Rugby League Analytics: A Possible Path Forward

Rugby league is a sport that does not seem to lend itself much to analysis, let alone analytics. But that is part of what has got me to thinking about it. Myth-making looms large in league: which of those myths would analytics confirm of dispel?

 

A search for ‘rugby league analytics’ does not exactly reveal a lot of findings to draw from. So what may such a thing look like? These are some of my thoughts, based on what I see as the logic of how the game works.

 

It seems reasonably obvious that a team stands a greater chance of scoring the more metres it gains. And a quick look at the metres gained for each National Rugby League team in 2015 suggests a strong positive relationship between metres gained and winning. Hence it seems that one could start to work out a player’s value by how many metres they gain, lose, or stop.

 

The importance of metres gained is why I lean towards the ‘spine’ being more important than the ‘pack’. The ‘spine’ seems to be the major influence of what I see as the most useful thing a team can do during its set of six tackles apart from score: get another six tackles. In particular a well-placed kick on the fifth tackle gives a team another six chances with which to gain metres or score, or better yet yield a try in itself.

 

Conversely giving away penalties seems to me particularly harmful, as it gives the opposition another set of six tackles and from further down the field. Although from looking over the team statistics for errors perhaps these more or less even out over time, unless one of your players is a particular hothead.

 

In terms of the ‘pack’ stopping the other team gaining metres seems to be their main contribution. I thought this may be hard to pick up in the statistics, since it is hard to tell the counterfactual of how many metres the opposition would have run if a team had not stopped them. I also thought tackles may not be much use in terms of picking winners, as it seems not to be in Australian Rules, as it may be more indicative of a team that tends not to have the ball. But again looking at the NRL team statistics from 2015, the more successful teams do look like they tended to have more tackles. They also look to have lower missed-tackle-to-tackle ratios.

 

I think then that a possible way to value league players is in terms of their net metres – that is, adding together their metres run or stopped, less their metres lost or given away. Obviously league is still a team game, and it may be hard sometimes to ascribe a particular play to a particular player. But even these simple steps may reveal a bit about the ‘true’ value of players.

 

Of course if anyone has done anything like this already, let me know …