How Are We ‘Keeping?

C:\Users\Alan\Documents\Football\Celtic Stats Analysis\Images 17-18\Astana H Gordon out again.JPG

Gordon reprises the “Astana out of box heading incident” – this one had a better outcome that in 2016!

It’s the international break, and the transfer window is now shut (hurrah!). Celtic content is thin on the ground. So, as we are 11 games into the 2017/18 season, here is the first in a series of quick (well, not 2500 words. Ed – get on with it) snapshots of where we are performance wise. I usually go with 10 games or 900 minutes as an acceptable sample size for data. Clearly not all players have played that amount, so please take care as small sample sizes may render some statistics as outliers at this stage in the season.


Gordon continues as first choice’ keeper and is the ONLY member of the squad to play every minute of this nascent season.

There are a number of data points I collect about the ‘keepers which I don’t often find a way to share within the match analysis articles. So here goes! Note unless stated, all values are number of occurrences per 90 minutes.

Goalkeeping Metrics

By far the most common means of save is to catch the ball, which has happened 13 times. Gordon has managed 1.91 saves per 90 minutes. He has made a total of 21 saves.

Of the 21 shots he has saved, the average xG (Expected Goals) of those shots is 0.44 per 90 minutes. Therefore, over the 11 games, the total xG for all the opposition would have been 4.82 goals.

Celtic have conceded 6 goals but I count 5 for this analysis as one was an own goal (Ajer A vs Astana) which means the original shot was off target therefore the xG was 0. The total xG for those 5 goals was 1.89 meaning you’d expect Celtic to concede about 0.17 goals per 90 minutes. In fact, we have conceded 0.55 goals per 90 minutes so far. This does not necessarily mean Gordon should have saved the shots, I don’t recall him not saving one that he should have. As I have mentioned before, the xG model I use is crude – so for example McLean’s shot for St Johnstone has an xG of 0.38 but was clearly a Big Chance. More data is required here to get meaningful results plus time permitting I will try and get use of a more sophisticated xG model.

Gordon is very secure on crosses into the box. Over the last 4 seasons, he fails to catch about 1 cross a season.

Key Defensive Saves and Errors

As I showed in the splendiferous article, Keeping Fine, Gordon’s error rate has been creeping up year on year.

This has crept up again to 0.36 per 90m, mainly due to errors kicking out from the backs. Rodgers wants the goalkeepers and defenders to play out, and sometimes that seems like an immutable law overriding all other defensive considerations. The goal conceded to St Johnstone being the nadir as his kick went straight to Craig back to McLean for a tap in.

The other major error leading directly to a goal was against Hearts when he went down very early at the near post as Isma’s shot went over him. A similar scenario unfolded for Twumasi’s second in Astana – another near post shot. Thinking back to Manchester City away last season when he perhaps went down too early to Iheanacho’s shot, there may be a development opportunity here.

Similarly, his key Defensive Save rate has been decreasing over the last 3 seasons. Last season it was 0.61 per 90m. At the moment, it is 0.27 per 90m. This is clearly influenced if there are less saves to make, which given Celtic’s dominance over the last year is a factor.


Gordon’s pass completion rate continues to rise – it was 78% last season and is now 87%. He is also completing more passes – 26.7 last season and now 28.2 per 90m. The fact that his number of clearances per 90m has gone down from 2.18 to 0.55 suggests that the manager is giving clear instruction not to boot the ball away. Whether, given the St Johnstone goal, there will be a recalibration of that approach remains to be seen. Celtic’s possession game is even more pronounced this season – they average 662 completed passes per 90m compared to 535 last season, even hitting 800 away to Kilmarnock.


Following on from Distribution it may seem odd to talk about Creativity in terms of the goalkeeper, but Gordon has been responsible for some fine long passes this season. One against Hearts was particularly memorable.

Gordon takes out the whole Hearts team to set Griffiths through on goal.

Although he has not registered an assist yet, Gordon does manage 3.81 passes per 90m that take out at least one opponent (I don’t count taking out the lone striker who is making no attempt to press). Indeed, 14% of all his passes take out at least one opponent. With Packing and Impect we now have some data and a language to balance the debate about Gordon’s kicking. Rodger’s clearly feels comfortable with the risk/reward ratio.

Early days but the signs were there last season of consistent performance but not necessarily improving performance. A bit more competition would be beneficial? Rodgers may argue he is getting exactly what he is asking for.

4 thoughts on “How Are We ‘Keeping?

  1. […] Gordon reprises the “Astana out of box heading incidentâ€? – this one had a better outcome that in 2016! It’s the international break, and the transfer window is now shut (hurrah!). Celtic content is thin on the ground. So, as we are 11 games into the 2017/18 season, here is the first in a series … Read More […]

    • Excellent appraisal of Gordon. My only observation over the past couple of seasons is the number of times he fails to come for a cross ball when he is in the strongest position to do so, given his height and proximity to the ball.

  2. I agree although it didn’t affect Celtic I blame Gordon for England’s equaliser in the 2-2 draw. The ball came a long distance and dropped in the 6 yard box to an unmarked Harry Kane – felt Gordon with his height and the landing position of the ball should have came and claimed it.

Leave a response

Your email address will not be published. Required fields are marked *