There was an excellent article in The Herald using data (crumbs!) to compare McGregor and Brown in the defensive midfielder position. I strongly recommend you read it. Sadly, it was uncredited but great to see stats-based content in the mainstream media.
The data used was from Wyscout which is a subscription-based service. As you know I capture my own data which may be better or worse (but rarely the same) as data collected elsewhere. But it is free and who doesn’t love free?!
Therefore. I would like to offer you my take on the McGregor versus Brown debate using the data I have which covers all matches in all competitions (I am sure Wyscout does the same).
McGregor’s stint as the holding midfielder really began away at Dundee with injuries simultaneously to Brown, Mulumbu, Kouassi and Ntcham. Up to the League Cup Final that is 7 completed matches which is not a large sample.
The role of the defensive midfielder or “6” is primarily about distribution, positioning and defending. And there are numerous indicators for each.
All values are Per 90 Minutes averages (unless they are percentages).
Distribution
I have McGregor completing more open play passes (67.14) than Brown (66.59), but only very slightly. McGregor does give the ball away with stray passes (8.14) at exactly double the rate of Brown (4.07) however. Brown’s overall pass success rate is 94% (including set piece passes now) compared to 90% for McGregor. Brown’s number is more akin to a central defender which is the type of passing accuracy you want to see from your defensive shield. That being said, there is a debate to be had on the merits of a more expansive (i.e. riskier) passing style against deep sitting defences in Scotland.
The Herald article allows you to interact to see different passing nuances (long passes, backwards passes, forward passes, passes into the final third and so forth). It’s really neat – try it. With an average of 844 passes attempted in any one Celtic game I don’t have the army of data inputters needed to tag each individual pass! But kudos for providing that data.
What I do done do is measure forward passes that take out opponents (packing). By this measure, there is a stark contrast between the pair. In the chart below Pack Passes90 is the number of forwards passes that take out at least one opponent per 90m. Pack Pass Impect 90 is the total score assigned reflecting the volume of position of the opponents you take out (more points for taking out an opposition defender).
Pack passes are inherently riskier, but take the team nearer the opponent’s goal with less opponents in the way, simplistically.
It is no surprise that McGregor is the more progressive passer by this metric but perhaps the differential in volume and effectiveness of pack passes is.
I track Secondary Assists which record the pass before the chance-setting pass. Also known as the Establishing Pass which is perhaps more descriptive. As most chances are created by attacking players, this concept acknowledges that many chances are as a result of passing moves and that the enablement of the pass perhaps started further back in the field. It allows a more nuanced assessment of combinations and contribution from all over the team.
In the 7 games McGregor played at “6”, he averaged 1.71 Secondary Assists per 90m compared to 1.13 by Brown. Not only did he create slightly more chances, the quality of the chances created was much higher. The shot resulting from the Secondary Assists from McGregor was 0.254 xG per 90m compared to 0.119 xG per 90m for Brown. And before anyone says “well McGregor takes corners so that explains it”, only 1 Secondary Assist was as a result of a Set Piece action in the sample. 2 of McGregor’s Secondary Assists were in Fast Break plays which are highly effective in providing goal scoring chances – Celtic have scored 14 goals from 70 Fast Breaks this season.
The point being, McGregor’s more forward-thinking style better facilitates getting the ball forward quickly to facilitate such scenarios.
Overall, McGregor completes 12 Pack actions (either passes, receives or dribbles forward that take out at least one opponent) per 90m compared to 8.36 by Brown. His average Total Impect score when playing as a defensive midfielder is 62 whilst Brown’s average per 90m is 32.
Defending
Since McGregor played the screening role Celtic only conceded 1 goal, and that was to RB Leipzig. On the face of it, the midfield has been screening the defence well. In those 7 matches, the opposition averaged 6.28 shots per 90m compared to 8.83 in the 23 matches previously. The opposition averaged 8 possession in the box with McGregor at “6” and 9.39 without. Of course, it is a team game and other changes such as the emergence of Benkovic, the return of Boyata and the whirlwind that is Christie are all factors. But the team didn’t struggle in this regard.
Back to the individuals, and the most telling measure of overall defensive performance is the Defensive Action Success Rate (DASR). See the Glossary for a full definition of that and indeed all the metrics used.
Surprisingly, McGregor has a 3% better success rate with his defensive actions. It has to be said he executes far less actions defensively than Brown. Brown is involved in 11.25 challenges per 90m compared to 8.57 per 90m for McGregor. This agrees with the Herald article. However, that article only considers the volume of defensive actions, not their outcome, which is why DASR is key.
And to layer on more surprise, McGregor wins back possession more than Brown. Excluding passes, McGregor has a Possession Win/Loss success ratio of 83% compared to 77% for Brown. McGregor is only dispossessed 0.86 times per 90m compared to 1.39 for Brown.
McGregor only misses tackles 0.14 times per 90m compared to Brown’s 0.43. And they both have similar Defensive Error rates with 0.11 for the captain and 0.14 for McGregor per 90m.
