Are Celtic playing slower this year under Brendan Rodgers?

There is no doubt that some harbour resentment at his return to Celtic, given the midnight flit to Leicester City in February 2019. That is, there will be no pleasing some.  Also, he is following a manager in Ange Postecoglou, who seemed to have discovered the cheat code for cracking Scottish domestic football, such was the elan with which the opposition was routinely put to the sword in a whirl of attacking football. The Australian is renowned for thrilling attacking football, as the English Premier League is discovering, as his Tottenham Hotspur career continues to build.

Also, Rodgers was undoubtedly focused on control in his first stint at Celtic Park. He used to talk about 'resting with the ball'. The connotation is that possession was not always with the intent to harm the opposition, but to allow the team to gather energy for the next onslaught whilst simultaneously tiring the opposition without the ball.

READ MORE: How Celtic beat Hibs, thanks to Idah, Ralston and Bernabei

When we interviewed Dominic Wells on the Huddle Breakdown - a very talented data analyst who follows Leicester City - he was clear that in the four years of Rodgers's time there, he had developed tactical flexibility in delivering three consecutive top-eight finishes, an FA Cup and a European Conference League semi-final. I was encouraged by this and have therefore remained sceptical about the tropes laid on him.

I have also written about how the overall speed of play in the Scottish Premier League (SPFL) has increased quite markedly, or at least, perhaps the Opta data underpinning this has changed in terms of capture and calculation method. Celtic do get the ball forwards much less quickly than the other sides in the league, but quicker than they had under Postecoglou.

As Celtic hit a mini downturn in performances and results since the January break, this question comes up again. So, how to address it using data?

Pass intensity

The Opta direct speed data was certainly surprising and raised eyebrows with the other analysts I shared it with. Make of that what you will.

As I have mentioned speed of play data whether it be the pace individual players move or pass is quite sensitive and not usually available through the public data providers. A simple proxy is the number of passes attempted per 90 minutes. The assumption is that the higher the number the quicker the players are looking to move the ball.

READ MORE: Detailed Celtic player ratings as Idah steals the show

What I have done is possession adjustment given that each season the team has a slightly different average possession metric. The possession-adjusted method is that used by Wyscout. Based on this method, here are the average passes attempted (from open play) per season.

Celtic sought significantly more control of the ball under Postecoglou. However, when adjusted for possession, Celtic attempt more passes from open play this season than on any I have records for – eight more per match than last season.

The 'Horseshoe of Futility'

I few years ago, concerning the way Ronny Delia’s team attempted to overcome low blocks, I coined the phrase the 'Horseshoe of Futility'. That is, to pass from side to side back and wide around the opposition defence looking for opportunities. It is a phrase some used to describe Rodgers’s unfolding Celtic campaigns in 2017/18 and 2018/19.

I propose a simple measure for this to enable a horseshoe indicator to be devised. What proportion of the attempted passes comes from the centre-backs? This is not perfect by any means. Sometimes Celtic have played with three centre-backs notably under Neil Lennon. Secondly, it does not indicate 'safe' passing per se, so there is an underlying assumption that the more the centre-backs have the ball the less, relatively, the more creative full-backs and midfielders have it. That may not always hold true. Here are the results:

This is okay, as far as my eye test goes. In the first two seasons under Deila, the football was simply very slow all over the pitch. Since then, the horseshoe indicator has been lowest under Lennon in 2019/20 and highest under Rodgers in 2017/18 and this season. But again, this is not indicating speeds as much as the perception that if the centre-backs are heavily involved in the passing, then the ball is perhaps not being progressed forward quickly.

My next thought is ‘What proportion of the passes are aggressive forward passes?’ I can measure this by showing the percentage of all passes that take opponents out of the game – pack passes.

This shows that it was perhaps fair to label Rodgers Celtic in seasons two and three as somewhat ponderous. The horseshoe indicator was 27 per cent yet only around 10 per cent of all passes were pack passes – forward passes taking the opponent out of the game. This season, despite the highest horseshoe indicator at 30 per cent, 14 per cent of all passes are pack passes. This is slightly below Postecoglou’s very aggressive passing style.

Summary

The reality is that Celtic appears to be moving the ball quicker than ever, it is simply that the centre-backs are more heavily involved perhaps due to the preponderance of man-to-man marking across the SPFL. If we add this to the Opta data evidence, I think it is difficult to support ‘slow and ponderous’ despite the horseshoe trope having some truth.

READ MORE: Celtic's new penalty expert Idah may mean trouble for Kyogo

My thoughts are that ‘slow and ponderous’ is getting mixed up with ‘inefficient’. And that is an analysis for another day!