March Madness

[shabby]

The NCAA basketball tournament is upon us. I fell in love with this event (I have no great feeling for the NBA) when I spent three weeks in Boston in March 2000.

I don’t profess any particular insight myself, in betting terms, but I know a man who does and he has some devotees, I think, on this forum. Nate Silver developed predictive models for Sports, Weather but principally polling forecasts for elections which generally turn out to be very accurate. His current book ‘ The Signal and the Noise’ is a fantastic read and explores everything from the financial crisis to baseball statistics. I heartily recommend it.

Back to punting, Mr Silver has developed a model for predicting the winners of the NCAA tournament that has done well in recent years and meets the needs of the craze of doing ‘the bracket’ as it is known in the states .
http://www.nytimes.com/interactive/2013 … html?_r=1&

What is interesting is that his predictions are at considerable variance with the odds offered by some bookmakers on this side of the pond. Paddy Power seem to be the biggest player over here and offer some interesting outrights and proposition bets. I have summarised the biggest opportunities (if Silver is right) and the variances. First the outrights.

Outright
Most odds are in line with NS ‘538’ model (+ a bit of margin) but the two at the head of the market could be overpriced.
Louisville are No.1 seeds in the Midwest with NS giving them a 22.7% of winning the entire 64 team tournament…broadly speaking that is 7/2 but Paddy Power go 11/2, (it was 6/1 yesterday).
Indiana are given a 19.6% chance of winning the tounament by Nate (4/1) but are 6/1 with PP.
A little bit of value there perhaps but nothing earthshattering…however Paddy Power reckon Indiana have a 7/5 shot of winning their quarter (bracket) whereas the Silver 538 forecast reckons they are a 10/11 shot.

I have backed both Louisville and Indiana outright and also Indiana to win the East at 7/5.

Propositions
Perhaps the best value bet though lies in the propositions. Paddy Power bundle the 4 top seeds in each quarter together and go 15/8 that one of them (Louisville, Indiana, Gonzaga & Kansas) will win the whole tournament. Paddy Power reckon there is a 34.8% chance of that happening. The 538 model however calculates that the chances of this happening are 56.8%, thats about a 4/5 shot. Thats a very big edge, if Silver is better than the Paddy Power compilers.

I have gone in to my maximum at 15/8 on the number 1 seeds with Paddy Power. betfair are currently 1.96 on this but with little liquidity.

This kind of statistical analysis is not for every punter but I like it myself, particularly in sports where I don’t consider myself and expert but do want an interest and that sums up the NCAA tournament for me.

Good luck

[shabby]

Well Nate Silver was indeed better than the Paddy Power compilers.
Louisville who Silver forecasted as a 7/2 shot before the tournament but were available over here at 6/1 won the whole thing last night beating Michigan 82-76 in high scoring and dramatic game.
The proposition bet at 15/8 was also landed (Silver suggested it was a 4/5 shot) with Louisville winning.
Indiana went out earlier so that bet went down.

6pts bet and 19.8pts returned for a tidy profit.

Once again I would recommend ‘The Signal and the Noise’ as a cracking read.

cheers

[Author]

Posts