How Postecoglou's Celtic stacks up against Rodgers and Lennon sides

This column will examine and compare the current Celtic vintage to recent teams and look to highlight some of the potential drivers using the 'Expected Trophies' and 'Jenga Tower' concepts shared earlier in the season.

First, let's take a look at a broad spectrum of performance metrics comparing this season with each of the last three:

It would only be natural for the 2020-2021 season’s comparison to have the most significant impact on our perception of this season’s performance levels and the radar comparing the two shows how dramatic the improvement has been.

However, this season also compares extremely well with the prior two seasons as well, including the extremely talented 2019-2020 team which dropped just 10 of 90 available points in the pandemic-shortened season.

Just as a reminder, metrics like xG are oriented to try and measure the relative quality of chance creation, while On-Ball Value (OBV) is designed more to decompose actual output. Here is a comparison of the 2019-2020 team with this season to date:

This radar offers some insight as to how the Broad Metric radar suggests a higher level of underlying performance overall via xG difference, yet the 2019-2020 side earned more points per game.

The two OBV-related metrics which stand out are the disparity in Shot OBV for Celtic, and then the Goalkeeper OBV Conceded. The 2019-2020 Celtic side enjoyed a season of good relative finishing, while also the added benefit of opposing keepers being extremely poor. Here are the shot maps for xG and post-shot xG (PSxG) for that season:

We can see from the graphics that the Shot OBV reflects the fact PSxG of 65.90 was higher than xG of 62.08, and that Goalkeeper OBV Conceded is reflected in the disparity between 82 goals scored compared to the PSxG of 65.90.

In contrast, here are the shot maps so far for this season:

Again, we can see how the relative OBV values are reflected in the xG and PSxG shot maps. The disparity in goals between 61 and 62 reflect an opposition own goal, which does not register PSxG.

The goal of this exercise is to examine how these various metrics may be analysed in a complementary fashion, but also hopefully display the potential value advanced statistics can offer in furthering the understanding and accurate attribution for performance levels.

For example, a Celtic manager is unable to control the level of performance that opposition keepers pose, which appears to have been a huge factor in results between the 2019-2020 team versus the current one. Now, let's return to the Expected Trophies and Jenga analogies.

While official disclosures will not arrive until the club’s financial accounts are released this autumn, the club’s wage bill will likely have dropped compared to the 2019-2020 season. As laid out in this column for The Celtic Way, the wage bill is by far the most important driver of club performance levels. This is important context given that the current team has been performing so well despite likely doing so with fewer resources.

After the wage bill, the next two components of the Expected Trophies analytical framework are recruitment and quality of coaching/management. These are far more difficult to attribute with precision, particularly given Celtic’s relatively opaque operating model in which decision-making for recruitment appears to have been fairly erratic. This is where the Jenga analogy enters the picture, as we consider the relative coherence in playing style, tactics and how player recruitment and utilisation either fit together well or not.

Given the recent news of Scott Brown’s departure from Aberdeen and related discussion of his legendary Celtic career, his utilisation over the prior few seasons is a good case study to further examine the Jenga analogy. This radar examines and compares Brown’s defensive actions during the 2018-2019 vs 2019-2020 seasons, with a particular focus on actions in opposition halves:

While there are obviously similarities, we can see from various metrics that Brown’s role, despite being two years older and in his mid-30s, resulted in him being more active in an advanced position into opposition halves. How did this manifest in measurable performance value?

Brown’s Defensive Action OBV was barely positive, and ranked in the 10th percentile for midfielders in the league, which was obviously less than ideal for a team of Celtic’s resources, and supposedly filling a role that focused upon defensive actions. However, this was not a new development, as shown here:

Despite a team defensive record which was excellent, Brown’s value for the 2019-2020 season via Defensive Action OBV was actually -0.02. Here is Brown’s OBV radar comparing his 2018-2019 season with this season at Aberdeen:

To Brown’s credit, despite playing at age 36, his Defensive Action OBV returned to being comparable to his 2018-2019 season. This is not to suggest his fit was necessarily a good or a bad one for Aberdeen this season though.

Outside of decision-making relative to managing injury risks, which has been examined here, the analysis of Celtic’s performance levels under Ange Postecoglou has been glowing and continues to suggest that his value is likely underestimated. Note the relative defensive value and efficiency from Callum McGregor this season compared to Brown’s 2019-2020 season:

The strength and resilience of the Jenga Tower will hopefully increase further if recruitment and resources trend in a positive direction.