Bright Starts: Predicting Players Primed for a Breakout Season

It’s true that young players develop at different speeds and there are lots of examples of “false starts” where young players make their debut and think they’ve cracked it only to be back warming the bench again a few games later. Therefore, predicting break out seasons can be difficult. However, predicting long term player success from relatively small samples of game time is one of the most interesting and effective uses of big data in football. Analytics FC Managing Director, Jeremy Steele, attempts to push the limits and unearth some top talent on the smallest of sample sizes.

The ability to “get there first” is a hugely important factor in player recruitment. Of course, some clubs have the financial clout which allows them to arrive late in the day and take a player from the under noses of others – just look at the Malcom deal last summer where Barcelona swooped in at the last moment leaving Roma fans waiting for his arrival at the airport and Sporting Director Monchi threatening to take Barcelona to court over the incident!

However, early identification and fast (but robust) due diligence can play a big part in a successful recruitment strategy, not to mention avoiding spiralling costs incurred by leaving deals until the last minute of a transfer window.

This way of thinking was an important part of the way in which we built the algorithm which underpins our TransferLab data scouting platform. We wanted to avoid the methods commonly used by football data analysts who tend towards sub-sets of the data to identify various single use metrics or smaller “chains” of events. The system was instead built in a way which maximises the information available to provide a framework with two key advantages when looking at young players who are breaking onto the scene:

1.   It uses every event captured within the data-set during each match and therefore results in robust predictive projections much earlier than other models, which often discard large amounts of match data to focus on one or two metrics.

2.  The nature of our algorithm means that outputs have a much lower dependence (although there are always some) upon the quality of teammates or overall team quality. Playing style effects are also less prevalent in our model.

These two factors are obviously key in providing much earlier identification of high performing players and also being confident that the performance is not just due to playing in a high performing team.

To illustrate the algorithm at work we’ve identified a few players who are showing signs of a “bright start” in their domestic league for you to check out next season and follow their progress:

Hugo Fernandez (Olimpia): Pacy and direct, Fernandez’s dribbling ability has been putting defences on the back foot in the Paraguayan Premier League this season. Fernandez’s expected goals numbers need some improvement but his creative numbers are very strong.


Jack Clarke (Leeds United): After only 863 minutes in the Championship, Clarke has already shown enough to have several Premier Leagues interested in his signature this summer. His underlying numbers are extremely good for an 18 year old plying his trade in one of the most challenging leagues in Europe.


Jules Keita (Dijon): After only 484 minutes in Ligue Un, Keita already looks a promising talent. A forward with quick feet, pace and no lack of confidence. Earlier this year he said, “Neymar is my idol. In Guinea they call me ‘baby Neymar‘….I could be better than him”.


Mohammed Ihattaren (PSV Eindhoven): After reportedly turning down moves to Chelsea and Manchester United to continue his development at PSV, Ihattaren made a big impact after breaking into the first team just after Christmas. The Dutch youngster is creative with his passing, dangerous in 1v1 situations, and may well be a future star for the Oranje.


Santiago Rodriguez (Nacional): The Young Uruguayan U20 international has made a big impact in his domestic league since making his debut in February. His ball progression and creative numbers are off the chart for a 19 year old.


Orkun Kökçü (Feyenoord): In only 532 minutes in the Eredivisie, Kökçü’s scored 3 goals and assisted a further 4. His data shows excellent all round output for a Central Midfielder. The early signs are very promising.


Peter Vindahl Jensen (FC Nordsjælland): Young Goalkeeper Jensen has actually not been on the winning side in any of his first six appearances in senior football. However, that doesn’t detract from his impressive performances between the sticks at FCN.


Of course, some of these early, bright performances could be “false starts” and these players might be perceived as a ‘risk’ in that sense. However, you are likely to get a much lower price moving early than waiting for a larger sample size, and so the uncertainty can be factored into any potential deal. If you wait too long, you might miss out!

For information on our data services, including our data scouting platform, TransferLab, visit

At Analytics FC, we provide software and data services to entities within football looking to realise the gains possible from analytical thinking.

Find out more about us, or get in touch if you have a question!

News, straight to your inbox

Provide your email address to subscribe and get email updates