FOLLOW

FOLLOW

SHARE

Data Science At The Rugby World Cup

How is the RWC and rugby in general using data in their sport?

30Sep

Rugby Union is - relative to soccer at least, where a game can be decided by a single lucky shot - a fairly predictable game in terms of results. Big teams are rarely beaten by the smaller teams because of the gulf between professional and amateur rugby. This makes South Africa’s defeat by Japan all the more stunning, particularly to the statisticians, none of whom will have predicted the result correctly.

These sorts of results do not, however, suggest that data is of no use in rugby union. It used to be that rugby coaches could get away with using personal experience and gut instinct for decision making. This is no longer good enough for most clubs, with many having adopted an evidence-based approach to performance that relies on data science and analytics.

Wearable technology has vastly increased the number of data points that analysts can use to leverage insights, to optimize both individual and team performance levels level. Players now wear RFID devices throughout both training and matches, so coaches can monitor their fitness levels and ensure they do not suffer from fatigue. They can also monitor areas of their game that need improvement. Such knowledge can help greatly with selection, helping to note which players are fresher and more ready for the game.

Data science has revealed a number of more general points about the game as well, many of which have been long believed by coaches but not proved. For example, teams that use a kicking game tend to be more successful. The less time you spend in your own half, and the more you play in the opposition half, the more likely you are to be successful. This seems like common sense, but by looking at how it is true - running the ball in rugby leads to more collisions and rucks which drain energy, and being turned over near your try line means more tries conceded - you can reinforce methods.

In terms of opposition analysis, statistics providers, such as Opta, can reveal a lot about an opponent’s playing style and find areas of weakness that can be targeted. How often a player kicks, the type of kick, and the area of the field that they tend to do it in can help predict their movements. This information does, however, still require a human to establish what the best method of defence should be, although they can look back at the data from past defence and run models to determine exactly how effective this will be.

Data science is also being utilized by the RFU in order to provide fans with as satisfying an experience as possible, installing sensors all across Twickenham to provide statistics in real time that can be aired on television to drive engagement.

For many of the sport’s purists, drilling the game down to numbers moving around a screen has become something of a bugbear. They argue that numbers are misleading, and they lead to the game becoming predictable. Teams can be unprepared for deviations from what the data expects, which can be easily exploited. Metrics such as tackle completion can also be misleading, with no consideration taken for power, and whether it is behind the goal line. England coach Stuart Lancaster has denied making substitutions based on GPS data, and the reluctance is understandable. It is also important to remember that sport is first and foremost entertainment, and should it turn into a purely kicking game, it is likely that it will turn many fans off, with most preferring the total rugby style of game where the ball is carried in hand more frequently.

Comments

comments powered byDisqus
Bottom of well small

Read next:

Data Mining In The Deep Web

i