How soccer (sic) scouts identify talented players - https://www.tandfonline.com/doi/full/10 ... 21.1916081 . Article here from academics in the Netherlands based on self-report studies with scouts. Key points being:

1. Scouts who observe young (under 12) players don't think they can reliably predict future performance later, yet still make recommendations to their clubs based on 'best current players' rather than on based on who might perform better when they get older - "selecting the best current players at a young age could harm the selection process". So the authors recommend a more responsive (de)selection process, including targeting older children instead.

2. Scouts, although they claim to use a structured analytical approach to identifying talent, this is not done uniformly, and they generally rely on intuition to select players at young ages but "[g]iven previous literature demonstrating that predictions based on overall 'intuitive' impressions are non-optimal in terms of reliability and validity, we recommend that scouts are trained in a more consistent use of the different aspects of structure when predicting performance". Selection is also generally based on technical skills and abilities, rather than other factors that are also key to top players (decision making etc.)

Perhaps reinforces what many of us may think about the recruitment of kids at such young ages, or the release of young players before they've had chance to fully develop. Also made me think about the limitations of AI inspired mechanisms to identify talent such as AI Scout which focuses entirely on technical skills in turn omitting lots of potential players who excel in other facets of performance.

Anyway, an interesting read over a Sunday morning coffee.

Very interesting indeed. It is very difficult to predict if a young player will make it based purely on their current technical ability. The best 11yr old in the world will likely never become the best 21yr old in the world. I often consider a more important factor is a players ability to take coaching.

I used to work alongside someone who was a youth footballer for Chelsea. He often talked about the one player who had a successful professional career. They were described as being one of the poorer players in that side. They did, however, have the ability to take on what coaches were saying, had a phenomenal work ethic, and was absolutely driven to succeed. He was released by Chelsea without ever playing a game but was given another chance at West Brom primarily because of those traits. Carlton Palmer.

It is by identifying weaknesses in systems that the system can improve, grow, and mature. This is the very definition of machine learning. What I’m seeing above is nothing more than data-based analysis that the system has weaknesses. Without having the system in the first place it wouldn’t have been possible to correlate or at least have an opinion as to its fallibility.

Once the weakness is identified you can take it into account, compensate for the weakness through coaches’ intuition and experience, and weight the probability of the successful correlation of the youth player themselves.

Without reading the full piece I hope it doesn’t suggest selection is better without the systems, leaving it to instinct/experience? Or does it suggest don’t start to assess at a younger age? If the latter clubs run the risk of “missing out” due to what would be the lack of a system that probability dictates would succeed more than it fails…..

The paper focuses on scouts who observe players from U12 up to U17 who had not yet been signed up to a professional club or academy. I think the ‘weakness in the system’, as you put it, is that these scouts generally felt that they were unable to identify whether the players they observed would make it until the players were older, and so they often reverted to recommending players based on observable characteristics such as technical skills. I don’t think the paper was suggesting a move to intuition and away from data, but pointing out that the current system is probably missing lots of the Carlton Palmers, or those with skill sets that arrive a fair bit later, so they don’t get the chance to be properly coached and assessed.