AI XI – Could technology be set to revolutionise Fantasy Football?

Andy By Andy

The domestic football season has finally come to an end, which means a well-earned rest from the action at Spark HQ, as our Fantasy Premier League (FPL) competition has now reached its climax. As I wrote last year, we have our own somewhat hotly-contested league at Spark; this season Michelle used her wits to lead from the front for almost the whole season, and deservedly took the crown after fending off late challenges from Gus, Chris and some idiot who kept on fielding Bournemouth players.

As we take a summer break, I’ve been considering how to do things differently and be more successful next season. Maybe technology could hold the answer to an improved performance?

Tech is certainly driving change: in real-world football, we’re about to see the roll-out of Video Assistant Referees at the 2018 World Cup, and things are getting even smarter in FPL, where Artificial Intelligence algorithms are being trained to play the game. Dr Gopal Ramchurn and Dr Tim Matthews, from the University of Southampton, created Squadguru to investigate whether algorithms were capable of driving success. After running simulations of matches to predict what could happen, it recommends any changes that should be made to the team. The trial has seen some successful results, finishing this season in the top 0.7% out of 5.9 million players worldwide.

So what does this mean for the future of FPL - could AI make human managers redundant? Well no, not immediately - surely it cannot understand off the field events at the moment – issues such as player fallings-out, transfer rumours, and lack of motivation can have a huge impact. The team at the University of Southampton understand this is the next battlefield - to make their algorithm even better, they want to build in a perception of human emotion. Once AI is able to build an understanding of these wider emotional factors, it would be in an even stronger position to win a game of fantasy football; looking from a wider lens, this could also lead AI to play a more integrated role in society.