By James Dailey
In the lead up to the Cup Final in December, I started a comprehensive comparative analysis between the squads of Celtic and The Rangers, while trying to attribute where they were performing better than our team. At that time, their cumulative xG was better than Celtic’s and their xG surrendered was only slightly higher. I quickly ascertained that their midfield had been significantly outperforming ours in attack, but that otherwise every other position group was superior in Celtic’s favor.
Us vs Them
I’ve been clear in other recent pieces documenting Scott Brown’s deficiencies in attack, but it is in comparison to The Rangers’ midfield which this becomes vivid. Stephen Davis’ offensive contributions to The Rangers’ attack is more than triple that of Brown’s to Celtic. Additionally, McGregor’s contributions to attack have been significantly neutered while playing next to Brown as part of the 4-2-3-1 formation which has dominated Celtic’s play until recently. The drop off has been so severe, combined with significantly improved play by Ryan Jack, that Jack’s underlying stats are comparable to McGregor’s this season. Here are the heatmaps for McGregor for this season versus last for SPFL games only:

The heatmaps clearly show how McGregor’s role has shifted, from more of an attacking box to box midfielder with relative freedom of movement, to a more constrained role.
The following xG and xA stats come from Ortec via Modern Fitba, while the remainder are courtesy of Wyscout. All focus on SPFL league games only and are per 90 minutes. The 19/20 season stats are up until the winter break, which I will explain shortly:

Fixed the Midfield?
One of my biggest criticisms of Lennon this season has been how I believe he has mismanaged our midfield, which is ironic given his legacy as a player. This has become particularly germane after having been dominated in the midfield during the two December derbies. Christie and Ntcham’s performance stats are also modestly better when they play as attacking 8’s rather than as 10’s, but it is McGregor whose contribution to attack has really been impacted. It seems like a very bad idea to me to shoehorn our best midfielder into a role which is suboptimal for his skills – particularly when the “payoff” seems to be negligible.
I may be persuaded if his defensive stats spiked and his more defensive role was contributing a lot in preventing goals, or if the shift unleashed other vital players to be able to contribute at higher levels. However, I just don’t see any evidence that this has been the case, as most of McGregor’s defensive stats are similar or even down from last season!
This next group of stats is admittedly from a very small sample size and targeting the four recent games in which Celtic has deployed a formation where McGregor played a significant portion of the game as a left central (MLC) rather than defensive midfielder, whether as part of a 3-5-2 or 4-3-1-2 (or 4-4-2 diamond as it is also known). Those four games were December vs Hibernian, and then this month versus Partick Thistle, Kilmarnock and St. Johnstone.

I focused on league-only stats prior to the winter break because, except for the single game versus Hibernian in December, Lennon deployed an almost universal 4-2-3-1 with McGregor and Brown in every other game – the sample is nearly “pure.” I included the Cup game versus Thistle due to how few games where we’ve deployed formations where McGregor was a MLC.
Conclusion
I’ve become resigned to the idea it is highly unlikely that Brown will be dropped from the first team this season. However, I believe playing a formation which unleashes our best midfielder will go a long way to securing 9 in a row. As I speculated recently in The Real Return of Lennyball, I am hoping Lennon returns to his “roots,” which should go a long way in fixing our midfield and unshackling McGregor. This makes three out of four games since the Winter Break where this has been the case.