One of the standard metrics I track is the difference between Celtic’s expected goals scored and conceded.
The rule of thumb is that the higher this value, the less likely that normal football variance factors such as opposition goalkeepers, poor finishing, bad luck, and refereeing, will impact the likely result adversely.
A differential of around +2 xG for versus against seems a good level of immunity from the vicissitudes of the beautiful game. No guarantees of course but decent insurance.
The winter break is as good a time as any to review where we are.
I’ve plotted Brendan Rodgers’ Celtic alongside Ange Postecoglou’s. A suffocating level of dominance in domestic games was something Postecoglou excelled at. Celtic attacked with such relentless intensity the opposition was usually so deep and harassed that attacking from them was threadbare.
Here is the two-and-a-half season trend.
The trendline (orange) is downward but remains in the sweet spot between +1.75 and +2 xG.
There are some “normal” variations:
- An annual September/October dip when European group stages football kicks in;
- An end-of-season lull when the title has been sown up;
- A new manager wobble at the beginning of each reign; and
- A February and March upsurge of form when other squads are stretched.
Specifics are the extreme start to the 2022-23 season under Postecoglou with a very strong start. The downturn once the league was secured that season was also noteworthy perhaps exacerbated by the knowledge Postecoglou was likely leaving.
The Rodgers upturn has been slower than Postecoglou managed upon taking over. Not so much a new manager “bounce”, but more a steady slog up the hill. The lull at the end of the graph coincided with the difficult Heart of Midlothian match and the derby where Celtic were on the wrong side of the xG “score” despite a relatively comfortable win.
My overall take from this is I’d expect the trend to buck upward in the second half of the season as injuries clear and the team continues to improve cohesion.
The gap between average xG for (2.84) and xG against (0.84) is bang on +2. But there is minimal margin for error given the league state and the standard of SPFL opposition.