Match Stats - Data Perspective

[Borys]

2nd time this season atleast (1st was vs BHA)

Right, I just looked at the graph but it doesn't show first few games. Corrected.

 

[Borys]

RB Leipzig 3-2 Manchester United ( Angeliño 2' Haidara 13' Kluivert 69'; Fernandes (P) 80' Konate (OG) 82')
xG: 1,9 xGA=1,4

• possession 55%, passing accuracy 80% (of 489)
• Shots 9/20, vs 5/10 for RB

Game timeline:
1. RB 1st goal 2' was well worked, great pass, great finish. Wan Bissaka very casual.
2. Matic interception, great pass from Bruno to Greenwood who goes for goal 8'. Big chance missed, Rashford would've been through vs keeper if Greenwood passed the ball. No chemistry?

3. RB score 2nd 13' after a series of mistakes. Firstly weird, very flat back line and AWB very far away from winger (pic rel); next AWB very slow to attack him and weak block; another fantastic first touch finish. RB worked very well to congest us in the middle by passing it right-left-right, they are well coached team.

4. Another big chance for RB 17', same scenario: goal out wide to the left, AWB not closing on winger; first touch cross to the middle, and shot goes just wide.
5. RB shot on post and offside goal 29'.
6. McTominay intercepts the ball and puts Rashford through, Marcus loses the ball 39'
7. McTominay and Maguire miss the header after beautiful Bruno cross 58'
8. Bruno hitting the bar around 65'
9. RB 3rd goal 69', Maguire leaving the ball for De Gea, and De Gea leaving the ball for Kluivert.
10. The rest is just a mess.

Groundhog Day? We concede early, we have multiple chances to score and miss all of them, opposition kills us.

Usual conclusions:

• we were all over the place defensively,
• conceded first two big chances RB had (33% conversion rate)
• everything goes past De Gea these days (1/4 shots saved)
• we had chances to score and even make it level, couldn't convert (10% conversion rate)

Unusual conclusions:

• AWB was absollutely terrible defensively
• Alex Telles had 52% of passing accuracy, was also dispossessed 2x, miscontrolled 1x. I am not a fan of him, glad to have Shaw back.

Controversial opinion:
I though our CBs were fine, maybe Maguire was at fault for the third goal but funny thing is, Leipzig CBs conceded a penalty kick and lost the ball to McTominay who created a good chance for Rashford. Our fullbacks were utter shit though.

One thing that I noticed is it seemed like there was a huge gap between back 3/5 and front 3. I'd like to understand what was the idea behind that formation and what our midfielders were supposed to be doing, because in certainly looked like they were coming wide.

Matic on the left, McTominay to the right. Who was supposed to be in the middle?

 

[Borys]

Manchester United 0 - 0 Manchester City,
xG=0,59 xGA=1,28 (although it seems a bit low, fbref says we got xG=1,0 and xGA 1,2 which seems about right)

• possession 46%, 84% passing accuracy (of 524),
• Shots on Target 1/11 (vs 2/9 for City)
• DDG with 2/2 saves

I was surprised with this to be honest, I don't see Rashford shot when released by Pogba on the map, and it seems like Mahrez and De Bruyne chances were a bit over-rated because there was a lot of players blocking the goal. Our best chance was probably when McTominay missed the ball - somehow it wasn't even shown in highlights (?) and is not included here - big weakness of analysing xG.

• Although there was only one like-for-like sub made (Greenwood->Martial), we didn't play with any set formation. Maguire and McTominay played on the same height (left and right respectively). I am a big fan of Scott on the right of back 3, he seems more natural in that position. He also had 88% passing accuracy (3rd best after Lindelof and Fred). Fred played as single pivot which doesn't suit him IMO, as his movement is very restricted (we should be going with Matic in that position IMO). Pogba and Bruno played all over the pitch so my best guess it was a 5-1-2-2 formation.

• no Cavani - no focal point, which has been also discussed on MOTD:

• Maguire and Lindelof won all headers (8/8). Maguire had a good game defensively, but misplaced a lot of passess (76% passing accuracy is very low for a defender). Lindelof on the other hand had the highest passing accuracy (94%).
• It's quite clear how high Ole rates Fred, another game where he could be substituted and he made him play to the end. He wasn't involved much in terms of passing, pressing and defensive actions - but hard to say whether due to injury, or due to restricted role of a single pivot.
• City didn't trouble us in the air, and they were close to conceding a goal on set piece on a couple occasions. I see this as a big weakness for them which can be exploited easily by all EPS teams. I am surprised then concede so little goals (8 clean sheets in last 10 games)
• City look good whenever De Bruyne or Jesus get the ball, the rest of the team is a bit meh. I'd say we're even with them in terms of quality, although they have more squad depth up front.
• I guess you could make a funny compilation of us trying to beat the City pressing. The idea was too obvious IMO, and I don't understand why we didn't want to use Pogba as target man to make it more "differential".

