A. I. Player Scaler

Universal performance equalizer for Apple to Apple player comparison

Universal performance equalizer for Apple to Apple player comparison

Each league around the globe has its own quality and level of play. Among leagues of various qualities and popularities, each team distincts itself with its own dynamics. But it all comes down to players to form the aggregated dynamics of a team or league. Each player possesses unique attributes, strength and different playing styles. Obviously, the performance of a player is dictated by the strength of his/her opponents and also teammates. Therefore, when it comes to define an unbiased comparison, all three of these dimensions need to be assessed.

Product Demo

Designing the Algorithm

Deriving this unbiased comparison of players is no easy task with only recent performance data of players among various leagues and requires the utilization of intelligent features which Comparisonator team derived them from GBE component of each player,

– International Appearance
– Amount of domestic playing minutes
– Minutes coming from continental tournaments
– Continental progression of players’ last club
– League position of the last club

8 different AI backed algorithm is executed to define player’s unique performance level. This can be seen as a global currency in circulation, where proper, unbiased comparison of players can be done.

Due to complexity of the task, only advanced machine learning algorithms can provide the expected performance. The program investigates every action of the player in the world, calculates every single parameter by adding 8 different weights to create weekly scenarios. Comparisonator separated this into 5 different separate scala

Comparisonator Main Index

Combinations of defensive, offensive, duels and passing actions form the Main Index. The artificially intelligent weight system implements two other characteristics to define the player level even deeper. If the player is a regular player and has more minutes compared to others in the same position, preferred understanding to use “total” collective index points to make a decision on the level. If the player spent less minutes on the field, then “AVG per 90min” can be preferred. By this way, player who has less minutes are within the calculation scala of players with more minutes, kind of a mechanical equalization process.

Offensive-Defensive-Duels and Passing Indexes respectively cover their own areas of parameter ranges, in order to understand specific performances.

You may click on the white paper section of Comparisonator website to understand this system scientifically

Comparisonator team is proud and happy to present its brand new home-bred football player AI Player Scaler developed by its AI engineers, a refined result of a 1 year R&D effort.