It may not be sexy but BracketBrains could improve your chances of winning that office pool for this year’s NCAA basketball tournament.
The site, developed by three Stanford grads (one math, one computer science, and one industrial engineering), analyzes data collected from over 10 years of game results to predict who might win the possible face-offs during March Madness. It’s intended for both casual and professional bettors who want assistance in making their picks, but who don’t have the time or expertise to cull and analyze all the data themselves.
It works by looking at 15 factors in particular, including teams’ average margins of victory, distances from home, and winning percentages. Users can tweak the weight placed on each factor to account for personal intuitions, observations, and theories. This personalization in turn will adjust the system’s predictions, making the site more interactive than a simple prediction sheet.
The current version of BracketBrains is basically version 1.1. Testers of the beta product reportedly did well for themselves over this past year: 22% won prizes in their bracket pools, and 70% of the most serious users thought it had improved their performance.
BracketBrains can be described as a “freemium” product; users can model one region for free, but they’ll have to go for one of three paid packages to access more functionality. 3,000 paying subscribers have signed up so far.
While BracketBrains uses algorithms and historical data to determine its predictions, another site we’ve reviewed called PicksPal uses crowd sourcing to make its picks. Visitors to that site must pay to access its predictions, which are determined by the service’s most prescient users.