SciSkill Index – Why and how
There are over a hundred thousand professional football players in the world and everyone has an opinion about each individual player. Five years ago, several clubs asked us to help them in their recruitment process, as they found it difficult to find the right players within their budget. The SciSkill Index created the opportunity for clubs to use a unifying index that evaluates the quality and potential of all players in over 200 competitions. This brief explanation will give an understanding about the methodology that we use to evaluate the influence of every individual player on their team.
Last updated: 12-12-2019
When we started the development of the SciSkill Index, our goal was to create an objective way of measuring the quality of all players around the world. To achieve objectivity, we knew that we needed the same input for all players, regardless of whether they played in the Chilean second division or the English Premier League. If you don’t use an equal input, you won’t get an equal output.
The information that is collected by data-gathering companies isn’t the same in every competition around the world, but certain data is always stored: the line-ups, the goals, the substitutions and the outcome of the match. With this input, we decided to focus on the influence that a player has on his team, rather than measuring what he is actually doing on the pitch.
The SciSkill rating measures the influence of a player on his team. It isn’t able to identify the personality or injury proneness of a player, but is able to calculate the influence that the player has on the team he plays for. The Index enables football specialists to use our SciSports platform to flag interesting players with just a few clicks.
Once we were able to estimate the current influence of all active football players, machine-learning algorithms enable us to estimate the potential of a player by assessing the growth of other players with a similar development path. The estimation is an indication of the average SciSkill of players that follow a similar career path. We are well aware that we are not fortune tellers and the life of a football player can change dramatically during his career, but the SciSkill Potential rating enables clubs to scout and compare young players with players in the peak of their career.
As mentioned before, we only use the basic information of each match, because we want each league to have a level playing field. Therefore, the input variables of the SciSkill information are:
- Line-up (including position)
- Type of match (e.g. league, cup, international)
- Competition strength
- Goals scored
- Red cards
The algorithm is an expectation-maximization model, which is an iterative machine learning algorithm that determines the quality of a player based on historical information. The current quality of a player is assessed by training the algorithm on historical data. When we know the line-up of a team, we can predict the outcome of a match (and we’ve outperformed the bookies during validation).
Example: Team A plays against Team B and the expected result of the match is 2-1, but the match actually finishes 3-1. Therefore, the attacking rating of all players of Team A would increase because they have overperformed. This improvement gets distributed amongst the players, in a ratio based on their attacking contribution per position. Similarly, the defensive rating of all players of Team B would decrease based on their defensive contributions.
The SciSkill Index was validated by outperforming the bookmakers. This means that we were better at predicting the outcomes of matches, based on the expected line-up of a team. A dozen clients from the football industry trust the SciSkill Index to assist them in flagging and identifying the right talents and players.