This analysis lark is much about pilfering and refining the fine work of others.
It’s all fair game so long as you attribute!
So, props and shout outs to @CannonStats on Twitter. You might want to follow them if your schadenfreude extents to giving support to Ange Postecoglou’s new big rivals.
In this tweet, the midfielders in the top five leagues (why always them?!) are plotted as regards their ball progression (progressive passes + ball carries) mapped against defensive actions. All per 90m.
The point of their diagram is to show that putative new signings Moises Caicedo and Declan Rice are uplifts on their current midfield as regards the combination of the above attributes. (Marco Verratti by the way – ooft).
Central Midfielder Framework
As central midfield continues to be one of my main worry areas going forward, I thought I’d pinch the framework and look back at central midfielders in the years I have comparative data for – 2019/20 to today.
The definitions will be a little different.
Instead of “progressive passes” I will use pack passes – forward passes that take opponent(s) out the game.
Ball carries becomes progressive runs which is pack runs (a forward run that takes opponent(s) out the game) + progressive carries (carrying the ball at least 10 yards in the opposition half).
All players played at least 900 minutes in that season.
The current midfielders are highlighted.
Analysis
I would like to highlight the following:
Nir Bitton suffered by not being Scott Brown and by the mistake of trying to turn him into a central defender. In 19/20 and 20/21 he mainly played as a centre back. In 21/22 he mainly played in midfield. Yet his data profile by these measures is almost identical. That possibly can be seen as vindication of playing him at centre back other than he wasn’t a very good defender! In 21/22 he was also playing under Postecoglou’s (remember that guy?) more aggressive vertical passing regime.
Moving back, Olivier Ntcham’s stellar 19/20 season was significantly more effective than any other current midfielder crop managed this season. Indeed, Ntcham’s “can’t be arsed” 20/21 season is still more effective than this years’ efforts. Whilst Reo Hatate has undoubtedly improved, he does not match Ntcham’s “poor” season.
Which reinforces the concern about the quality of the current squad members.
Callum McGregor inhabits his own little zone in the bottom right corner. One characterised by much ball progression but low defensive activity.
That’s ok, we know he is the tempo setter. What worries me is that McGregor’s performances in this regard have dipped below the previous two seasons. My Spidey senses are twitched when a player approaches his 30s and performance regression is evidenced. I’ll signpost this now so folk won’t get upset if this thorny subject is raised again over the coming seasons.
Up to the top left and the defensive monsters with less creative ball progression.
What is surprising perhaps is that Matt O’Riley profiles more like Scott Brown than he does McGregor! I have long banged on about O’Riley’s defensive capabilities and believe he is the only one of the current crop capable of performing at Champions League level in this regard. For example, in pressures and pressure regains he is second only to Daizen Maeda in the squad.
His overall performance levels have dipped versus last season which is also a concern, and I would like to see him offer more in the ball progression stakes. His ball progression should be higher than Brown’s was, for example.
As with Turnbull/Mooy/Hatate vs peak Ntcham, so O’Riley pales slightly next to Ryan Christie. It is another indicator we need to upskill this key area.
Finally, one of the on-loan players is Ismaila Soro. He is the busiest defensive player in the last four seasons and his ball progression is no worse that the current midfielders. Could he make an unlikely redemptive return?
Summary
Central midfield remains a major area of concern for me as regards comparison with a Champions League benchmark.
Based on core skills of defensive actions and ball progression, we have not improved the midfield over the days of Brown, Christie Ntcham and even Bitton.
Additionally, there are concerns that McGregor’s performances may be regressing (that could be reversed of course).
Hatate, O’Riley and Turnbull are all young and bright players and have growth in them, so there is hope still.
But I remain fixed it is an area Celtic need a significant quality uplift.
The drop off in numbers might be down to the patterns Ange had the team playing. Everything seemed to be circulated around the back line out to the wings most of the time
Fascinating.
Excellent work. My thoughts include that Matt O’Reilly was very good as the number 6, particularly against Shakhtar at Celtic Park and Callum is targeted by man marking and seems to have to fight to find space before being able to lift his head to make forward passes. I love both players but I thought the midfield missed Mooy near the end of the season.
West Ham are the English Ross County but having strong athletic players does seem to help then in midfield.
Interesting stuff Alan, again.
One clarifying question, I would ask is whether our midfield sun players coming on for the last 15 to 30 minutes score higher on these metrics than starter midfielders.
What percentage of Bitton, N’tcham and Christie’s figures were earned as subs against citing defences?
A fair question and I have not checked but unlike the wingers and forwards less sun stat padding . If anything that would apply to Turnbull
Very fair point
While useful I’d argue that these two metrics don’t show the full picture.
For a start I think you get “stat players” who are very aware of their numbers and maybe artificially enhance them by racking up unnecessary passes etc or don’t take any risks.
Players sitting deeper I would be expecting to receive far higher packing scores than those in more advanced roles. Both because there’s more players in front of them and also because there tends to be less pressure or congestion directly around them. Those further up the field have less bodies to hit and are perhaps more likely to try passes thay lead to turnovers.
The same with runs.
Defensive actions is interesting but does not account for things like blocking passing lanes or occupying space effectively.
I would be of the opinion that as good as McGregor is at setting a pace he’s not hugely involved in the creation or prevention of goals. That may be by design.
The numbers are probably harshly impacted by how few times Kyogo sees the ball as well. Guys like N’tcham were hitting Eduoard or Dembele who constantly sought out the ball. Far easier to ramp up your score with those type of strikers.
Although we don’t play with a real number 10 these days I’d like to see Haksa tried in an advanced central role or even how his numbers would stack up. Don’t see him being a success in a wide role with us.
Does the above table include their packing score overall or just the positive?
Just the number of pack passes and you are correct to highlight as eg Starfelts packing score would be much lower that O’Rileys on an average per pass basis