With Alistair Johnston and Alexandro Bernabei recently passing the 900-minute threshold for meaningful data, I wanted to try out a new view of overall attacking threat.
Since the World Cup break Celtic have scored 57 goals in their 18 matches as the cohesion of the team and strength of Ange Postecoglou’s system has proven consistently too strong for SPFL level opposition.
All players are expected to contribute to this outstanding goal threat to some extent.
To assess who is contribution what I like to combine metrics that measure quantity and quality.
In this case, quantity is represented by my aggregated Celtic Attacking Threat (CAT) Score. This counts key attacking events such as shots, touches in the box, successful assisting passes to give a volume of attacking activity.
Quality is represented by expected scoring contribution which is simply xG + xA – expected goals plus expected assists.
Defenders
I have lumped the defensive players together.

The centre backs are, as expected, at the bottom left with little of note other than Carl Starfelt’s propensity this season to be on the end of corners.
Of the full backs, note that Josip Juranovic was the most efficient with the highest xSC off the lowest CAT score (slight penalty xG boost remember).
There is not much to separate the other full backs which may reflect system over individual given they are all quite different styles of player. Note that Bernabei has a high volume of contributions and the highest xSC. He has tended to play against weaker opponents to be fair.
Midfield/Forwards
Probably more interesting are the midfielders and forwards.

This highlights the key role Giorgos Giakoumakis had as a “super sub”. In some respects, Liel Abada has taken on that mantle with a ridiculously high 1.15 xSC per 90 minutes off 11.27 CAT which is the only player over 10 – next highest is Sead Haksabanovic on 9.89.
Callum McGregor’s role is probably unique so no harm there.
The other central midfielders are clustered and Matt O’Riley is noteworthy given he played 12 matches in McGregor’s role which should serve to supress his attacking stats.
You can see that Reo Hatate, for all the aesthetics and moments of magic, is still to achieve the level of output of his peers.
Of the wingers, Diazen Maeda is similar although we know he brings other unique defensive attributes to the team.
As mentioned, Haksabanovic racks up impressive attacking numbers but the presence of he, Giakoumakis, David Turnbull and Abada should alert you to another important dynamic.
The Sub Dividend
As I have mentioned many a time, the five-sub rule is a massive tactical advantage for Celtic given the depth and quality of their overall squad.
Essentially, Postecoglou changes the front five every match from 55-75 minutes. Teams with 20-30% possession have a new, equally skilled but fresh set of problems to deal with on tired legs and minds.
This is reflected in the data.
Of the attacking players, eight of the 11 have a higher xSC coming on as a sub than as a starter. Giakoumakis’ xSC as a sub was an incredible 1.46.
Here are the % of minutes played per player as a starter:

Those at the bottom of the chart, with 30-40% of their minutes as subs are also near the top of the average attacking threat chart above. No coincidence.
Equally, there is almost a tax on starting in terms of stat suppression. Hence Jota and Kyogo Furuhashi not storming the charts! Giakoumakis’ productivity data was “higher” than Kyogo’s overall, but Kyogo’s are almost exclusively as a starter against fit, alert opponents and before being taken off when the game opens up.
The good news is that the system Celtic play means there are plenty of attacking threats across the squad.
When quoting attacking numbers though, be mindful of the advantage sub minutes have in terms of the ability to rack up good stats against tired defences.