Since Brown has been missing Celtic continue be involved in around 80 challenges per match. So, the amount of challenging to be done hasn’t changed but obviously McGregor is less involved in that aspect. Perhaps the prodigious work rate of Christie alongside McGregor is providing balance to the team. Christie is involved in an incredible 16.94 challenges per game. That is 4 more per 90m than the next highest, Ajer (12.91).
Positioning
A defensive midfielder must have positional discipline as he is the shield in front of the defence and the central position is the most productive to attack from.
Both players are bypassed by passes and dribbles at around the same rate (i.e. packed). Brown is bypassed 8.52 times per 90m and McGregor 8.57.
Contrary to the Herald article, I have McGregor completing more Interceptions (3.43 per 90m) compared to Brown on 2.95. This may be down to definition of Interception and the games included (I have included all Celtic matches not just SPFL games).
McGregor even manages to block more shots – stopping 0.28 per 90m compared to 0.16 by Brown.
Where Brown does have the edge is that although McGregor wins the ball back in his own defensive third more often (1.43 time to 1.39), the captain loses possession in his own defensive third less often (0.21 per 90m compared to 0.57 for McGregor).
I have an aggregated metric called Possession Effectiveness Index (PEI) that takes all a player’s actions and calculates to what extent the team retains possession considering the outcomes of all those actions. Brown’s PEI is 93% whilst McGregor’s is 88%. Overall then Brown is more secure retaining possession across all actions.
Conclusion
My confirmation bias in starting this article was that I would compliment the Herald article with a slightly different take on some other metrics but reach much the same conclusion – i.e. Brown’s contribution is underestimated and the numbers shows what he brings to the team as the premier suiting midfielder. That is probably still the case but I have under estimated the extent to which McGregor has adapted his game to the demands of that position.
What an intelligent footballer he is.
What gloried exponents of the number “6” position such as Makelele and Kante have shown is that size and physique are not essential to be successful in that role. In fact, speed of thought and movement allow balls to be intercepted, nicked off opponents, rather than relying on crunching tackles and a lottery as to where the ball ends up. And speed of thought allows quick transference of the ball forwards.
It may be peak Brown would easily out-number McGregor in the analysis. But as shown in The Declining Influence of Brown, this is not peak Brown.
This is based on a small sample and includes matches against some hapless opponents as well as RB Leipzig. But the signs are McGregor could be the Brown replacement in our midst. Especially with Christie working prodigiously hard alongside him taking on much of the heavy defensive lifting.
Always enjoy reading your analysis and appreciate your efforts. HH
Very kind – thank you
It’s not in the statistics but the reason Calum is better is the overall psychology of the team. The mostly no purpose of Scott passing is retention which inculcates the mindset of the team where as McGregor creates a forward positive attitude all over the park, ain’t rocket science. There is a role Scott can play but it has to be selective?????
I think the positives of having a progressive player at 6 are clear so long as it does leave the team more defensively exposed. So far so good.
Great article, Love your work.
HH
KLV
Thank you
As always, well written article with substantive data to explain the narrative.
Am I allowed not to like the outcome – i like the idea of more peak Broony…
I haven’t gone back and checked whether he has improved since my rather doom laden early season article (*cursory glance* – he has) but the overall fluidity of the team suggests a step forward. Pro football is tough.
An excellent article. I really enjoyed it. Is there any metric for number or touches of the ball prior to passing or even the amount of time spent with the ball prior to a pass?
Whilst it might be my imagination, I feel McGregor takes less touches and plays the pass much quicker than Brown. Hence keeping up the momentum of our play and stopping defences from getting into their positions.
Not really but the fact both McGregor and Brown complete the same number of passes per 90m probably dispels that perception. However McGregor clearly completes more forward passes.
Peaky Broony vs Connor mcGregor’s wee brother.
Another excellent article and one I was hoping to read.
Have you done on on – Is Scott Sinclair recovering form or on a slow downward slide?
Excellent stuff yet again bud and quite revealing regarding the speed at which McGregor has undertaken this new role and has quickly established himself as a key player in a role which has been the Domain of Scott Brown for several years now?
McGregor is an intelligent footballer who has the ability to play in any role across the midfield.
His the are now getting very rare in the modern game as Managers tend now to have players focus on one role now.
He’s multi talented,never hides,always makes himself available regardless of the level of opposition or stage he’s on and seems to have that in built natural ability to find or make space even in a packed midfield.
When you consider how highly Scott Brown is regarded as a Midfielder it is quite a compliment to McGregor that the Manager has opted to stick with him in crunch games like Rosenborg and Killie at the weekend.
I totally agree with you regarding height and physicality in this position after all Marco Verratti is only 5’5” tall and is probably one of the very best at this role at present in European Football.
Positional awareness and timing are just as if not more important than how hard or how often you can tackle as a defensive mid.
The stats seem to back up the current thinking that Browns successor is already at the Club and Broonies days as a first pick are now numbered.
http://www.football-analysis.com/the-regista-and-the-evolution-of-the-playmaker/
For those remotely interested in the role of the deep lying playmaker.
Thanks. Another article tomorrow perhaps labours the point but the transformation in the team is startling.