How AI is changing the face of sports

Yesterday was Deadline Day for Premier League transfers, the culmination of a frantic effort from clubs to bolster their strengths, address their weaknesses and fill the gaps in their line-ups. During period like this, understanding exactly what the team needs – and exactly which players can give them what they need – has never been more important for clubs.

2021-08-31

AI’s biggest success stories usually come from situations like this: taking processes that are already hyper-optimised and finding an extra inch of efficiency, insight or scale that provides a new competitive edge.

But what role can AI play in the world of professional sport? How can teams turn insight into action?

Using AI to optimise rugby team tactics

Rugby coaches prepare their teams for matches by analysing the opposition’s tactics and using publicly available data on the movement of players around the field during a game.

A premiership rugby club was looking to find an extra competitive edge with their analysis and asked us to build them a model that would provide deeper, more sophisticated insights into their tactics.

To do this, we built a database that tracked more than a million actions taken by players on the pitch. We were then able to build a machine learning model that uses the performance of previous lineout routines to predict the success of a particular routine based on:

Its position on the touch line
The number of players involved and
The areas of the pitch where the routine is targeted.

We created a dashboard that coaches were able to easily interpret and filter by types of plays and their location on the pitch – giving the team in-depth insight into the factors that determine their success and failure in the course of a game. The tool is now routinely used by coaches during preparation for games.

Using AI to recruit up-and-coming rugby players

Teams on the lookout for new players have an astounding abundance of data on the performance of every individual professional player throughout every game at their disposal.

This abundance of data can make scouting for new players incredibly time-consuming. Recruitment can also be biased – it’s easy to gravitate towards the most well-known players, based on their fame or previous performance.

The club asked us to develop a machine learning model that would help increase the efficiency and effectiveness of their player recruitment process. We created a model that allows club members to find suitable players based on having similar characteristics of any chosen professional rugby player.

With this capability, the club can start with a player – either an existing, high-performing club player or a ‘wishlist’ player – who fits their desired profile and use their statistics to quickly create a shortlist of players who fit a similar profile. Vitally, this also allows clubs to identify players who have previously been undervalued and, therefore, might have been overlooked during a manual search.

The club now routinely uses the model as an integral part of its game preparation.

The future of AI in sport

These applications of AI are only the beginning for sports teams; the industry is so rich with data that the development of new, more innovative ways to win a competitive edge are all but inevitable.