Overall I made a point in another thread that Ole is happy with a dull, goalless draw against top sides if it ends without injuries and not a lot of effort put in, which is what happened yesterday for both sides. If that's the strategy, we will be able to judge it one month from now. We will play Sheffield, Leeds, Leicester, Wolves, Villa, Watford and Fulham by then, so I believe the table won't be so "flat" anymore.

EDIT: I forgot to add, as one of the biggest critics, I was impressed by Pogba. No wonder he plays good when he doesn't have a set position and when he doesn't have the time to think.

 

[Physiocrat]

Good stuff again @Borys

Can you access an average position map for all players and translate it into a formation image? I've seen some shows do that and Zonal Marking does it. That would give the best impression of the formation. For what it's worth, MOTD had us as a 433 as Bruno as the central striker. When I saw that I was dubious unless they thought Bruno was a false 9

 

[Borys]

Good stuff again @Borys

Can you access an average position map for all players and translate it into a formation image? I've seen some shows do that and Zonal Marking does it. That would give the best impression of the formation. For what it's worth, MOTD had us as a 433 as Bruno as the central striker. When I saw that I was dubious unless they thought Bruno was a false 9

Found it:

I prefer looking at heatmaps, they give a better insight. For example McTominay average position yesterday was pretty much the same to what he usually does, BUT it felt like he moved around much less than usually. Same applies for Fred. Heatmaps confirm that.

Best example is how similar positions Cavani and Rashford get (on average), even if they are moving around in a completely different way.

 

[Physiocrat]

Found it:

I prefer looking at heatmaps, they give a better insight. For example McTominay average position yesterday was pretty much the same to what he usually does, BUT it felt like he moved around much less than usually. Same applies for Fred. Heatmaps confirm that.

Best example is how similar positions Cavani and Rashford get (on average), even if they are moving around in a completely different way.

Great stuff. I agree heatmaps give you more detail but for a one image it gives you a nice idea of what the setup is. It also shows how narrow the attack, it is like a narrow 424 with no pronounced CF.

It also shows how wide the CMs are which is what you commented on against PSG. If you continue with these posts it would be much appreciated if you put the average position in too. It will help a general understanding of the way the team plays

 

[P-Nut]

That image is like the opposite of what you'd teach kids playing football, wide at the back and compact up top.

 

[Physiocrat]

That image is like the opposite of what you'd teach kids playing football, wide at the back and compact up top.

True, but TBF we had a decent amount of possession so the back four being wider isn't that bad. What you would want to see is the average position in and out of possession. That would give you a better idea. That said the attack is clearly really narrow

 

[Borys]

True, but TBF we had a decent amount of possession so the back four being wider isn't that bad. What you would want to see is the average position in and out of possession. That would give you a better idea. That said the attack is clearly really narrow

It would be perfect if you could at least set a time for which you want to see heatmap, or like you said, on/off the ball for each team. Keeping all that in mind, we just need to be very careful drawing conclusions based on average position.

 

[Physiocrat]

It would be perfect if you could at least set a time for which you want to see heatmap, or like you said, on/off the ball for each team. Keeping all that in mind, we just need to be very careful drawing conclusions based on average position.

True. Let's hope they bring out an in and out of possession heat map. It shouldn't be difficult given they have the heatmaps to begin with.

 

[UNITED ACADEMY]

That image is like the opposite of what you'd teach kids playing football, wide at the back and compact up top.

The main reason why our defense was stretched wide yesterday because it seems Bissaka was instructed to press Cancelo when City had possession which forced Lindelof to go wider to watch Sterling when Bissaka decided to press cancelo.

 

[Borys]

The main reason why our defense was stretched wide yesterday because it seems Bissaka was instructed to press Cancelo when City had possession which forced Lindelof to go wider to watch Sterling when Bissaka decided to press cancelo.

It certainly looked like that was happening, then McTominay dropped to fill that void between Lindelof and Wan Bissaka and we played back 3 some times. I have to admit I like that transition.

 

[Sunny Jim]

Good stuff again @Borys

Can you access an average position map for all players and translate it into a formation image? I've seen some shows do that and Zonal Marking does it. That would give the best impression of the formation. For what it's worth, MOTD had us as a 433 as Bruno as the central striker. When I saw that I was dubious unless they thought Bruno was a false 9

Aktualny i think Bruno would be a gratis false 9. And it would allow to accomadate our midfield stars.

 

[Green_Red]

@Borys where do you get your data for analysis?

 

[Borys]

@Borys where do you get your data for analysis?

See the OP. Ask if there's something more specific.

 

[Green_Red]

See the OP. Ask if there's something more specific.

