I like to take the opportunity to look at longer Celtic trends during the preseason before the real stuff kicks off again. At that point, we disappear down the rabbit hole of individual games and performances.

So, it’s refreshing to take a step back and, putting Scottish Cup wins and Scottish Premiership titles aside, study the underlying performance trends.

One of the long-term trends I track is the expected goal difference.

That is: the difference between the xG for and against. I do this on a six-match rolling basis to smooth out the spikes caused by the occasional 7-1 win.

Also, I include Scottish Premier League games only as the variability in European-level matches is too great.

I can illustrate this further by showing you the pattern of xG difference between Scottish and European football.

The orange line shows the average points (three for a win, one for draw) achieved given the difference between xG for and against in European competition.

Around the point of parity (-0.5 xG difference to +0.5 xG difference) Celtic achieves between 50 percent and 60 percent of points available.

That does not change until the xG difference is higher than two, when it rockets up to near 100 percent.

On the other side, once xG difference is against Celtic, the percentage of points won plummets sharply to less than ten percent once the difference is between one and 1.5 xG in favour of the opponent.

Broadly, the kinks are due to the extremes in opposition strength encountered in Europe (from e.g. Finnish sides to Real Madrid). But, also broadly, this looks like a “normal” distribution of points achieved given xG differences.

The green line is SPFL matches.

This shows once Celtic gets to one xG more than the opposition, they are winning 80 percent of the points. Once Celtic gets to two xG more, it becomes 90 percent.

Even with a negative xG difference, Celtic win 70 percent of the available points.

Opponents must generate good shot quality and get lucky to beat Celtic.

The wider point here is that there are very different model outcomes. Stating the bleeding obvious, switching between SPFL and Europe is tricky.

Back to Brendan Rodgers and his performance relative to xG difference.

2023-24 SPFL Season

Celtic’s last campaign was profiled as follows.

The trendline was downwards across the season but never fell below one xG positive difference in Celtic’s favour throughout the campaign.

It was a difficult start with away matches against the stronger sides in the division – Aberdeen, Kilmarnock, Heart of Midlothian, Hibernian and Rangers.

Also, September and October is when league games are interspersed with Champions League group matches. These do extend into December but by then you are into the cadence and your fate is probably known.

This was a period when new faces took time to integrate. It also coincided with a rash of injuries.

Across early November to early December, the season peaked as regards domestic xG differential. The fixtures were reversed, with the more difficult teams at home.

Notably, Aberdeen struggled with a Thursday European night to Sunday SPFL trip to Celtic Park combination and were thumped 6-0.

Hibernian, so stubborn at Easter Road, succumbed 4-1 at the home of the champions.

It was downhill from mid-December to the end of February. Rodgers was disappointed with the January transfer window (in) activity. Injuries were lower in total but focussed on key players – Cameron Carter-Vickers and Reo Hatate particularly.

Read more:

Dispiriting losses to Kilmarnock and Heart of Midlothian were in this period.

But as often happens given the depth of Celtic’s squad relative to the opposition, once Rodgers settled on a core of 17 fit and able players, performances improved to the end of the season and sufficient to clinch the title comfortably.

How did this pattern compare to previous managerial reigns?

2021-22 to 2023-24

Here is the trend against SPFL opponents since the start of the Ange Postecoglou period. I would argue that this period saw a coherence of attacking football that was a benchmark setter in modern times.

The trend line is downward across the three seasons. It dipped most alarmingly at the end of Postecoglou’s managerialship. The combination of tying up the league early and, perhaps, the players sensing the impending departure of the boss, are likely contributors.

Postecoglou and Rodgers have had significant squad rebuild to undertake at the start of their reigns. This inevitably leads to a slow start and periods of sub-optimal performances.

There is also the early season impact of juggling European campaigns interspersed with taking on domestic opponents likely at their peaks before squad injury impacts are felt.

Celtic then usually peak in the late year period where other SPFL clubs struggle with heavy domestic timetables, and then late in the season when Celtic’s larger squad and focus and experience of winning titles come into play.

There are patterns within patterns, therefore.

Overall, the “decline” in xG difference on a six-match rolling basis is within over 1.5 xG and 2 xG in the hoop’s favour.

However, I don’t like declines in performance indicators.

What I would hope is that Rodgers is allowed to get the overall squad quality back to the levels even pre-Postecoglou. I’d argue the failing 2020-21 squad, and certainly the season before, under Neil Lennon, had greater inherent “talent” – a qualitative judgement for sure.

Postecoglou fitted round pegs into a rigid tactical attacking masterplan of round holes. System strength more than compensated for individual talent weaknesses.

Rodgers’s style of football is less positional and more reliant on individual decision-making (we are talking about degrees here).

Overall, I’d like to see two outcomes.

Firstly, a reverse of xG difference decline accepting this trend is within tolerance levels (average remains over 1.5 xG).

Secondly, a way to even out the peaks and troughs – in simple terms, consistency. This will be harder to achieve given the European calendars, and the ever-present injury threat.