(12-02-2021, 09:33 PM)Derek Hardballs Wrote:(12-02-2021, 08:47 PM)Birdman1811 Wrote:(12-02-2021, 08:45 PM)Derek Hardballs Wrote: Identifying players through stats is a very blunt tool. All seems rather one dimensional, players aren’t statistical data. I know it’s trendy, as is the pressing style of play but it seems like recruitment by numbers.
It was a big part of what we did under Ashworth. Of course the stats are very detailed and you also get full personality reports on players too.
Out of interest, how do you suggest doing it?
Do you really need a dossier full of analysis to work out if x, y, z player fits into the way you want to play? I don’t need to count how many passes Sawyers made to tell me he is good at passing, I don’t need to realise he was also a bit lightweight and wouldn’t fit into my plans if I wanted to play an aggressive, pressing style of play. I don’t need stats to tell me Gallagher would be a better fit for example. If I was looking for a natural goalscorer i would look for ahem a player that has scored lots of goals first and foremost. I doubt Lineker would have been signed by many clubs now because his stats would show a player with unbalanced stats.
I would also trust myself and those I surround myself with opinion over any numbers presented to me via a trendy database. I would also want to talk to the player in person not rely on others to judge their characters and motivation.
We signed plenty of poor players when Ashworth was at the club, his hit rate was no better than GM’s in fact probably worse and one relies on a laptop if what you say is true the other relied on the opinion of his Dad driving a Cortina![]()
Stats should not be the driving force behind identifying players. I know it’s old fashioned but ultimately creating a team is an art form not a science. Otherwise the sport would become a glorified Championship Manager algorithm.
Teams across Europe that have become primarily stats focused have generally been successful. The likes of Brentford, Mitjylland, the Red Bull clubs all have recruitment systems based primarily upon data science and statistics. Working heavily with data science myself, you're correct, it's not an exact science, but it is a very successful one when applied correctly. On an individual level players succeed or fail all the time for a multitude of reasons that can't be modelled. However, the idea behind it is that if you identify the right players based upon a set criteria, you'll generally have an upward trend over time with the right models and analyses.
You use the examples of not needing stats to establish that Sawyers is a good passer, but the stats echo that opinion, so instead of having scouts watching multiple games of every player accross the globe, they can quickly use the statistics to instantly tell them what their eyes would recognise after watching multiple matches. Of course the technology is such now that we can instantly have access to ground covered, pressing intensity, pass completion, any metric you can ever think of.
Also, as has been seen on here countless times, the human mind is prone to bias. Conor Townsend has always been a competent left back for us and it took some years to stop referring to him as league one level. The statistics strip away that emotion.
I apologise if I sound patronising, its not my intention, but if you think that it's a one dimensional tool then maybe you don't realise the levels of analyses that can potentially be carried out with the available data. It's not just about on pitch performance, you can also make predictions about how a player is going to adapt to a move based upon their personality, personal circumstances, historical transfers etc. As I've said, some things can't be modelled, but the amount of available data for footballers is monumental. In the higher leagues every movement has been tracked. Surely it makes sense to use what is available.