Thanks

 

[Borys]

So you manually compile the stats? I was thinking maybe you downloaded them as json via a web api or something. I've been looking for a good free stats api for ages for the Premier league but have never found one.

I use Power BI for most stats which I update every week like xG to compare teams, goalkeeper stats, and so on. Using fbref tables is very easy to connect as data source. So it's only "click refresh button" job for me, once the query and modeling is done. But after the game, I look for some interesting stats regarding each player.

 

[Physiocrat]

Aktualny i think Bruno would be a gratis false 9. And it would allow to accomadate our midfield stars.

Yeah, that might work although I'm not sure Martial, Greenwood or Rashford are the best off the ball runners needed for a false 9

 

[Borys]

What happens if possession-based team faces counter-attacking side? We've seen an excellent example yesterday with Liverpool hosting Tottenham. If you ever wondered how xG timechart would look like for that game, here it is:

Tottenham scored 1 goal from 3 big chances they had (or should I say all chances). Liverpool scored 2 from multiple half-chances (11 shots on goal which is A LOT for any EPL game, we average 5-6 per game). Very interesting game to watch.

 

[Borys]

Sheffield United 2 - 3 Manchester United (McGoldrick 5' 87'; Rashford 26' 51' Martial 33')
xG=1,79 xGA=1,24

• possession 60%
• shots on target 8/14
• saves 4/6

• I was surprised to see the xG so low, it's probably because Rashford chance (offside) and Martial chance (he wasn't quick enough to take a shot) were not counted in
• 4 lucky goals, overall we made it difficult for ourselves (as we tend to do) with Henderson error.
• midfield duo let Sheffield just run with the ball at times. Basically all typical issues, Matic dribbled past, Pogba not tracking back or giving the ball away. Second tier midfield it is for me, unaccepteable on any serious level. Lucky enough, SHU are were not a serious threat.
• Matic had a poor game in general, 0/2 successful tackles, dribbled 2x, dispossessed 3x, although high passing completed % and 75% headers won. Pogba was most useful for long passing and defensive headers. His pirouettes are effective in 30%, some think it's enough to make him first choice midfielder, I think it's just enough for Sheffield level opponents. He did have a good game defensively though.
• it was the first time this season gk error led directly to a goal
• SHU scored 2 goals for the first time this season (!)
• our approach was quite interesting, a lot of movement up front and some good runs from Martial and Rashford. It was pretty much our main and only tactic in the first half. Sheffield played risky high line, which was weird as we were caught slightly offside 4/5 times before Rashford scored so they had a lot of time to change shape. As I've mentioned many times, lack of movement up front seemed to be more a result of instructions rather than "laziness" of our forwards. If you compare our play to Liverpool or Tottenham, we tend to play it safe (seems like we are more afraid to lose the ball)
• we conceded from the corner, again. I can't find a stat where I can compare number of games from set pieces, so if anybody knows please PM me.
• I thought Lindelof was OK, won more aerials than Maguire (5 - 71% vs 4 - 57%), more tackles (4 vs 3), and more clearances (4 vs 3). Harry had 2 interceptions so I guess that makes it pretty even. I don't think Victor got bullied last night as some suggested on the forum.
• Early days for Telles, but looks... well, "weak" springs to mind. Lowest pass success on the field (71%, last game also lowest 52% against RB), won 43% of aerial duels. I am not particlarly impressed by any of his attributes including crossing (good technique but random aim), although it might be more effective with Cavani on the pitch.
• Special mention for Marcus Rashford, for who I think it was first really good game. He converted 2/4 chances he had, made a couple of good runs, seemed like a real threat and very fit. One thing I noticed that on 2nd goal he didn't go for power, like he usually does. For some players it means gaining some confidence, we'll see if I'm correct on that. Martial is still out of form though, got a little bit lucky with the goal.
• Greenwood was OK, he's getting sharp and by far the best option on the right wing, but he needs to learn how to cross and when to pass. He's too focused on taking a shot even when there are better options around. Same applies for Mata, both could've set Bruno for a great chance.
• I liked how Greenwood and Fernandes operated, it seemed like we had the correct balance with some threat from the right.

PS
Stats on fbref are not updated so I don't have all data, will update tomorrow if something interesting pops out.

 

[Borys]

Manchester United 6 - 2 Leeds (McTominay x2, Bruno x2, Lindelof, James; Cooper, Dallas)
xG=4,78 xGA=1,63 (fbref says 3,8 which is second highest this season for EPL)

• After last game vs Sheffield I said I was surprised they didn't react to our through ball attempts after we were caught offside 4 times. Yesterday Leeds basically made the same mistakes again and again, leaving holes in midfield. Bruno was dragging Leeds DM all over the pitch, leaving space for others to run into.
• proper box to box performance from McTominay, most distance covered with the ball running forward (Maguire is usually 1st by some distance); 2/3 succesful dribbles; 2nd highest passing accuracy (85%); interesting that he didn't have a good game defensively looking at the stats.

