Phil Clarke looks at the most important statistics in rugby league and how they can help us predict the outcome of a game
Phil Clarke explains how the new 'performance gauge' checks to see if the score reflects the statistics.
Last Updated: 20/03/13 9:30am
I have to hold my hands up and say that I was the driving force behind the attempt to bring predictive analytics into the Super League coverage and I also admit that I got it badly wrong. The Margin Meter that we used in games didn't work. We made several errors and were foolish enough to think that we could extrapolate the current stats over the remaining time of a game to forecast the final result or margin of victory.
We received some strong criticism over our attempt and have spent the last twelve months trying to learn from the mistakes that we made. Other sports like football, rugby union and even rugby league in Australia are also attempting to use a similar approach to give the viewer greater insight into the game. Basically we want to know if the numbers on the scoreboard are backed up by the stats. Is the current score what you'd expect bearing in mind the balance of play?
At the games, we try to measure the attacking, defensive and possession statistics and see if they support what the scoreboard says.
We have an expert helping us with all the calculations as I'm not bright enough to do the maths. Bill Gerrard has a wealth of knowledge in sporting statistics and has done work for Billy Beane in the past (see Moneyball). His analysis of the hundreds of previous games, and the stats from them, tells us that the fundamentals of the game explain around 80 per cent of the outcomes. This is a high level of accuracy to calculate what the scoreboard should be.
We've learned what the Key Performance Indicators are and how to 'strip the dread from the data' as the author of a book called 'Naked Statistics' describes it. Breaks, metres and tackle success are the three most important statistics in analysing the game.
Rugby league is a complex sport and I'm not trying to simplify it down to a few numbers. I realise, appreciate and understand the creativity and effort required to break through your opponents defence. I also know how important 'breaks' are in determining where in the league table a team will finish after 27 Rounds. In addition to 'metres made' and 'tackle success', they are a great guide to show us which team will finish in which places when we get to the end of the season.
Some of you may have seen the performance gauge in the bottom of the screen on live games recently. We've tried to show both the most important numbers and represent them in a traffic light system, with green meaning good, amber as average and red as poor. A team with three green lights playing against a team with three red ones will almost always be in front on the scoreboard. But there are some occasions where one team may have had an interception try and one through another freak event, be leading by ten points, but being outplayed in the main statistics. The performance gauge checks to see if the score matches the momentum of the match. Most of the time, the scoreboard reflects the stats and it's just a matter of time before it catches up.
Switching to green
Winning teams are ones that have the ability to switch on the green lights when they need to. They have the ability to change the momentum of the match and accelerate when the opportunity arises.
The digital age in which we live means that there is more information and data available than ever before. We're just attempting to strip the 'noise from the signal', as Nate Silver describes it. With all of the things that happen in the game, we're trying to pull out the most important ones and provide a snapshot of the statistics in an easily digestible way.
A Swedish mathematician called Andrejs Dunkels said that it's easy to lie with statistics, but it's hard to tell the truth without them. I want to ensure that the comments I make are evidence based and not just a gut feeling.