Midfield Creativity, Stats and Comparisons

Ekeke

Full Member
Joined
Aug 1, 2006
Messages
48,832
Location
Hope, We Lose
There seems to be an awful lot of players missing

De Bruyne is the highest from CM. Gross from Brighton is 2nd, Klitch at Leeds is 3rd, Gundogan 4th, Bernardo Silva 5th

Pogba 6th, Moutinho 7th, Thiago 8th, Tielemans 9th, Curtis Jones 10th

Fred would be 14th by my count just behind Kovacic in 13th

via https://fbref.com/en/comps/9/gca/Premier-League-Stats
 

andersj

Nick Powell Expert
Joined
Aug 7, 2004
Messages
2,690
Location
Copenhagen
"In modern football the game module is now changing its function. From a static disposition of the players we are arriving at a dynamic occupation of the positions functional to the principles of the game model «

As we will see below, the meaning of role is changing in modern football. It is no longer a fixed position that identifies the characteristics of a player, but more and more the different functions and therefore the tasks that a player performs in the match identify him. Therefore, the characteristics of the players are enhanced through the tasks he is called upon to perform."


This is an extract from Pirlos Pro Lisence thesis. I think it is quite relevant today when we end up comparing KdB, a player with very little responsibility for defending, to more classic midfielders.
 

justsomebloke

Full Member
Joined
Oct 25, 2020
Messages
1,136
Lastly I also looked at the more defensive stats. The first graph is successfully winning the ball in four different ways - interceptions, tackles won, dribblers tackled (which seems to be counted separately than other tackles) and successful pressing.



Here you can see Ndidi, Kante and Fred as expected in the top 3, but then Jorginho perhaps surprisingly in 4th ahead of Allan.

Other than Fred, our other three midfielders are right down near the bottom. McTominay is 20th, Pogba 21st and Matic 23rd out of the 26 players I checked. Liverpool, City and Chelsea all don’t have a single midfielder who wins the ball back as little as any of our three. I think most people expect Matic and Pogba to be near the bottom but it’s surprising that McTominay is that low. Considering he’s supposed to be one half of our hard-working defensively solid midfield pairing, he’s not actually doing as much defending as expected.

It does appear that most teams have one midfielder that wins it back a fair bit more than his partner. Kante or Jorginho with Kovacic, Ndidi with Tielemans, Hojbjerg with Sissoko, Allan with Doucoure, and to a lesser extent Rice with Soucek. I guess that’s the balance we have when Fred’s in there, but we probably rotate more than any of them so when he’s not playing we do appear to miss that ball winner. Liverpool (who have the highest average), City and Wolves spread it out a bit more.

On a related note here’s successful and unsuccessful pressing as a total, as even ‘unsuccessful’ pressing probably forces the opponent to pass it backwards or sideways half of the time so it’s still impacting the game.



Allan comfortably out in front with the amount of attempts, Fred comfortably in second, while Jorginho and Ndidi have slightly more successful attempts than the top two due to a better success rate. After that there seems to be two groups of players. One, led by Henderson, Phillips and Kante, that still does a lot of pressing but not quite to the level of the top few, and another group that are a fair bit behind. Our three of McTominay (18th), Matic (23rd) and Pogba (24th) are right down the bottom of the second group.

I also checked the success rate for different types of tackles, largely because I wanted to see if Fred and Pogba were as low with their attempts as I think many expected (Fred because he attempts so many and goes in too hard, Pogba because of defensive laziness).



Fred is on the low end of things, but not by as much as expected. Most players seem to have a press success of 30-34%, so Fred at 29.8% is only just behind. Allan (the only player to attempt more than him) has a lower rate with only 28.8%. General tackles and tacking dribblers are similar - probably on the lower side but not particularly low (ahead of both Matic and McTominay at both). Interestingly Pogba has the highest success percentages for all three categories for our team.

----------------

Overall

Going purely by the stats, the three stand outs seem to be Thiago, Jorginho and Fred for a decent spread of creativity, control and defensive work. Which could be used to argue that these stats mean jackshit, as Thiago is struggling to fit into Liverpool, Jorginho was arguably behind Kante and Kovacic in the pecking order under Lampard (although he’s been a starter so far for Tuchel), and some of our own fans firmly dislike Fred. On the other hand it does agree with my ‘eye test’ that Thiago is actually doing quite well on the individual level for 80% of the time but he just doesn’t suit how Liverpool are currently playing (and the other 20% is him making mistakes and being brushed aside), Jorginho is a good player if they set things up properly which they may be starting to do, and while Fred isn’t perfect he is important as our best all-rounder.