By the way, I always said we're not playing double pivot but rather two box to box midfielders. Neither of Fred/McTominay really feel comfortable sitting deep.

• Bruno has been decent again Personally I think he's worth sticking to 4-2-3-1, interesting to see how centrally he played against Leeds:

• One worrying thing is how Leeds targeted Wan Bissaka - seems like they crossed the ball into his area on purpose (7 aerial duels, lost 4; for comparison Shaw had only 1 all game). This is our weak spot. He had the highest number of touches which is surprising, it seems we prefer playing from the back through the right side (was also visible against City). He also made a bad pass which led to dangerous situation at least a couple of times.
• We are still missing big chances, Martial especially. Cavani should've done better as well.
• We are still conceding goals from corners.

Regarding our last discussion, funny how average position graph looks for Leeds team as from what I remember, attacked mainly through the wings:

This week it's also time to give Miss of the Week Award, which for the second time this season goes to Anthony Martial. Anthony not once, but twice missed a great opportunity to score so he's earned it:

I think United janitor needs to take a close look at that spot on the grass!

 

[sebsheep]

Manchester United 6 - 2 Leeds (McTominay x2, Bruno x2, Lindelof, James; Cooper, Dallas)
xG=4,78 xGA=1,63 (fbref says 3,8 which is second highest this season for EPL)

• After last game vs Sheffield I said I was surprised they didn't react to our through ball attempts after we were caught offside 4 times. Yesterday Leeds basically made the same mistakes again and again, leaving holes in midfield. Bruno was dragging Leeds DM all over the pitch, leaving space for others to run into.
• proper box to box performance from McTominay, most distance covered with the ball running forward (Maguire is usually 1st by some distance); 2/3 succesful dribbles; 2nd highest passing accuracy (85%); interesting that he didn't have a good game defensively looking at the stats.

By the way, I always said we're not playing double pivot but rather two box to box midfielders. Neither of Fred/McTominay really feel comfortable sitting deep.

• Bruno has been decent again Personally I think he's worth sticking to 4-2-3-1, interesting to see how centrally he played against Leeds:

• One worrying thing is how Leeds targeted Wan Bissaka - seems like they crossed the ball into his area on purpose (7 aerial duels, lost 4; for comparison Shaw had only 1 all game). This is our weak spot. He had the highest number of touches which is surprising, it seems we prefer playing from the back through the right side (was also visible against City). He also made a bad pass which led to dangerous situation at least a couple of times.
• We are still missing big chances, Martial especially. Cavani should've done better as well.
• We are still conceding goals from corners.

Regarding our last discussion, funny how average position graph looks for Leeds team as from what I remember, attacked mainly through the wings:

This week it's also time to give Miss of the Week Award, which for the second time this season goes to Anthony Martial. Anthony not once, but twice missed a great opportunity to score so he's earned it:

I think United janitor needs to take a close look at that spot on the grass!

Would their wide players switching sides mess with the Leed's average positions?

 

[Borys]

Would their wide players switching sides mess with the Leed's average positions?

Yeah for sure. In extreme scenario you can have two conservative wingers staying on the sideline, swithing positions on the halftime and ending up with average pos in the middle. Heatmap would show it correctly.

They can give some indication of how high midfielder played for example but not much more. It can lead to wrong conclusions, I don't use them.

 

[Borys]

(EFL Cup) Everton 0-2 Manchester United (Cavani, Martial)
no xG available, if anybody knows where to find it for EFL please PM me

• 65% possession (!)
• 8/19 shots on target vs 3/7

Probably the best game for us this season, no insight is available so I want to focus on formation which I believe was very interesting. Re-watching the game, I'd represent the formation (while attacking) like below:

• basically we played one man midfield in Pogba. He was positioned between Everton forwards and midfielders, and he had acres of space around. What's interesting is he did a lot of standing still - you can see the dots on the graph which suggest he wasn't very mobile. Compared to Matic, he had significantly less passes (47 to 93).

What struck me is how little we used that to pass between the lines. This is pretty typical formation for us vs Everton:

One quick pass to PP and we're through with 4 players behind him. Like I said, acres acres of space for Pogba and yet we rarely utilized that. It seemed like a good role for Pogba too.

