a lot under Amorim for everyone looking into every detail. But saw this on Twitter and thought it was worth a post.


Obviously the defensive side was horrendous for us, but was surprised to see we were at the 20th percentile for getting shots off on counters which is extremely anti-United and the opposite of what you'd expect our players should actually be good at. Of course shit set pieces too. Two key areas to focus on for Amorim going in possession.

The charts are quite biased. Look at shots for example. On a cursory glance it looks like Sporting have double what United have. But when you look at the number, it's 17 vs 14. Same with many other of the stats. It's created to look like the gap is bigger than it is by using scales that don't start at 0.

 

[bosnian_red]

The charts are quite biased. Look at shots for example. On a cursory glance it looks like Sporting have double what United have. But when you look at the number, it's 17 vs 14. Same with many other of the stats. It's created to look like the gap is bigger than it is by using scales that don't start at 0.

Huh? It's a radar chart, the thing to look at is the percentiles your team ranks at and the base value is a certain percentile. No use in starting it at 0, no team is averaging 0 shots per game.

 

[Chipper]

Interesting about Everton winning the ball that much.

Surprised the xG being so low, felt like we created quality chances obviously.

Also interesting to see our defensive action height being low like that, compared to how high we were vs Bodo. Might be a reflection of how little Bodo pressed vs Everton. The next progession should be keeping a high line despite being pressed meaning we control the game in the opposition half.

Quite a wide variance between xG models for this game.

Opta had us lower, 1.20 - 0.65.

Understat = 1.76 - 0.67.

xG philsophy (another twitter guy) and Fotmob extremely low. 0.79 - 0.76 & 0.73 - 0.65. Don't know what's going on there!

Shot count was 11-8 in our favour which is a bit low for a PL game. The average team is having over 13 a game.

There's all those little moments where a shot doesn't happen but you have the ball in a promising situation not picked up by xG. Although that's what xThreat is. Felt we had a few moments like that and that it should have been higher.

Yes, hopefully can we progress towards what you say. That's got to be the ultimate aim.

 

[NZT-One]

Quite a wide variance between xG models for this game.

Opta had us lower, 1.20 - 0.65.

Understat = 1.76 - 0.67.

xG philsophy (another twitter guy) and Fotmob extremely low. 0.79 - 0.76 & 0.73 - 0.65. Don't know what's going on there!

Shot count was 11-8 in our favour which is low for a PL game.

There's all those little moments where a shot doesn't happen but you have the ball in a promising situation not picked up by xG. Although that's what xThreat is. Felt we had a few moments like that and that it should have been higher.

Yes, hopefully can we progress towards what you say. That's got to be the ultimate aim.

Looks quite fair to me. I mean, xG above 1.5 is decent and even though we scored 4 times, there wasn't much apart from that (no criticism). The biggest chances were the goals, all from deep inside the box, yet you'll still "only" get like 0.4 for it. Rashfords shot was borderline hit-and-hope into a packed 6 yrd box, maybe some of the models handle that even more different than Understat (who have that shot as a 0.04) because of the deflection.

My feeling is we have to make sure to find good fits for the WB positions because they'll be key as they seem to be the natural escape route for Amorims idea. One of them should always be free enough to play into and as soon as the rest of the team realizes that there is always somebody, we'll make better use of the space, finally challenge defenses more and everything will settle from there.

 

[Valencia Shin Crosses]

Interesting about Everton winning the ball that much.

Surprised the xG being so low, felt like we created quality chances obviously.

Also interesting to see our defensive action height being low like that, compared to how high we were vs Bodo. Might be a reflection of how little Bodo pressed vs Everton. The next progession should be keeping a high line despite being pressed meaning we control the game in the opposition half.

Yeah Everton were dominating us in duels for a decent amount of that game

 

[bosnian_red]

Quite a wide variance between xG models for this game.

Opta had us lower, 1.20 - 0.65.

Understat = 1.76 - 0.67.

xG philsophy (another twitter guy) and Fotmob extremely low. 0.79 - 0.76 & 0.73 - 0.65. Don't know what's going on there!

Shot count was 11-8 in our favour which is a bit low for a PL game. The average team is having over 13 a game.

There's all those little moments where a shot doesn't happen but you have the ball in a promising situation not picked up by xG. Although that's what xThreat is. Felt we had a few moments like that and that it should have been higher.

Yes, hopefully can we progress towards what you say. That's got to be the ultimate aim.

Fotmob is just being fecky, if you click through our shots it totals higher but their totals aren't updated for some reason. Also Zirkzees goals for example showing extremely low when they should be around 0.3 each if not more

 

[Borys]

We lost the ball quite a few times because of some freak events like sliding tackle and the ball getting stuck between the feet. I think we just need to calm down a bit and we'll be fine.

Of course we're now scoring more than creating what was bound to happen, so Amorim is being lucky in that way, but this is very needed for many players at the moment.

 

[quadrant]

Fotmob is just being fecky, if you click through our shots it totals higher but their totals aren't updated for some reason. Also Zirkzees goals for example showing extremely low when they should be around 0.3 each if not more

It’s done that once before this season, in a European game, can’t remember which one. I believe they get their stats from Opta anyway