BEBR launched a new cool product: Combining advanced match statistics with Machine Learning models to automatically identify which actions on the field have the biggest impact on the outcome of the game. We can then use this to accurately evaluate the match impact of individual players, objectively compare players with each other by swapping one player for another, and use the impact scores along with variables such as age, years left on the contract and social media metrics, to accurately predict the transfer value.
The most commonly used source for the market value of players is transfermarkt.com. How is it possible that in this multi billion dollar industry, the most commonly quoted predictions are based on a “community-based gut feeling''?
Data & Performance
BEBR is doing it the data driven way. We have been collecting and engineering terabytes of data from football players, coaches and clubs. As an example, one of these data sets contains all match events and player locations every tenth of a second! By feeding this data into our Neural Network, we have managed to get some insane results. At the time of writing, our predictions already outperform the competitors (a.o. Transfermarkt) market value estimations by over 40%!