From outside the ‘big four’ of Man Utd, Man City, Chelsea and Liverpool (the teams that are financially strong enough to refuse any realistic bid), the Wolves duo of Moutinho and Neves seem to be the best all-rounders. I’m always surprised that Moutinho never played for any of the top clubs, he feels too good of a midfielder to have spent his career in Portugal, France and now Wolves for the last few years of his career. If we could get him for cheap I’d still be tempted to sign him even though he’s 34. I’m actually a bit surprised Neves is up there as I thought he was struggling a bit this season and had never really played to his potential (maybe harsh considering he’s still only 23).

Other decent all-rounders seem to be Partey, Hojbjerg and perhaps Phillips.

Tielemans and Ndidi have an interesting yin and yang dynamic, with Ndidi being arguably the best defensively while the worst creatively of all the midfielders, and Tielemans next to him pulling up good creative stats but one of the lowest defensively. Of course his ‘poor’ defending stats are still similar to McTominay, Pogba and Matic for us. If he got a move to one of the top teams would his creativity decrease without Ndidi covering his back, or would it increase with better players in front of him? Although Leicester would argue they are one of the top teams now...

On the negative side of things, McTominay is surprisingly low on most of these stats. I knew he was on the lower side compared to our own midfielders but hadn’t really compared him to players from the other clubs before. I’m hoping that now that he seems to be getting back into good form (after 1-2 months of poor to average form) we’ll see some of these stats improve. Of course stats don’t win games, so if he does enough to help the team put the results on the board that’s the main thing. Some of the stats do agree with the eye test that he could improve how much he gets on the ball and then what he does with it though.

Lastly, what the hell is Soucek supposed to be good at? He scores a decent amount of goals and is good in the air, but otherwise he’s a stand out for all the wrong reasons in these stats. In the first category of actions leading to a shot he’s 2nd worst only to Ndidi, he has significantly the lowest distance passed (both in general and forward) and ball carries forward, he attempts the least amount of passes and has by far the worst pass success rate of them all, and his total defensive wins is the third worst. I haven’t actually taken any note of him the few times I’ve watched West Ham play, is he actually as bad as what these stats make him look? It seems to be working for them with their excellent season so far, but damn.

Ok, now that I’ve done this wall of text, I expect everyone to ignore this thread and it will die with no responses at all. :lol: There's no real rhyme or reason to most of it, just stuff that I looked up as I wondered what it would look like.
Really really interesting, many thanks for taking the time to write this. It offers statistical substance to the intuitive judgment that it is both very doable and potentially very impactful to improve our central midfield pairing.
 

MadDogg

Full Member
Joined
Apr 24, 2002
Messages
12,423
Location
Manchester Utd never lose, just run out of time
My only point is the selection of statistics here will favour two things. First the players who are designated playmakers in their teams. And second they will favour possession-based teams rather than those who have a more transitional style. So for example, McTominay doesn't come out great in the first case and the West Ham players don't come out that well in the second. There are no real off-the-ball statistics and I presume that's because of what is available?

Given players are only on the ball for something like 2 - 4 minutes of the 90 in every match, so much of the game is about what they do off the ball in creating space and tracking runners. The Athletic did a piece recently on Lingard's physical efforts this season - sprints, top speed, high intensity runs, accelerations, distance covered, etc - which helps to show his contributions and why he has found himself in goalscoring positions so often. I imagine McTominay would come out quite well on some of those metrics. And probably the same for Soucek, although I can't admit to have watched that much of West Ham this year.
Correct. FBref doesn't really have off-the-ball stats, and I remember trying to find somewhere that did have total distance covered recently but couldn't find any. It seemed to me that those kind of stats are hidden behind pay walls.

There seems to be an awful lot of players missing

De Bruyne is the highest from CM. Gross from Brighton is 2nd, Klitch at Leeds is 3rd, Gundogan 4th, Bernardo Silva 5th

Pogba 6th, Moutinho 7th, Thiago 8th, Tielemans 9th, Curtis Jones 10th

Fred would be 14th by my count just behind Kovacic in 13th

via https://fbref.com/en/comps/9/gca/Premier-League-Stats
This was specifically aimed just at deeper midfielders so the likes of De Bruyne and Silva wouldn't be included. Basically players who play in a midfield two, or as one of the deeper two in a midfield three. Attacking midfielders are obviously expected to have much different stats, likely much more creative but much less defensive work (or their defensive work will be much higher up in the field) so doing a direct comparison doesn't really work. I was uncertain whether to include Gundogan as City play that specific formation with only one real deeper midfield, but if I do this again at the end of the season I might include him as he probably is the more defensive of the two attacking ones.