Now one thing is clear to me, we don't pass the ball between the lines and don't play one touch football unless we are ahead and comforteable. Only time I remember us doing it effortlessly was when van de Beek played in midfield against Başakşehir. It was like night and day difference, and surprisingly we never seemed to play him there again. Against Everton he had a role which I can't specify to be honest. It seemed like he moved a lot on up and down on the left wing, however he (just like Pogba) was constantly ignored:

Obviously, in both cases the ball went back to our CB because it seemed like our idea to get the ball forward was to get Maguire/Matic/Bailly run with it. Which again is weird and it suggest we are not coached to pass the ball around or Ole doesn't trust the pass and move style with our players.
The second idea was to pass from defense directly to forwards, which worked well when Martial was on the pitch. Someone called him "attacking pivot", which seems about right. He also receives most line-breaking passes in the league:

• Matic and Bailly played as left/right CB respectively. Matic was stepping up from defense into midfield but you can see his movement was rather constrained too:

• Greenwood had a very difficult task and I believe his effort has been a bit underrated. He was basically up against half the Everton defense with only Tuanzebe and Bruno to support. I think he did well as he stretched back 4 and pulled defenders to the side, which is as good as it gets as far as our RW options are considered.

I'd like us to play that formation again as it seemed to give us total control of the game. I am not sure what was the reason for that, but we looked totally sound defensively and created some good chances. What I'd do is swap van de Beek with Pogba in that role on the left (Donny is no threat outside the box and we already played 2 strikers + Bruno), and use Matic in midfield. He is by far the best suited player to that role of single midfielder as I trust him on the ball and defensively. In that case we can play 3 at the back, including Bailly (who looks comforteable on the ball) or McTominay (so that we can make a quick shift into 3 man midfield).

 

[Borys]

Leicester City 2-2 Manchester United (Barnes, Tuanzebe OG; Rashford, Bruno)
xG=2,2 xGA=1,2

We basically kept them quiet all game (xGA<0,5 until 90th minute!) which is as good as it gets against Leicester away. That said, they work really hard and they've been rewarded (once again it seems).
I might add, although this might be a bit controversial, that I don't blame our back 4 for any of two goals conceded. First should've been either blocked by McTominay or saved by keeper, but it was a very good shot with weeker foot, and I can understand why Scott didn't want to risk a foul just outside the box. Second was just pure class from Vardy (see pic below) and a bit of luck. Maybe Bailly could've done better but it's unrealistic not to concede any half chances against top opposition.

Anyway, we're not here to analyze micro-events so lets get down to business.

• 46% of possession
• 79% of passing accuracy. We tend to have 83-85% so that's below average. Leicester had 78% so basically both teams focused on disturbing opposition play. It was not a high-quality game overall.
• 4/9 shots on target. Decent conversion rate overall for us I believe, Rashford scoring 1/3 is basically his normal service. Martial scored 1/1 but unfortunately for us was slightly offside.
• Leicester had 1 shot on target according to fbref.

Obviously we can moan about Rashford missing two chances as a reason we only got 1 point from that game - but he is not a clinical striker and never was (he is pretty much average finisher basing on np:goals-expected goals; Martial was pretty good on that last couple of seasons but is on negative side so far, which is a big dissapointment).

Now focusing on stats perspective, let's compare Leicester vs other top opponents:

• Leicester (H) vs ManUtd: xG=1,2 Shots on target 1/10 goals:2 (including OG)
• Leicester (A) vs Tottenham: xG=1,5 shots on target 3/16 goals:2 (including OG!)
• Leicester (A) vs Liverpool: xG=1,3 shots on target 4/11 goals:0
• Leicester (A) vs ManCity: xG=2,6 shots on target 4/4 (!) goals:5 (including 1 penalty)

So... we had the best stats in terms of xG and shots on target. Leicester record against top sides is pretty ridiculous really (and we were the only ones who faced them at King Power Stadium)- especially conversion rate is incredible. I can see why people are so annoyed with our defense, but we were extremely unlucky to concede two goals all things considered. It's hardly a reason to turn everything upside down and sell Maguire sack Ole.



In other aspects, it was a scrappy game. Leicester worked very hard to close our front 4. We don't move the ball fast enough to keep it for longer period of time, so not only we had less possession but also our midfielders had low number of touches:

Pretty low %of passes completed too. Leicester DMs (midfielders?) did much better job and I thought they were impressive both on and of the ball:

In general we can compare Leicester and ManUtd as both teams operated in 4-2-3-1 with 2 midfielders sitting and one attacking mid. I'd say pretty even contest between both "midfield 3". Note that the number of defensive actions is twice as high as usually - the game was mainly played in midfield and neither team could really progress forward with ease (this is also confirmed by low number of touches by forwards for each team):


We are deep into the season and I see two patterns emerging in terms of our game plan:
a) counter attacking
b) direct pass from defense to forwards

Point a) is not particlarly interesting as it basically suits our strengths. But b) is something I've noticed recently as we tend to pass the ball between defenders at very slow pace, one striker drops to receive the ball between the lines and create an opening (Cavani goal vs Everton, Bruno goal vs Leicester in last two games). We've seen Cavani doing that yesterday, but Martial is particlarly effective at that:

Martial receives the most line breaking passes in the league. I wish I could check how Cavani and Rashford do in that regard, but I've found only this graph.

