In 2017 we analyzed all youth players of the Sudamericano U15. Now, two years later, it is interesting to see what happened to them and how that relates to their FBM contribution statistics. We look at all youth players who played at least 3 matches. Normally, we want at least ten matches for the most probable estimation of the worth of a player. Fortunately, one of the strongest features of Bayesian statistics is that even with a few data points Bayesian statistics is still able to draw valid conclusions. That means that Bayesian statistics is ideal for estimating which youth players have the best chance of making it.

In 2017 we analyzed 193 players born in 2002 or 2003. They played between 3 and 7 matches. All youth players get the following probabilities assigned:

Player | Score | Overall | Attack | Defense | Transition & buildup | Reliability | Number of games analyzed | |

C. de Oliveira Costa – Kakà | CF | 6.29 | 99.17 | 96.83 | 75.49 | 96.08 | 91.77 | 5 |

Score is a summation of the other five probabilities. These five probabilities are:

- Overall is the probability that a player is able to contribute to the team in general.
- Attack is the probability that a player is able to contribute to the attack.
- Defense is the probability that a player is able to contribute to the defense.
- Transition & build up is the probability that a player is able to contribute to transitioning & building up.
- Reliability is the probability that the overall, attack, defense and transitioning & buildup probabilities remain the same in the next match. Yet, it also is an indication of how reliable the player is.

Score is a number running from 0 to 10 with players approaching 10 will be the best players in the toughest competitions. So the score of 6.29 for Kakà in the Sudamericano U15 is quite good. We use a score of 2.5 to distinguish between players who are more likely to make it in pro football. Players not able to score 2.5 points are less likely to make it.

In the Sudamericano U15 of 2017 there were 93 youth players who scored 2.5 points or more. There were 100 youth players who scored less than 2.5 points. Two years later we looked at whether they played at all in 2019 and if so how many minutes they played and how valuable the team is that they play for. Here are the results (we have also included the data for the top 30 youth players):

Criterium | Top 30 youth players according to FBM player score | Youth players who scored 2.5 or higher | Youth players who scored less than 2.5 |

Players still playing | 86.67% | 77.42% | 63% |

Average value of the team the player plays for | 7.7 million | 3.6 million euro | 3 million euro |

Average minutes played in 2019 | 601 | 512 | 356 |

As you can see youth players who score well in FBM contribution statistics have a higher chance of still playing two years later. They play for more highly valued teams and they play more minutes.

The above results were achieved by only looking at the FBM player score. Basically, this means letting the computer decide who is the better player. When we actively evaluate these youth players ourselves, we look more closely at the FBM contribution statistics. We remove the attackers who scored 2.5 points or more, but who also have an attack probability of less than 50%. And we remove the defenders who scored 2.5 points or more, but also have an defense probability of less than 50%.

When we evaluate players as described we only keep 78 of the 93 youth players who scored 2.5 points or more. Their results are as follows:

Criterium | Top 30 youth players according to FBM player score | Youth players who scored 2.5 or higher | Youth players who scored less than 2.5 |

Players still playing | 90% | 83.33% | 63% |

Average value of the team the player plays for | 7.7 million | 4.3 million euro | 3 million euro |

Average minutes played in 2019 | 608 | 542 | 356 |

When you compare this second table with the first table, you can see that by not steering blindly on FBM player score, but actually looking at the underlying probabilities, allows you to select the most promising youth players even more accurately. What this means for clubs is that they can have their youth players analyzed and use FBM contribution statistics to determine which players are best to continue working with.

## Leave a Reply