Tuesday, September 25, 2007

Bayesian Superbowl & POTUS outlooks

Tradesports provides probability markets in Super Bowl and Conference victories. It logically follows that a team can only win the Super Bowl if it wins its conference, i.e. p (B | A) = 1. Bayes’ Theorem provides the probablity of a team winning the Super Bowl upon winning its conference, i.e. p (A | B).

As of 1245pm Tue Sep 25 2007, here’s what p (A | B) looked like:

78% PATRIOTS
75% TEXANS
62% CHARGERS
62% COLTS
57% JETS
56% STEELERS
55% RAVENS

Notice anything interesting? Yep, they are all AFC teams. The Jaguars and NY Giants are both the highest, at 50%. (NB: Liquidity is still building up, and these are calculated using mid-market, which may be at varying levels of freshness).

UPDATE: Eddy Elfenbein does the same thing for 2008 candidates for President:
67% CLINTON
65% GORE
54% EDWARDS
50% MCCAIN
50% OBAMA
46% GIULIANI
39% THOMPSON
39% ROMNEY
35% PAUL

The Dems look to be the AFC of the election cycle.

No comments:

Post a Comment