To sum it up, although I do not like how our midfield role is very limited (yesterday it was Fred&McTominay, previously it was Pogba as I've described in previous post), it is quite effective. We created enough chances (some due to Fred and McTominay quick passing by the way, they can do it once we intercept the ball in midfield, but it's quite clear they are told to keep it simple in build up play). We also kept Leicester very quiet all game (again, 1 shot on target all game). If it wasn't for Tuanzebe unlucky deflection at the end, we would be praising Ole for great tactics, moaning we could've buried them with 2 more goals.

I believe we're heading in the right direction, unlucky draw yesterday but overall it's hardly a reason to panic over our defense. In next days I will investigate deeper why De Gea is second worst goalkeeper in the league judging by % of shots saved. But I'd like to understand how expected goals for the team this and previous seasons work so need more time on that.

I'd like to give Rashford another Miss of the Week Award for this shot which he somehow was able to put over the bar:

Rashford has joined Martial at the top with 2 awards each, following by Bruno with 1 as he's dropped off recently. To be honest I'd bet on Rashford winning this competition by some margin as he tends to go for power more often than not, which is not as effective as choosing a corner and sliding it across keeper but lets wait and see.

EDIT:
As I've just learned from @sullydnl the stats on quality of chances conceded (in De Gea thread) are from fantasyfootball site, so unless he wants to comment on that, I'll try to find it on fbref.

 

[Borys]

Manchester United 1-0 Wolverhampton Wanderers (Rashford 90+')
xG=1,48 xGA=0,41

• 58% possession, 85% passing accuracy
• only 3/11 shots on target
• De Gea made 5/5 saves - mostly straight at him but it's an improvement for sure

We used similar formation to the Everton game (even heatmaps look similar for each players, only Rashford didn't hold the width as Donny did). It seems to be the way to get more "control" of the game which is widely considered on this forum as "not giving the ball away". Too bad we actually didn't really know what to do with it and it felt like Rashford and Bruno did an awful lot of running/sprints. The first half we were quite good I thought, one big chance (spike on graph) created by Greenwood but not converted by Bruno.
In a way we did control the game more because Wolves xG was the lowest so far this season (Chelsea game excluded), although it's hard to say if that was because of our setup or they were simply exhausted due to a short break.

• usual tactic of 4231 when defending, switching to 3 at the back + wide fullbacks when attacking (Matic dropping into defense). We use it whenever we play second string team (Telles, Matic, Pogba). I do like that approach although I still don't quite understand what is Pogba role as one man midfield.
• Rashford played van de Beek role of forward dropping deeper to collect the ball. He's been quite effective in that role and likes freedom (11 attempted dribbles, 4 successful).
• I've been very impressed with Matic, who is by far our best midfielder. Obviously his passing stats will be pumped by being our defensive pivot but he did well to pass his way out of trouble on a couple occasions (94% completed passes). Good defensive stats too.
• Paul Pogba on the other hand did not have a good game in my opinion. He had 4 miscontrols (highest for both teams), and disspossesed 2x (highest for United). I might be a bit harsh on him (as I'm not his fan) but I really don't understand why is he still playing so much in midfield, considering a) he will be leaving soon and b) he's not even that good in there.
• In my opinion we should be shooting from distance more often. We are not coached to the level we will create clear chances with small play, and considering we change 7 players between games, it'll be very hard to learn. Small number of shots is a bit worrying and I feel like we have players who can hurt opposition from distance. Low number of shots is something we need to fix, this is not the way to play against opposition sitting deep, especially when we have 2/3 strikers lurking in the box.
• I still maintain Greenwood is our best option on the right, he created our best goal scoring chance. He's not at his best clearly, probably lacking some confidence but I wonder if fitness somehow isn't an issue, he has rather low number of touches (35; 24 passes) so not quite sure he can never finish a game.

We have a number of options in midfield, I'd strongly suggest using Matic as first-team midfielder. Personally I don't see why people say "his legs are gone", while I see him running as fast as anyone, keeping up with Traore. He's perfectly fine when playing once per week. It would help us a lot with keeping the ball in midfield and we can pair him with Fred, which would be the safest bet.

If Ole is adamant at playing back 3, we can also just use McTominay/Fred as part of back 3 and shift Matic higher up to the position Pogba's been playing in that setup. We definitely want Matic more involved in the buildup play, not only starting from the back 3. I'd love to see him with van de Beek (which would be by far our most advanced technically midfield duo on paper), but at this point I don't see Donny featuring in midfield unless there is an injury.

 

[Chesterlestreet]

This is a fine thread @Borys - just wanted to say that.

More of this, less of...a lot of other things.

Anyway, I largely agree with your Matic point - at least in theory. The problem is that Matic has shown, pretty much conclusively, that the level of his performances tend to drop when he is used frequently over time. It seems to me to be a case of starting him very carefully - i.e. NOT to treat him as a default starter.