Pogba has played quite a few games in one of the more attacking positions so when I did these stats I manually went through and worked everything out by only using the games he played in midfield. And in general I basically just picked the players from the top teams plus a couple of others that I was interested in seeing how they compared (Phillips, Neves, Moutinho and Ward-Prowse). If I do it again I've already realised I need to add Bissouma and Andre Gomes, but if there are any other noteworthy midfielders that people want to add I can do that as well.
 

Ekeke

Full Member
Joined
Aug 1, 2006
Messages
48,832
Location
Hope, We Lose
Correct. FBref doesn't really have off-the-ball stats, and I remember trying to find somewhere that did have total distance covered recently but couldn't find any. It seemed to me that those kind of stats are hidden behind pay walls.


This was specifically aimed just at deeper midfielders so the likes of De Bruyne and Silva wouldn't be included. Basically players who play in a midfield two, or as one of the deeper two in a midfield three. Attacking midfielders are obviously expected to have much different stats, likely much more creative but much less defensive work (or their defensive work will be much higher up in the field) so doing a direct comparison doesn't really work. I was uncertain whether to include Gundogan as City play that specific formation with only one real deeper midfield, but if I do this again at the end of the season I might include him as he probably is the more defensive of the two attacking ones.

Pogba has played quite a few games in one of the more attacking positions so when I did these stats I manually went through and worked everything out by only using the games he played in midfield. And in general I basically just picked the players from the top teams plus a couple of others that I was interested in seeing how they compared (Phillips, Neves, Moutinho and Ward-Prowse). If I do it again I've already realised I need to add Bissouma and Andre Gomes, but if there are any other noteworthy midfielders that people want to add I can do that as well.
Oh okay. There arent that many that play that role then I guess, but yeah City do to an extent. I think theres another element which is whether a team is trying to have most of the ball or hit teams on the break. As if you're going to hit teams on the break then your DM is going to release the ball earlier to the attacking players. That part makes sense with Fred for me, theres a decent amount of the time where when we win the ball he will find a pass in the buildup for us to then counter from deep. But as I read it the stat says a player can be awarded both of the "2 actions before a shot" right? So if Fred passes the ball to Rashford on the halfway line and he runs it from there, then eventually goes past a player and then shoots - Rashford would be awarded 1 of those actions for the one successful dribble before the shot, but Fred's pass way earlier before all that work would be considered the 2nd action?

It seems like a really general stat if thats the case and I'm not sure what it really says when it can be something completely rudimentary from your own half and then another player drives up the pitch, beats a player and then shoots.

Because how often do players actually go past 2 players before shooting in order to be credited with both those actions themselves? A lot of the time defenders are backing off scared to make the challenge and get it wrong, or the player running with the ball moves the ball in a direction away from the player rather than directly dribbling them. So some players who are considered really good dribblers dont beat as many players with "dribbling" stats as we might think. For example our highest is Rashford with 2.4 dribbles per game and then its Pogba with 1.8. In order to be given both points for the SCA that would have to be 2 in a single run before shooting. So whoever passed the ball to them on the halfway line is probably going to end up with credit. Then again that would also count as a key pass so I guess thats not much more useful.
 

justsomebloke

Full Member
Joined
Oct 25, 2020
Messages
1,136
@MadDogg

I thought that an interesting implication of your analysis was that it seems to indicate that it's more the creative side of the midfield pairing that needs upgrading than the defensive side. Would you agree that's a valid conclusion?
 

TheMagicFoolBus

Full Member
Joined
Nov 5, 2016
Messages
3,017
Location
Lisboa, Portugal
Supports
Chelsea
This was specifically aimed just at deeper midfielders so the likes of De Bruyne and Silva wouldn't be included. Basically players who play in a midfield two, or as one of the deeper two in a midfield three. Attacking midfielders are obviously expected to have much different stats, likely much more creative but much less defensive work (or their defensive work will be much higher up in the field) so doing a direct comparison doesn't really work. I was uncertain whether to include Gundogan as City play that specific formation with only one real deeper midfield, but if I do this again at the end of the season I might include him as he probably is the more defensive of the two attacking ones.

