Why Arsenal are phasing out human scouting (Part 2)

Francis Cagigao Had it Coming (Part Two). by AI

(In this part, we look at how human or manual scouting is being gradually phased out by technology).

Why do Arsenal, for instance, reportedly value Matteo Guendouzi at around 40 million euros? And why do the speculated buyers include some of the biggest clubs in Europe?

Simple answer: Guendouzi is a statistical phenomenon. He tops the list of ball-progressing midfielders for his age group and he’s also at the very top across the open list. According to the metrics, Guendouzi is almost as good as the likes Rodri, Jorginho, Granit Xhaka at moving the ball up the field. And he’s just 21.

Clubs nowadays can keep track of a player in Poland’s Third division through a computer. While still not fully realized, sophisticated data analysis can gauge the level of a player with great precision and even more advanced crunching can help determine if a player would be a good fit for your team approach. Mohammed Salah at Liverpool is a prime example of this.

Clubs still predominantly take a more holistic approach to recruitment, with the analytics and scouting departments working together to produce a comprehensive view on targets.

Football is becoming less scouting-reliant. Big data analytics will only get better. And just as how European scouts are aware of the very best young talents at each age group, analysts also do. Clubs give their players at every level performance-tracking wearables. From training to friendlies, every pass and every move is recorded and analysed. The future is almost here: players in the most obscure places will soon be trackable. Why send a scout to Congo to watch a special 15-year-old when your computer could do the same?

If that was just it, Francis Cagigao can make the case that Arsenal still need a complete approach to identification and recruitment that still requires him at the club.

Is that the case, though?

Agboola Israel


  1. I don’t think the restructuring of the scouting department is what bothers people the most but the timing of it,also I think companies must use different criteria,data….when calculating a player’s valuation because I’ve checked and for the same players you have 10,20 even up to 35/40M difference in valuation when it comes to Sancho for example??

    1. Predicting value is like predicting weather. You use statistics to compare to how much similar players were sold for, but ultimately it’s determined by human beings with varying motivations. If nothing else, market values are great for telling us how overvalued a player is when rumours circulate.

  2. Good stuff AI. You may consider me an old cynic, but why, with an array of stats and info readily available, did Arsenal sign the likes of Mustafi, Xhaka and Socratis?At the end of the day it must be down to human error , which will always be with us of course.

    1. At the end of the day scouts can only recommend a list of targets to be signed…. Which players to be signed is decided by the management

    2. I agree it is a good article Grandad. Even with all the data and gizmos in shoes etc the computer doesn’t hire the players

    3. Please watch the movie “Moneyball”. A scout cannot predict entire career of a player. Stats are better predictor than scout. This stats based recruitment is introduced by Oakland A’ and perfected by Boston Red Sox whos owner also owns Liverpool F.C. where same tactics is used. Liverpool is getting the benefit of this approach. Why not we also take the same approach?

      1. We use similar analytics methods but like someone commented recently,sorry I can’t remember who it was,but they wrote that you can have all the data in the world but it is what you do with it that counts/makes the difference which I thought was a good ,valid point!

    4. Very good point Grandad makes about human factors,I don’t think we can predict how a player is going to adapt to a new league,new team& style of play and for foreign players a new country and everything it entails!

  3. You do have to balance the eye and the stats. “Numbers don’t lie” but they sort of do at the same time. Guendozi doesn’t look great to my eye, based on watching Arsenal games.

    But the numbers can take away a lot of the leg work I guess – narrow down where a human is required for the review and assessment. Will be interested to see the third part.

  4. It’s seems that we have forgotten the role of agents and human relationships in the equation.

    Some agents can sell by personal relationship with some management – we have to admit.

  5. Francis, how is Ajax doing their scouting and recruiting right now? They seem to be on quite a roll.

    Is it a combination of metrics and eyes, agent relationships?

  6. Having worked with computers in Civil Engineering and Project Management since the days of the slide rule, they are not independent of human input.
    The old saying “BS in, BS out”.
    To assess the data of the mythical youth player from the Congo, depends on the data collected by a human being, inputed into a “black box” to be analysed and comparisons/recommendations then come out the other end. The quality of the output is very much dependent on the integrity of the input data, ie the ability, judgement, football knowledge of the data collector (scout?)
    My concern is that Arsenal is now putting its faith in super agents, such as Kia Joorabchian to select the players in the shop window; narurally the players will be his, with the push being for those in which he has the most financial interest, without consideration of the benefit to Arsenal.

    1. Agreed. 👍

      It is just human nature, for those businessmen to look after their interests and only their interests.

  7. Stats alone can and often do completely defy the logic that eyes and a keen analytical brain bring far better to the case. Stats but never human logic will tell us that Guendouzi is a top “ball progessor” to use the earticles term. But his constant sideways and back passing does not impress a top scout nor an experienced fan. When eyes and brian on the one hand and stats on the other hand say different things, -AS THEY OFTEN DO – then give me human ability every time.
    Stats used to put the statuesque Mertesacker consistently among the Prems best passers. They never said though that almost all those accurate passes were five yards or so to Koscielny and back, ad infinitum, while in his own half and under no opponent pressure. Did they! Stats, SOME OF THEM, AND ONLY WHEN WHEN CORRECTLY USED, have their place, though a minor place ONLY!

  8. Stats and technology are only as good as the guys interpreting them.

    vaR is the perfect example. We all know if it has improved the game.
    Last year it WAS rubbish.

    Still need scouts to take into things like opposition, games at home and away, weather, early or late part of the season, compliments on the team etc..

    Stats do not paint the complete picture.

  9. Summed up perfectly Jon.An example of how stats can be totally misleading.Player A dribbles past three players, then passes to Player B, who taps the ball two feet to Player C, who scores.Who gets the assist?Based on the computer programme Player B, but as a scout or spectator we know who made the goal.

Comments are closed

Top Blog Sponsors