In a 4-2-3-1 where one of the CM pair drops down, I'd choose Matic as the default starter every time (him being the one dropping down) - and I would also choose either Fred or McTom over Pogba per default.

But "default" doesn't say much - Matic can't play as much as certain other players who are also "default" starters based on ability/top level.

 

[Borys]

This is a fine thread @Borys - just wanted to say that.

More of this, less of...a lot of other things.

Anyway, I largely agree with your Matic point - at least in theory. The problem is that Matic has shown, pretty much conclusively, that the level of his performances tend to drop when he is used frequently over time. It seems to me to be a case of starting him very carefully - i.e. NOT to treat him as a default starter.

In a 4-2-3-1 where one of the CM pair drops down, I'd choose Matic as the default starter every time (him being the one dropping down) - and I would also choose either Fred or McTom over Pogba per default.

But "default" doesn't say much - Matic can't play as much as certain other players who are also "default" starters based on ability/top level.

It might be true in general, but at this point Matic is performing as good as ever and this month he's played:

• 8 Dec 60' vs RB
• 17 Dec 90' vs Sheffield
• 23 Dec 90' vs Everton
• 29 Dec 90' vs Wolves

Ole squad management has been good this season, and so far it seems like Matic can play full game once per week no problem. Also, we have troubles in early stages of the game when opposition is pressing us hard, so I believe we should consider this even if he can only play first 60' as he would really help us with keeping the ball.

I wish 5 subs would be available again - probably it would've solved this issue for us.

 

[anant]

Manchester United 1-0 Wolverhampton Wanderers (Rashford 90+')
xG=1,48 xGA=0,41

• 58% possession, 85% passing accuracy
• only 3/11 shots on target
• De Gea made 5/5 saves - mostly straight at him but it's an improvement for sure

We used similar formation to the Everton game (even heatmaps look similar for each players, only Rashford didn't hold the width as Donny did). It seems to be the way to get more "control" of the game which is widely considered on this forum as "not giving the ball away". Too bad we actually didn't really know what to do with it and it felt like Rashford and Bruno did an awful lot of running/sprints. The first half we were quite good I thought, one big chance (spike on graph) created by Greenwood but not converted by Bruno.
In a way we did control the game more because Wolves xG was the lowest so far this season (Chelsea game excluded), although it's hard to say if that was because of our setup or they were simply exhausted due to a short break.

• usual tactic of 4231 when defending, switching to 3 at the back + wide fullbacks when attacking (Matic dropping into defense). We use it whenever we play second string team (Telles, Matic, Pogba). I do like that approach although I still don't quite understand what is Pogba role as one man midfield.
• Rashford played van de Beek role of forward dropping deeper to collect the ball. He's been quite effective in that role and likes freedom (11 attempted dribbles, 4 successful).
• I've been very impressed with Matic, who is by far our best midfielder. Obviously his passing stats will be pumped by being our defensive pivot but he did well to pass his way out of trouble on a couple occasions (94% completed passes). Good defensive stats too.
• Paul Pogba on the other hand did not have a good game in my opinion. He had 4 miscontrols (highest for both teams), and disspossesed 2x (highest for United). I might be a bit harsh on him (as I'm not his fan) but I really don't understand why is he still playing so much in midfield, considering a) he will be leaving soon and b) he's not even that good in there.
• In my opinion we should be shooting from distance more often. We are not coached to the level we will create clear chances with small play, and considering we change 7 players between games, it'll be very hard to learn. Small number of shots is a bit worrying and I feel like we have players who can hurt opposition from distance. Low number of shots is something we need to fix, this is not the way to play against opposition sitting deep, especially when we have 2/3 strikers lurking in the box.
• I still maintain Greenwood is our best option on the right, he created our best goal scoring chance. He's not at his best clearly, probably lacking some confidence but I wonder if fitness somehow isn't an issue, he has rather low number of touches (35; 24 passes) so not quite sure he can never finish a game.

We have a number of options in midfield, I'd strongly suggest using Matic as first-team midfielder. Personally I don't see why people say "his legs are gone", while I see him running as fast as anyone, keeping up with Traore. He's perfectly fine when playing once per week. It would help us a lot with keeping the ball in midfield and we can pair him with Fred, which would be the safest bet.

If Ole is adamant at playing back 3, we can also just use McTominay/Fred as part of back 3 and shift Matic higher up to the position Pogba's been playing in that setup. We definitely want Matic more involved in the buildup play, not only starting from the back 3. I'd love to see him with van de Beek (which would be by far our most advanced technically midfield duo on paper), but at this point I don't see Donny featuring in midfield unless there is an injury.

