Returntomeanology

There is a statistical concept known as the return to mean. It is a little counter-intuitive, in that it says that when you have two tall parents, you should expect a shorter kid, not a taller one (and vice versa). In other words, when you’re at the extremes of a given distribution, you’re more likely to return to the average than drift further out. I sometimes find myself wondering about this in the context of the great game of league.

Are the Titans/Dragons/Dogs really that good? Or will they regress a little during the playoffs? Are the Eels really in comp-winning form or will they return to a slightly less stellar mean, as the pathetic Tigers recently have? Are the Broncos as bad as their mini-train wreck a few rounds back suggested, or are they more like a top 5 team? Are the Sharks and Roosters really that bad? Yes, they really are.

Even statistics have their limit.

Underneath all the noise of attitudinal corrections and injury crises, kind draws and away ground hoodoos, can we discern an underlying trend in a team’s performance? And what’s the proper time period to understand a team’s form? 10 weeks? One season? Several seasons? Who could be the judge of a team’s ‘true’ trajectory anyway? And is that what coaches really get paid for? To change their team’s mean and understand the other team’s? Of all these deep thoughts, one spoke loudest in my mind: why spend time working it out myself when I can hire a consultancy to do a worse job for more money?

Ike Arnby Levitzer Consultancy’s report has just landed on my desk, complete with a summary of the key findings. (Tell me this, why should a study have ‘key’ findings? Can’t they all be key? Why should there be something that ‘it all boils down to’? Maybe when you boil it down it just burns and gets stuck to the pan, isn’t that possible? As humans, we love seeking the ‘essence’ of things, but sometimes, as when I was trying to bake a vanilla cake on the weekend, there just is no essence.)

Key findings, based on where team sits on ladder (ie not based on any recent winning/losing streak): Teams performing above their mean include the Dogs, StGeorge (just), Penrith and the Knights. Teams performing below their mean are Sea Eagles and Warriors. Titans, Storm, Broncos, Eels, Tigers, Cowboys, Souths and Raiders are all sitting about spot on their mean. I couldn’t get a read on the Sharks or Roosters, the consultancy wanted extra money to sit through their games.

What do I make of their analysis? What does it tell us about the semi-finals? Or season 2010? Unfortunately I ran out of money to pay a consultancy to answer those questions, so if you have any idea please drop us a comment.