Celtic used three goalkeepers in the 2017/18 treble winning season. Overall Celtic conceded nearly a goal a game once again as in the last three seasons. Europe and Scotland provide different challenges however.
The Glossary is a safe pair of hands regarding any unfamiliar terms.
Appearances
Celtic’s Number 1 remains the 35-year-old Gordon. He competed in 45 of the 61 matches Celtic played. De Vries started the season as back up but the on loan Bain took over that role by the seasons end.
Goals Against
Given the different challenges presented by European football (it’s much blimin’ harder!) I have split out the European versus domestic data.
Gordon conceded more goals in 12 European matches than in 33 domestic appearances and De Vries 3 from 2 European games compared to only 4 from 8 domestic games. Which in microcosm shows the difference in difficulty.
Gordon has the highest goals conceded per 90m of 0.944 compared to 0.737 for De Vries and 0.714 for Bain who only played in Scottish matches.
But you know that this does not tell us anything about chance quality. If we compare the Expected Goals of all the shots that were on target versus the number of goals actually conceded then it shows Gordon let in 8.92 more goals than expected. Bain conceded 0.718 goals more than the model predicted whilst De Vries conceded 1.201 LESS goals than expected.
Normalising this to a per 90 minute average, Gordon conceded 0.2 goals per 90m more than expected. Bain conceded 0.103 more and De Vries saved 0.272 goals per game more than expected.
Saves
Gordon saved 14.877 goals over the season which is 0.334 per 90m. However his overall save rate of 70% of shots on target is down 6% on last season. De Vries and Bain had higher save success rates at 73% and 74% respectively although both made less overall saves per 90m.
It is likely with such small sample data that Gordon’s numbers are more reflective of the fact he played all the Champion’s league Group Stage matches whilst the other two’s data sample is very small. That being said, his overall save % is down markedly on last season. Although I do not have the Expected Goal data for last season, this is a worrying sign of regression.
De Vries’s performance in St Petersburg did not improve confidence and mirrored his spell in goal last season where he seemed unable to get to saveable shots. Despite this his domestic form was good.
Bain made a steady start although firm conclusions are difficult without him being tested in the European arena.
Distribution
Starting attacks and maintaining possession are cornerstones of the Rodgers philosophy.
Gordon averages around 6 more passes per 90m than the other two but again this difference may be due to him seeing more of the ball in the European games. Bain achieves the highest pass accuracy and 86%.
Of course this says nothing about the impact (or Impect) of the passes. That is, how many opponents have their passes taken out the game?
This is where De Vries comes into his strengths. With the lowest pass completion of 73% he is taking more risk with his passing but 22% of all his passes successfully take out at least one opponent – over 10% higher than his peers. Indeed his Passing Impect score is 7 more than Gordon and 14 more than Bain per 90m. Rodgers likes his ability to play penetrative passes from deep. Bain’s passing is by far the most conservative but then he was trying to win a contract!
Crosses
Despite received wisdom (I see many comments criticising Gordon for not commanding his box), Gordon catches significantly more crosses than the others. And no there weren’t all in the recent Scottish Cup Final! De Vries caught only a third of the crosses Gordon dealt with.
Key Saves
Key Defensive Saves can be any action that prevents a clear chance or goal. But for a goalkeeper it is most likely to be an unlikely save.
Gordon remains capable of big saves as shown in the League Cup Final from Moult’s diving header at 1-0.
Bain did not need to make any extravagant saves in his spell.
Defensive Errors
Goalkeeping errors tend to be costly but they can also be wayward kicks out which result in a chance being created.
Gordon’s defensive error rate is low compared to the centre backs where anything less than 0.5 per 90m is very good. Many of his errors did come from poor passes. Celtic conceded 3 goals from Gordon errors compared to an xG of 2.09 for those chances.
De Vries only recorded 2 errors but one of them was to let in Kuzyaev’s long range striker in St Petersburg.
Bain made 4 errors but none resulted in a shot for the opposition.
We are dealing with very small samples here.
Summary
Whilst De Vries, at 37, has shown what skills he may have brought to the team a few years ago in his prime – especially the long passing, and capability against modest opposition. However, he struggles in Europe and his relative lack of agility is shown in his low cross interception stats. I don’t see him as a realistic challenger for the Number 1 jersey.
Bain has been brought into challenge Gordon and De Vries and seems to be the backup. His stats so far are modest against domestic opposition. Too little data to reach firm conclusions.
Gordon continues to adapt and improve his passing game and is still capable of outstanding saves. However the drop off in overall save % is worrying as is the differential between goals conceded and expected goals. A temporary blip or signs of permanent regression?
This may be another area exercising the scouts during the very short Summer break.
I enjoy reading your reports on all the different subjects and applaud your hard work. I believe you have come to the conclusion that D de Vries is not good enough for Europe, which I would say also. I would like to see Connor Hazard become third choice behind Bain and Gordon. I think Bain is a very good find as he is quick of his line with good decision making skills.
Thanks as always Iain. I’m not sure about Bain. When He was at Dundee I always thought he played shorter than his height if that makes sense. He has been calm though but untested in Europe.