Pogba has played quite a few games in one of the more attacking positions so when I did these stats I manually went through and worked everything out by only using the games he played in midfield. And in general I basically just picked the players from the top teams plus a couple of others that I was interested in seeing how they compared (Phillips, Neves, Moutinho and Ward-Prowse). If I do it again I've already realised I need to add Bissouma and Andre Gomes, but if there are any other noteworthy midfielders that people want to add I can do that as well.
Brilliant post mate - planning a refresh end of season by any chance?

Would be curious about Wijnaldum and Ndombele actually - though the latter might be difficult as I believe he's played as a 10 a fair bit this year.
 

MadDogg

Full Member
Joined
Apr 24, 2002
Messages
12,423
Location
Manchester Utd never lose, just run out of time
Oh okay. There arent that many that play that role then I guess, but yeah City do to an extent. I think theres another element which is whether a team is trying to have most of the ball or hit teams on the break. As if you're going to hit teams on the break then your DM is going to release the ball earlier to the attacking players. That part makes sense with Fred for me, theres a decent amount of the time where when we win the ball he will find a pass in the buildup for us to then counter from deep. But as I read it the stat says a player can be awarded both of the "2 actions before a shot" right? So if Fred passes the ball to Rashford on the halfway line and he runs it from there, then eventually goes past a player and then shoots - Rashford would be awarded 1 of those actions for the one successful dribble before the shot, but Fred's pass way earlier before all that work would be considered the 2nd action?

It seems like a really general stat if thats the case and I'm not sure what it really says when it can be something completely rudimentary from your own half and then another player drives up the pitch, beats a player and then shoots.

Because how often do players actually go past 2 players before shooting in order to be credited with both those actions themselves? A lot of the time defenders are backing off scared to make the challenge and get it wrong, or the player running with the ball moves the ball in a direction away from the player rather than directly dribbling them. So some players who are considered really good dribblers dont beat as many players with "dribbling" stats as we might think. For example our highest is Rashford with 2.4 dribbles per game and then its Pogba with 1.8. In order to be given both points for the SCA that would have to be 2 in a single run before shooting. So whoever passed the ball to them on the halfway line is probably going to end up with credit. Then again that would also count as a key pass so I guess thats not much more useful.
The way I look at it is that it should be taken in the same kind of way as an assist. Some assists/key passes/SCA will be something that any random player will do. A very easy pass to a player who then does all the work to actually take the shot. So picking out an individual match and saying he got 1 in that match doesn't mean much. However over a longer period it should average out so that players with a higher number will have 'deserved' it. For instance quite a few of Bruno's assists are either lucky (I remember two where he's been stretching and just managed to get a slight touch that has taken it to the scorer) or quite easy. However he obviously deserves to have a ridiculous amount of assists, and indeed the act of being in a position to have that lucky/easy contribution is a skill in itself. SCA will be the same. The longer the period the stats are from the more accurate they'll be. With Fred I do expect that his stats will have dropped if I do them again at the end of the season. He was already a bit higher in the creative stats than I would have expected, and he's now been fairly average in the second half of the season (with the occasional great match and the occasional terrible match thrown in).
 

Ekeke

Full Member
Joined
Aug 1, 2006
Messages
48,832
Location
Hope, We Lose
The way I look at it is that it should be taken in the same kind of way as an assist. Some assists/key passes/SCA will be something that any random player will do. A very easy pass to a player who then does all the work to actually take the shot. So picking out an individual match and saying he got 1 in that match doesn't mean much. However over a longer period it should average out so that players with a higher number will have 'deserved' it. For instance quite a few of Bruno's assists are either lucky (I remember two where he's been stretching and just managed to get a slight touch that has taken it to the scorer) or quite easy. However he obviously deserves to have a ridiculous amount of assists, and indeed the act of being in a position to have that lucky/easy contribution is a skill in itself. SCA will be the same. The longer the period the stats are from the more accurate they'll be. With Fred I do expect that his stats will have dropped if I do them again at the end of the season. He was already a bit higher in the creative stats than I would have expected, and he's now been fairly average in the second half of the season (with the occasional great match and the occasional terrible match thrown in).
I think Fernandes means those ones though. He gets a slight touch into a danger area expecting the striker to be there
 

Desert Eagle

Punjabi Dude
Joined
Sep 25, 2006
Messages
10,118
Great work Maddog. Appreciate the time and effort put into this. Shows what a lot of people have been saying; Pogba is incredible but defensively suspect, fred is solid if not spectacular and mctominay is a clear weakness.
 

