A quick piece on Michael Johnston given the furore following the St Mirren game.
According to FotMob he led the team with 0.46 xG and 0.26 xA giving an Expected Scoring Contribution (xSC) of 0.72. This is slightly below his average of 0.8. And, of course, this yielded 1 assist and 0 goals – leaving his tallies thus unchanged for the season.
Two warnings:
His sample this season is very low – 4.88 matches worth of minutes (900 is my threshold)
I haven’t captured the St Mirren game data myself but the points being made below are unchanged
Attacking Winger Scoring Contribution
Here is a very simple of view of the primary wingers in the seasons I have xG and xA data for.
It plots Actual Scoring Contribution (Goals + Assists per 90m) against Expected Scoring Contribution (xG + xA per 90m).

There is potentially a lot to unpack on this chart but don’t get distracted! Other articles may follow. Let’s focus on Johnston.
He only has two seasons where he has racked up enough minutes to qualify for this chart. In both 18/19 and 19/20 he completed around 12 matches worth of minutes.
As a young player his numbers would place him in a banding that includes the worst of Roberts (17/18 injury hit), the worst of Forrest (17/18) and Morgan (hmmm). I would say that is OK for a young player.
This season (top left corner), his xSC is up there with the best seasons Celtic wingers have had in recent times (Forrest (19/20) and Elyounoussi (20/21) - both had 0.81 xSC). Johnston’s current xSC of 0.8 (slightly lower post St Mirren) is therefore impressive.
What is perhaps surprising (not yet astonishing due to sample size) is he has 0 goals and only 1 assist this season off such a high xSC. That assist came when he was playing off the right wing against Real Betis as opposed to his normal left wing position.
Conclusion
You can explain this with “bad luck” to some extent.
But such a disparity between Expected and Actual, even at such small sample sizes, must surely come into the “poor execution” territory. Remember, xG models are built on what an “average” player would achieve.
Maybe this is confirming Johnston’s final shot and final pass are just very poor quality relative to the positions he finds himself in. The eye test would support that.
It is also perfectly feasible he will go on a run of assists and goals similar to 19/20 when he racked up 0.94 SC per 90m compared to an xSC of only 0.52.
We can certainly hope so.