While I'm a big fan of xG, this game perfectly highlights the shortcomings of that. Bruno+ Cavani shot as per understat got a combined xG of .86. With Bruno shot getting 0.45 and Cavani shot getting xG of 0.74. I don't have an issue with the math or anything here, but the fact that Cavani's shot was weighed at 0.74. Sure he was 5 yards out, with just a defender to beat, but the model fails to take into consideration his initial position which was back to goal

 

[Borys]

While I'm a big fan of xG, this game perfectly highlights the shortcomings of that. Bruno+ Cavani shot as per understat got a combined xG of .86. With Bruno shot getting 0.45 and Cavani shot getting xG of 0.74. I don't have an issue with the math or anything here, but the fact that Cavani's shot was weighed at 0.74. Sure he was 5 yards out, with just a defender to beat, but the model fails to take into consideration his initial position which was back to goal

I don't quite understand your point. In the end it was a header few yards from goal and more centrally than Bruno. In my opinion fair rating. Or am I missing something?

 

[Pagh Wraith]

I don't quite understand your point. In the end it was a header few yards from goal and more centrally than Bruno. In my opinion fair rating. Or am I missing something?

He's talking about the rebound after Fernandes' shot was saved where Cavani was facing the goal with his back. Something Understat's very basic model fails to account for. As for his other chance, a header can practically never have a 0.74 rating (pundits and commentators constantly overrate the chance of headers being scored). Here's the simplest of all xG models to play around with: Torvaney | xG

You can see how drastically the values diminish when you set header to yes.

 

[Borys]

He's talking about the rebound after Fernandes' shot was saved where Cavani was facing the goal with his back. Something Understat's very basic model fails to account for. As for his other chance, a header can practically never have a 0.74 rating (pundits and commentators constantly overrate the chance of headers being scored). Here's the simplest of all xG models to play around with: Torvaney | xG

You can see how drastically the values diminish when you set header to yes.

Right, I see it now. I never realized that was even a shot

One of xG weaknesses then for sure. Just like shot on target with many bodies blocking the sight, I believe it should be rated pretty low too. I remember City had a couple of highly rated shots like that.

 

[Pagh Wraith]

Right, I see it now. I never realized that was even a shot

One of xG weaknesses then for sure. Just like shot on target with many bodies blocking the sight, I believe it should be rated pretty low too. I remember City had a couple of highly rated shots like that.

Not an xG problem per se. Statsbomb's model accounts for all of that. Number of defenders, pressure on the shooter, even the goalkeeper's vision cone and height of the ball when hit.

 

[Borys]

Not an xG problem per se. Statsbomb's model accounts for all of that. Number of defenders, pressure on the shooter, even the goalkeeper's vision cone and height of the ball when hit.

Nice. There will always be some exceptions, but current model is pretty good already for some insight. I wish they could add some parameters like highest xG per shot or average xG from all shots, combined with number of shots you could see how the team plays.

 

[anant]

I don't quite understand your point. In the end it was a header few yards from goal and more centrally than Bruno. In my opinion fair rating. Or am I missing something?

Yeah, I'm referring to the rebound, not the 1st header. That should ideally be assigned a xG for an overhead kick rather than a shot from foot.

Not an xG problem per se. Statsbomb's model accounts for all of that. Number of defenders, pressure on the shooter, even the goalkeeper's vision cone and height of the ball when hit.

Not sure if goalkeeper's vision are taken into account or not. Although, it seems they do take into account deflections (as Rashford has an xG of 0.3 on statsbomb compared to 0.1 on understat)

Think Statsbomb have given that shot a value of ~0.3 (his total is 0.6, and assuming the header earlier in the game is having similar value across the 2 models).

 

[Chesterlestreet]

I wish 5 subs would be available again - probably it would've solved this issue for us.

Damn right.

It's ridiculous.

 

[Obiorahking_]

Thoughts on build up on left vs right?

 

[Borys]

Thoughts on build up on left vs right?

Pretty much the same as always, more attacks coming down the left side but I'd say that's mostly because of Rashford who has his favourite "starting position" there. On the right Greenwood was moving a lot rather than standing, but also he had relatively low number of touches. What's interesting is heatmaps show a lot of action on the wings but no focal points which suggest Cavani was moving a lot as well.

Looking at Wolves heatmap it certainly looks like they've preferred right side so that would be a factor also.

 

[Pagh Wraith]

@Borys Here's a screenshot of the match report from xgstats.com where every shot is watched and recorded manually (by a human).



I think that's more in line with reality having watched the match.

 

[wolvored]

Personally I think Xg is a nonsense way to analyse individual chances. No two chances are ever the exact same.

There are too many variables in football for this to be the basis for an accurate measurement of whether or not each chance should be converted.

I know f all about xg and for me this is the most sensible post on here.
Fergie being interviewed after we beat Arsenal. "Sir Alex, Arsenal had 65% possession." "Aye and we won 4-2" Only stat that matters.