My journey with football analytics started as a hobby early in the 2019-2020 season and accelerated as a welcomed distraction during ten weeks of lockdown in spring of 2020.

Like many Celtic supporters, my attention had become focused on the prospects of achieving 10-in-a-row, and at the time, I took to my friend Alan Morrison’s blog and Twitter to scream into the void. I was concerned about potential complacency at Celtic and the growing threat emerging from Rangers. 

My analysis of the 2019-2020 season suggested that the final 80 to 67 points gap (Rangers had a game in hand) was not reflective of the underlying dynamics between the two sides. Unfortunately, many of the concerns laid out were subsequently realised, and the table and relative disparity experienced an extreme flip for the 2020-2021 campaign.

Once again, the 25-point final gap that season was nowhere close to representing the underlying dynamics of the two teams, and Celtic have obviously been able to return the league trophy to Parkhead over the past two seasons.

While the recent loss to Rangers was a dead-rubber game with a partially rotated squad, it did not occur within a vacuum. In my view, the current 10-point gap in the league table is once again wider than the likely underlying club dynamics.

READ MORE: Why Celtic's McGregor has an air of Sergio Busquets about him

Fortunately, my assessment still retains Celtic solidly at the top, but the distance below may be uncomfortably closer than the 10 points suggest. First, let us return back to that 2019-2020 title race with a radar comparing various metrics for both teams up until the winter break:

To that point, the title race was nip and tuck, but on balance, the underlying data hinted that Celtic were performing at a higher level and that Rangers were enjoying a combination of scoring goals well above their xG and conceding fewer as well. After the break brought with it a rash of injuries to the Rangers squad after what had been a period of relative health, and their goal rate for and against relative to xG mean reverted sharply on top of their xG differential falling:

Celtic continued to chug along, but the concern was that some reasonable upgrades to Rangers’ squad depth and the potential for lower injury rates could combine with risks of Celtic taking even a small step backwards in performance and jeopardise the 10-in-a-row season.

Regrettably, those factors played out to various degrees, but as the next radar suggests, a 25-point gap in the final league table was nowhere near reflective of the underlying team dynamics. Rangers outperformed xG significantly both for and against while also enjoying extreme variance in the number of net penalties and red cards, with 10 penalties for, one conceded, six red cards by opponents and zero against. In contrast, Celtic’s penalty advantage was seven to five but had three reds against versus zero for opponents. Combined, that left Rangers +7 in penalties and +9 in red cards.

Celtic Way:

A variety of developments this season have me again nervous that the league table may not be a strong indication of the relative probabilities for Celtic to win next season’s league campaign. Firstly, let us compare the periods prior to and subsequent to this season’s unusually timed World Cup:


Celtic’s relative performance levels were utterly dominant over the first fifteen league games prior to the World Cup, both with results and underlying dynamics, including scoring more than a goal per game. 

However, Celtic’s chance creation declined materially on average following the World Cup, and Rangers’ improved marginally via improved set piece output. In addition, as the following on-ball value radar from the post-World Cup period suggests, goalkeeper play has had a material impact on the season to date:

Allan McGregor and Jon McLaughlin were pretty awful overall this season, and Rangers’ 13th percentile goalkeeper OBV includes Robby McCrorie’s two contributing starts out of the twenty games. In addition, Celtic enjoyed the 98th percentile level of ineptness from opposition keepers over the same period. 

READ MORE: Why Celtic need to find a long-term Joe Hart replacement

Another standout in Celtic’s favour in the OBV radar is the relative advantage in pass OBV. This is where the attacking prowess and creativity of the likes of Jota and Matt O’Riley are reflected relative to a Rangers’ attack that has suffered from a creativity deficit from attacking players and is largely reliant on their full-backs delivering crosses into opponents' boxes. 

So, what has me a bit anxious? Like Celtic over the summer of Ange Postecoglou’s arrival, Rangers appear heading towards a significant squad overhaul. Three of the biggest culprits contributing to the goalkeeper and attacking creativity deficit are likely to be departing: McGregor, Ryan Kent, and Alfredo Morelos. While there are reasonable doubts as to the relative financial landscape for each club in the upcoming transfer window, productive replacements for the three may be within Rangers’ grasp even if finances are limited - such is the nature of how bereft their playing has been of attacking prowess. 

In addition, the decline in chance creation from Celtic during the post-World Cup period did not emerge in a vacuum. Domestic opponents appeared to adopt various iterations of similarly defensive tactical approaches looking to, and enjoying some success, in stunting Celtic’s attack on the margin. 

While these remain a big if, it will be difficult not to upgrade on McGregor for next season, and just one or two reasonable attacking upgrades could be enough to keep the title race closer next year. Also, with Rangers mysteriously enjoying another +10 season for net penalty kicks, a continuation of that sort of good fortune could make for an even more uncomfortable journey. 

Fortunately, Celtic are likely to continue upgrading the squad this summer, so catching the champions is unlikely to be aiming at a static target.