I was asked by a Patreon to look at Lincoln City goalkeeper George Wickens and analyse his performance development over the last three seasons.
I went into it with no preconceptions. All I wanted to do was present three sets of numbers, side by side, and make an assessment, as requested by a reader. Of course, assessing a goalkeeper can be difficult because the obvious statistics are influenced by everybody in front of him. Clean sheets and goals conceded tell us something, but neither separates the goalkeeper’s contribution from the protection his defence provides.
Compare George’s figures from 2024/25, last season and our Championship campaign so far, and some encouraging patterns emerge. There are nuances, particularly around the smaller sample this season, but there is evidence to support the argument that he is progressing at a good rate. The eye test might tell you that anyway, but why not use numbers as well?
The clearest positive is his save percentage. Calculated from the saves and shots against in the last three seasons’ league figures, George stopped approximately 67.4% of the shots he faced in 2024/25, increasing to 71.4% last season and 77.1% this time around. That is almost a 10% improvement from his first season to the current campaign. Put another way, he has gone from saving roughly two-thirds of the shots he faces to more than three-quarters, with the latest increase coming after our move into a higher division.
The workload makes that particularly interesting as well. Last season, George faced 2.80 shots per 90 minutes and made 2.00 saves, whereas this season those figures have risen to 4.07 shots and 3.14 saves. That’s to be expected, but it does pour shade onto the argument that he looked good last season because his defence was so good. Yes, he was called into action less, but the numbers show that when he is needed, he steps up.
He is facing around 45% more shots and making 57% more saves. Yet his goals conceded have increased by only 16%, from 0.80 per 90 to 0.93, which suggests he is coping well with the additional demands. Compare this season with 2024/25, and the picture is similarly encouraging. He is facing around 22% more shots and making approximately 40% more saves, while conceding 15% fewer goals per 90.
We cannot isolate every part of that improvement from the team around him, but those are positive figures for a goalkeeper adjusting to Championship football. More work has not brought a proportionate increase in goals against, and his save percentage has continued moving in the right direction.
Expected conceded goals, or xCG, offer another useful comparison. In 2024/25, George conceded 1.10 goals per 90 against an expected figure of 0.97, leaving a gap of 0.13. In other words, ‘could do better’. Last season, that gap narrowed considerably, with 0.80 conceded against 0.78 expected. This season, the figures are identical at 0.93, meaning his actual concessions match the model’s expectation.
That is evidence of a better relationship between expected and actual outcomes, although it does not show him outperforming the model this season. Equally, we should avoid treating that difference as an exact calculation of goals George has personally saved us, because as we know xG is, at best, a divisive metric.
Speaking of dividing xG, if we divide xCG by shots against gives an expected value per shot of approximately 0.29 in his first season, 0.28 last season and 0.23 this season. Yep, my head hurt writing that, so I get how you feel reading it. It means the average shot he faces currently carries a lower scoring expectation this season. Basically, we’ve been seeing a lot more pot shots and long-range efforts, but those do still need saving.
Distribution provides a more mixed picture. His long-pass accuracy improved from 59.1% to 64.7% between his first two seasons, despite the number attempted rising substantially from 10.18 to 17.56 per 90. This season, he is attempting slightly more again, at 18.14, but accuracy has dropped to 47.4%. That works out at approximately 8.60 completed long passes per 90, compared with 11.36 last season.
It is a clear reduction in completion, although the figures cannot explain how much reflects his kicking, the targets available or the opposition challenging for those balls. If he is delivering the balls into the right areas, but this season the strikers are not winning those headers, then his accuracy numbers are affected. There is also that smaller sample size: 90 minutes aiming for Ben House and Rob Street instead of Freddie Draper or Mikkel Ladefoged is likely to drag down the overall numbers more significantly in an eight-game sample than a 46-game selection.
Short passing has remained much more reliable, with 95.7% accuracy this season despite an increase in volume. His exits have also risen, from 1.47 per 90 last season to 2.56 this term. That shows greater activity away from his line, but without success rates or errors, it cannot establish better judgement.
For me, the strongest argument remains the combination of workload and save percentage. George is being asked to do more, at a higher level, and is stopping a greater proportion of the shots coming his way. Of course, we need a longer Championship sample before making sweeping claims, but these numbers give us a reasonable basis for seeing a goalkeeper developing with the demands placed upon him, and that should encourage us as the season progresses.
Comments Welcome!