I will freely admit that I’m not a knowledgeable person when it comes to the actual game of football. As much as I enjoy watching it, I cannot analyse it like the likes of Ryan Deeney, Blues Breakdown or Blues Analytics can on Twitter and I’m happy to trust in their expertise.
However, as a maths nerd I believe that with the right kind of objective data, I can maybe make arguments that have some basis in fact.
I’m a fan of sports like baseball which use deep statistical data and I wanted to see what I can do of the same ilk.
I spent the weekend updating a spreadsheet I first made last year, tracking the performance of players on a match-by-match basis.
I hoped by doing this I could find some trends in the data to try to back up what we’re seeing with our eyes and to give us a better understanding of how Blues are performing.
You can see the raw data at this link if that kind of thing interests you.
The data is taken from whoscored.com, who use an objective algorithm to judge player ratings. You can read more about that works at this link.
Best and Worst performers
As Blues fans we all have our favourites and scapegoats within the team.
One of the enduring things about football is that it’s a game of opinions and as such one man’s meat can be another man’s poison.
However, I think with a large enough statistical dataset it’s possible to be a little bit more objective about how good players are, and how consistently they perform for the team.
To try to work this out, I have taken all the ratings for Blues players from league matches this season and have worked out the mean and median scores.
I’ve also kept track of the highest and lowest ratings a player have picked up so I can calculate the difference between them.
The mean or average rating for a player is the simplest thing to work out and gives you an idea of which player has performed the best.
This is calculated by adding all ratings together and dividing it by the number of ratings.
I’ve also calculated the median rating, as mean scores can be skewed by outlying great or awful performances.
The median score is the score in the middle between the best and worst performances, which when compared to the mean helps us to understand any skews.
Using this data we can infer the following:
The highest rated player we have is Ethan Laird, who in 10 games has an average of 7.18.
This suggests that when Laird is fit he absolutely must play and that Blues must do what they can to stop him getting injured.
The best performing player over the whole season is Siriki Dembele, who has an average of 6.97 over 26 games.
However, from looking at the median scores, I’d suggest the player with the most consistent good performances is Juninho Bacuna, who has a mean and median of 6.96 over 29 games.
Bearing in mind the number of differing positions Bacuna has played this score truly shows his value to the team currently and has made a mockery of my personal feelings that Bacuna is somewhat inconsistent.
One of the problems with the whoscored dataset is that ratings are almost always lower for players who come on as subs, and as such if a player has made a lot of sub appearances it will be difficult for them to have a high average rating.
This is why I’m a bit reluctant to point out specific players as being “the poorest”, as there are mitigating factors.
That being said, the stats do seem to confirm the eye test that Scott Hogan (6.19) and Oliver Burke (6.27) have not performed well at all this year.
This backs up the feeling from most Blues fans that the forward line is the most pressing need for Blues to invest in – and that the fact we didn’t (or couldn’t) in January may well cost us dearly.
Performances under different managers
One of the things I wanted to do with this data was to see if I could conclusively prove if the team were performing much worse under Rooney compared to the Eustace era.
To try to calculate this, I averaged out player ratings for each game and then averaged that rating out over the period each manager has been in charge.
The conclusions that can be drawn from this data are quite interesting.
Under Wayne Rooney, the average rating for a player in any game was 6.54.
This compares unfavourably both to John Eustace this season (6.69) and last season (6.63) – but not as unfavourably as I thought it would.
The best rated team performance this season was under Eustace (Bristol City away, 7.26), but that hasn’t matched the best of last season (the 2-0 win away against Hull in October 2022, 7.41).
Strangely, the lowest rated team performance this season – Southampton away at 6.21 is more than a quarter of a point higher than the worst performance of last season (Burnley away, 5.94).
This means statistically while Blues haven’t quite been as good as they were last season, they have been more consistent in their performances.
I also split the player data so that I could compare how they performed under various managers.
I did this because while Blues undoubtedly did worse under Rooney, I was sure that some players had benefitted from Rooney picking them when he was manager.
The main beneficiary I believed was Jordan James, and the stats back it up – under Eustace this season his average rating was 6.22 while under Rooney it was 6.69.
This will undoubtedly have been helped by James starting more games under Rooney but I do think Rooney’s tactics leant themselves to picking James more than Eustace’s did.
It’s interesting to note that both Cody Drameh and Juninho Bacuna did better under Rooney too, although by smaller margins than James.
It did feel like some players regressed under the reign of Rooney and the stats back that up too.
From what I saw I thought Dion Sanderson struggled with the way Rooney set up the team and his stats show that his performances declined. Under Eustace his average rating was 6.74 while under Rooney it was just 6.40.
I also thought Ivan Sunjic struggled with the way Rooney wanted to play and his stats show he regressed too, going from an average of 6.72 to 6.46.
It’ll be interesting to see how those figures change as Blues play more games under Tony Mowbray.
Decisions for the future
As the season draws to a close there are a number of players whose contracts with the club come to an end.
This has naturally come up in discussions between Blues fans online and while there are a number of external factors which will come into play, I think it’s interesting to use the stats to see who might be worth giving an extra year.
The first decisions will be on loan players – whether Blues will make any attempt to convert them into permanent signings.
I’ve long thought that neither Emmanuel Aiwu nor Oliver Burke have been worth keeping hold of and the stats seem to confirm that neither have performed well enough to consider signing.
While the option to sign Aiwu probably ensured there was no way we could send him back, I do wonder if Blues wanted to send Oliver Burke only to be denied.
Tony Mowbray made a vague remark about players having to agree to be moved on and not doing so which prevented us from letting anyone go; without any knowledge at all of who he was talking about part of me does wonder if Burke was asked if he would terminate his loan and he declined.
The stats also show the absolute need for Blues to sign at least a couple of forwards in the summer.
Hogan’s performance ratings match the feeling I had that the former Brentford man was done as a Blues player while Lukas Jutkiewicz might also be a victim of his advancing age now.
The stats have actually convinced me that even though I’m unconvinced that Junior Dixon is good enough based on what I’ve seen from him playing, I’d rather him in the squad than Hogan now.
I don’t think that will happen, but I do think that we’re more likely to see a false nine being deployed by Mowbray as he did at Sunderland, especially if and when Stansfield needs a break.
From the out of contract players, I don’t think many can expect to be offered a new deal based on these stats.
The only players I’d consider based on the stats have other issues which might preclude them getting deals; Keshi Anderson probably needs to show he’s injury free while age might do for John Ruddy and Kevin Long.
As much as I’m reminded by the famous quote about the three kinds of lies being “lies, damned lies and statistics”, I do think that they can help us to be objective in our opinions.
If I can find the time and motivation, I’m planning to go back through older whoscored data to build datasets for previous seasons so that I can build better comparisons between players and managers.
The data I have right now show that there are some definite causes for concern within the squad and the performance levels under Mowbray in the league confirm his concerns about the lack of quality coming off the bench.
However, I think we have to be positive and accept that Mowbray has spent the last few weeks working out who he can trust to perform in his squad.
I think we’re now at a point where Mowbray has a good idea who he can play and who he can’t – and those who haven’t performed are going to have to get used to watching from the stands on a Saturday.