MadDogg

Full Member
Joined
Apr 24, 2002
Messages
12,423
Location
Manchester Utd never lose, just run out of time
@MadDogg

I thought that an interesting implication of your analysis was that it seems to indicate that it's more the creative side of the midfield pairing that needs upgrading than the defensive side. Would you agree that's a valid conclusion?
Going purely off stats I would say that the main thing I see is that we only have one good all-rounder who scores high in most of the main attributes - Fred. The other three are all quite low on the defensive side of things, and two of them are also quite low on the creative side (Pogba being the exception there of course).

I'm not sure if my conclusion is based purely on stats or whether it's biased because it's how I feel in general from just watching us. But I would say that I feel we need another good all-rounder. Somebody who is better defensively than Pogba, Matic and McTominay (Scott is debatable as it 'feels' like he does more than what is shown in the stats), but who also contributes more in a creative and playmaking sense than the latter two. Even if they aren't as good creatively as Pogba (not many are), if they are better defensively they then should be able to play that position all the time instead of having to be moved further forward like we have to with Pogba due to his defensive liability. Somebody who would end with a similar stat spread as Jorginho would be ideal in theory IMO.

Obviously stats are only half the story though, and I feel what I wrote above is pushing the limits of usefulness. In general I do feel we are missing somebody who is solid at both defensive and creative side of the game though, and at the moment we are constantly having to decide what we're going to sacrifice each match. We're having to sacrifice too much defensive stability to get creativity, or sacrificing too much creativity to get defensive stability. There should be a better balance than what we currently have available.

Brilliant post mate - planning a refresh end of season by any chance?

Would be curious about Wijnaldum and Ndombele actually - though the latter might be difficult as I believe he's played as a 10 a fair bit this year.
I do hope to do it again at the end of the season. It took quite a long time to do it originally but that's mostly because I kept having to constantly redo players over and over as I decided to add new stats. Now I've got the format (and I realised I can use the Scouting Report option on FBref to get the stats a lot faster) it should be easier.

I actually did include Ndombele in my personal stats a month or so after posting these, going through and using only his central midfield performances. Defensively he was about mid-range (ever so slightly below Moutinho), slightly above average for shot-creating actions (once again about the same as Moutinho if you removed the latter's set pieces). Quite low with Key Passes, Expected Assists and Passing Distance (both total and forward) though. Interestingly he did have the highest (in terms of percentage) amount of short passes and, alongside Doucoure, the lowest amount of long passes.

Wijnaldum is another I wasn't sure whether to include at the time as he was, while not an actual attacking midfielder, probably the most attacking of the three man Liverpool midfield. I might add him though as in any other team he'd probably just be one of the deeper ones, and indeed he has been playing that role more often this season.
 

TheMagicFoolBus

Full Member
Joined
Nov 5, 2016
Messages
3,017
Location
Lisboa, Portugal
Supports
Chelsea
I do hope to do it again at the end of the season. It took quite a long time to do it originally but that's mostly because I kept having to constantly redo players over and over as I decided to add new stats. Now I've got the format (and I realised I can use the Scouting Report option on FBref to get the stats a lot faster) it should be easier.

I actually did include Ndombele in my personal stats a month or so after posting these, going through and using only his central midfield performances. Defensively he was about mid-range (ever so slightly below Moutinho), slightly above average for shot-creating actions (once again about the same as Moutinho if you removed the latter's set pieces). Quite low with Key Passes, Expected Assists and Passing Distance (both total and forward) though. Interestingly he did have the highest (in terms of percentage) amount of short passes and, alongside Doucoure, the lowest amount of long passes.

Wijnaldum is another I wasn't sure whether to include at the time as he was, while not an actual attacking midfielder, probably the most attacking of the three man Liverpool midfield. I might add him though as in any other team he'd probably just be one of the deeper ones, and indeed he has been playing that role more often this season.
Makes sense! And actually I could have my timelines muddled but I think that scouting report option may have been released subsequent to your original post? Feel like it's only a couple months old.

Also very interesting to think about ways to visualise the passing profiles you've provided in terms of the breakdown between short / medium / long passes & their success rate, especially if chances can be worked in there. Maybe something like average distance per pass plotted against chances created would give a decent picture of creativity & degree of difficulty? Would be nice to work in pass success rate somehow though. I'm quite curious at the similarity in passing profile between Kovacic and Rodri actually! Wouldn't necessarily have expected that given the eye test personally.
 

hungrywing

Full Member
Joined
Jul 6, 2009
Messages
9,069
Location
Your Left Ventricle
Saw this a few days ago; didn't quite know where to put it.


