Ian Sibbit on Twitter:

Ian Sibbit
Margin meter has just predicted grazed knees when I play at widnes.

Being vaguely serious for a sec, it looks like this might be the academic that Phil Clarke was talking about: [url=http://business.leeds.ac.uk/about-us/faculty-staff/member/profile/bill-gerrard/Bill Gerrard[/url. If it is an economist who's come up with the maths behind the margin meter, then it'd be no surprise if what we've got is an economist's model system that attempts to predict future performance on the basis of past performance - just like for the markets. And we know how well they work. It's like picking your fantasy team on last year's tries and appearances.

RL and markets share a feature that makes prediction difficult: they're not simple systems, and there are strong - apparently - random (in jargon: stochastic) effects. That's not to say that games are always unpredictable, though - we've all seen scorelines go exactly how we'd expect. We can pick the likely top four in the league with some confidence. You can build a pretty good fantasy team if you pick last year's kickers (as these are stable points-getters, probably more stable than try counts for a player).

Intuitively, to me, the margin meter is unlikely to be a reliable predictor early in the game - how many times have we seen a strong performance from a lower table side (or GB/England) for 60min, only for the last 20min to decide the game? You might expect there to be some sort of weighting that early events count for less than later events in the game. We might also need to take into account that there is knowledge other than that which comes from the measures Phil Clarke described in his introduction: we 'know that Australia are fitter and stronger', we 'know that Widnes lack discipline', we 'know that Wigan have good goal-line defence', 'we know James Child is refereeing' (in jargon: prior knowledge), and so on.

It doesn't help that the margin meter only reports a single number with no indication of error. "Widnes by 6, +/- 24" might be what the model is actually predicting, but we wouldn't know that from what they show.

Anyhoo - I wouldn't believe its predictions for a second, but I'd be interested to see more technical information about the maths.

Very true, though those previously bought a trophy instead of a plastic pitch

Wonder if this has really been brought in by Skybet?

I can imagine loads of people betting in play with the Margin Meter only to be bamboozled by a scoreline that goes completely against it.

Margin Meter was a success, not from the point of view of predicting the score, but in generating media attention via TWITTER.

It got people talking about the game itself and rugby league.

wonder how wrong the margin meter will be in tommorows game haha

FYP

The use of technology fot technology's sake, pointless

Its embarrassing, imagine tuning in being a first time viewer of SL and seeing that.

Presumably a cardboard one. While he alternates between muttering to himself and shouting at passers-by.

2 litre bottle of unbranded cider and smelling of urine

It was brilliant with 14 minutes left and Wakefield winning by 8. Up pops Phil with his margin thingy saying Wakefield will win by 3 points and Clarke states "it's predicting the points to flow for Wakefield now"