Brendan Rodgers' Celtic must be more efficient in attack

I’ve written about the defence often so far this season. But especially in domestic football - since losing Jota and Aaron Mooy - as well as a long-term injury to stat monster Liel Abada, it is arguably the attacking potency that is more of a concern. I say “concern”, in the spirit of looking for continual improvement and for Celtic to always be the best version of themselves.

Before we start, there are discrepancies between different data vendors and how they characterise Celtic’s shooting data. StatsBomb seems to hate Celtic and paint a picture of abject misery – more on that later. Opta via FotMob are more sanguine and has calculated that Celtic easily have the best xG per shot average in the league. Then my own data, which is a bit in the middle between the two. This often seems to happen.

READ MORE: Celtic's Brendan Rodgers has more to do to match predecessor

SPFL Shot Data

The following charts show you the number of shots per game for Celtic and their opponents by volume of shot and xG of that shot.

Celtic take far greater numbers of low xG shots than their opponents, which isn’t unexpected. However, at around the 0.35 xG per shot range, the two lines coalesce, and Celtic are taking no greater volume of high xG shots than their opponents. The lines should emerge eventually simply because the higher the xG shot the lower the incidences of such shot.

For comparison, here is last season’s distribution:

Here we see that Celtic outperformed the opposition at every xG value of shot down to around 0.7, which is near enough the value of a penalty.

In other words, Celtic were consistently able to generate a higher volume of higher-quality shots than their opponents in the SPFL compared to this season.

All Shots, All Competitions

I’ll now shift to my own shot data which covers all competitions.

The following maps out three shooting metrics at team level.

1. Shot accuracy – the percentage of shots on target.
2. On-target conversion – the percentage of shots on target resulting in goals.
3. All-shot conversion – the percentage of all shots resulting in goals.

This shows that Celtic are putting a higher proportion of shots on target than in any of the last two seasons under Ange Postecoglou. Despite this, they are converting significantly fewer shots on target (down from 43 per cent to 28 per cent) and fewer shots overall (down from 16 per cent to 12 per cent). If we look over the last 10 seasons, the average accuracy is 38 per cent, on-target conversion 34 per cent and all shot conversion 13 per cent.

READ MORE: Liam Scales' amazing journey to Celtic: Mentors on his meteoric rise

However, Postecoglou’s finely tuned attacking patterns resulted in an efficiency of goal threat we want to see continue. Given Celtic are being more accurate, and the conversion rate is low, could that hint at simple bad luck, a streak of poor finishing, or opposition goalkeeping excellence?

Possibly elements of all of those, but let’s consider the average xG per shot and where Celtic are shooting from.

Shot Location and xG Averages

This maps the percentage of shots that are being taken from inside the box and the average xG value per shot across all competitions.

Under Postecoglou, Celtic achieved the highest average xG per shot values in the ten years I have been collecting the data. It is perhaps no surprise this has dropped off given the personnel who have left, the injuries, the depth in the striking positions, a tricky fixture list and adapting to a new manager.

However, are the team helping themselves if the shots taken from outside the box are on the increase, or at least shots taken INSIDE the box are decreasing to be more accurate (down from 11.43 to 9.45)? It is clearly preferable to take a higher proportion and volume of shots in the box, given their higher xG value.

Player Improvements?

We are still very early into the season, but which players can perhaps recalibrate their shot decisions to take less from low xG outside-the-box positions?

Looking at players who have completed more than 180 minutes so far, here are the % of shots from outside the box and average xG per shot:

Perhaps the starting front three of Abada, Daizen Maeda and Kyogo Furuhashi are the most efficient in terms of a lower ratio of long shots and higher average xG per shot.

O’Riley is clearly having a productive season and his breaking into the box more regularly is bearing fruit. In terms of an improved profile, midfielders David Turnbull, Callum McGregor and Reo Hatate should all aspire to have shot profiles more like O’Riley’s.

And recent new start Luis Palma clearly likes shooting from extravagant angles and distances. Two early Celtic goals will no doubt encourage him. But I wonder if his luck/skill (you decide) will hold and he will beat the xG models, or whether he will revert to the mean over time.

Summary

Whilst I believe the StatsBomb data provides the bleakest possible picture of Celtic attacking/shot potency, there is no doubt, averaging across a number of data sources, that shot selection and average xG per shot is down on last season. We are still in the small sample territory, however.

Under Postecoglou, the team achieved hitherto unreached heights compared to ten years of data, so I wouldn’t be getting too despondent yet.

READ MORE: Why Rodgers Celtic won't fear new Rangers boss Clement

It would be good to get back to those levels of productivity and an easing injury list and fixture difficulty list may see that transpire.