EDIT: There's another thread analyzing DMs. It seems there's a disturbing(?) trend of Declan Rice not being a standout performer according to a lot of these analyses.
 

UNITED ACADEMY

Full Member
Joined
Aug 14, 2018
Messages
8,468
Great work Maddog. Appreciate the time and effort put into this. Shows what a lot of people have been saying; Pogba is incredible but defensively suspect, fred is solid if not spectacular and mctominay is a clear weakness.
Anyone who pay attention on the game can see it. It's just people tend to think if you are scoring goals mean you are better thus people tend to think McTominay is better.

However, this is down to coaching as well and playing style. Ole's playing style is direct, forward and likes to attack quickly, unlike some managers prefer possession. I could argue that some of those players in stats would improve in their forward passing and progressive ball if they play under Ole right now.
 

TrustInJanuzaj

Full Member
Joined
Mar 26, 2015
Messages
5,795
I’m not surprised that McTominay performed badly statistically, he’s still the weak link in this team. He tends to get lumped in with Fred which is more unfair on the Brazilian than anything because he performs a certain job to a very high level. McTominay has had a decent season but he just doesn’t have the talent to be anything more than a squad option here. The goals have clouded some opinions too. Yes it’s nice that a midfielder chips in but if it’s to the detriment of the core elements of his game who really cares?
 

TheMagicFoolBus

Full Member
Joined
Nov 5, 2016
Messages
3,017
Location
Lisboa, Portugal
Supports
Chelsea
Saw this a few days ago; didn't quite know where to put it.


EDIT: There's another thread analyzing DMs. It seems there's a disturbing(?) trend of Declan Rice not being a standout performer according to a lot of these analyses.
Interesting as well. And not surprised about Rice; apparently Lampard was borderline obsessed with signing him to the point where it became genuinely aggravating for Marina et al - Chelsea's analysts & scouts didn't look at him favourably either.
 

Desert Eagle

Punjabi Dude
Joined
Sep 25, 2006
Messages
10,118
Anyone who pay attention on the game can see it. It's just people tend to think if you are scoring goals mean you are better thus people tend to think McTominay is better.

However, this is down to coaching as well and playing style. Ole's playing style is direct, forward and likes to attack quickly, unlike some managers prefer possession. I could argue that some of those players in stats would improve in their forward passing and progressive ball if they play under Ole right now.
I agree style is part of it. Mctominay, James, AWB and Maguire are part of a pattern. Van der beek who doesn't fit the pattern has struggled to get playing time. This is why people myself included disagree with a lot of Ole's choices because in modern football unless you're a specialist you have to be very good at most aspects of the game. I dislike Jordan Henderson as much as any united fan but him and fletcher are the types of players we need. Solid all rounders who have very good football iq and very high intangibles. Thats the only way you get by without elite skill in the modern game at the highest level.
 

tjb

Full Member
Joined
Sep 6, 2013
Messages
1,961
One thing not being highlighted here but is significant is the quality of the teams they play for. Higher quality teams attain more possession and opportunities for ball progression. In addition, we have less 50/50 type of games as teams tend to sit back when we play at home, which allows some of our players the opportunities to move the ball more than their peers and hence statistically it appears higher.

Mctominay hides during games and it reflects even worse here because it indicates that despite having more opportunities to use the ball, his statistical figures are still low.

Fred's problem is that despite being a good pressing midfielder, he is not so good technically. His passing range outside of the area around him is poor, his first touch is poor, he can't drive with the ball and isn't strong enough to hold the ball or barge an opponent off the ball. If you watch Chelsea, Kante may not be the most complete midfielder, but he is at least average in all those facets of the game. In addition, despite the fact that Fred can get interceptions and occassionally is able to win the ball in a 50/50 duel, his tackling technique is poor aswell, add that to his lack of physical strength, and you have a player that can't really be relied upon to win back possession without support, unlike Kante. Finally, positioning isn't a factor that can easily be measured by stats. Fred's positioning when the opponent has a counter attacking opportunity is pretty poor and is the core reason why most wouldn't call him a defensive midfielder. It's easy to get past our midfield when our attack is high. Teams always seem to be able to get by with one or two passes. Where most teams have midfielders that can delay or move the attack to less dangerous regions, ours just seem to allow it happen.