/* ]]> */
Mar 192014

This is my perennial warning cautioning against going with the favorite: if you picked Florida as your champ, you probably already lost.

When you think probabilistically about your chance of winning a pool, with no skill your chance is 1/n, where n is the number of entries, including yours.

Given the traditional scoring of 1,2,4,8,16,32, the thinking goes that it behooves you to pick the champion, because that’s where the points are.

While that’s true, the problem with this thinking is you’re competing with all the other bandwagoners who also pick the same favorite. So your chance, with no skill, is no longer 1/n but 1/(1+Florida_picks)*Florida’s chance of winning it all.

Florida’s chance of winning it all ranges somewhere between 13% and 19%, but the percentage of ESPN and Yahoo pool entries that have Florida as the champ range between 29% and 37%.

Assuming Florida does win, whether you win the pool depends on how well you did in Rounds 1-5 versus the other 1/3 of entrants who also picked Florida.

Let’s say you’re in a pool with 24 other participants. Most likely, about 8 of them picked Florida. So your chance of winning is approximately 15%*1/9= 1.67%.

Consider instead (for a 24-person pool) picking Arizona, which has a similar Win Probability but only about 7% of entrants have picked them, or about 2 other entrants in this 24-person pool. Now your chance of winning is 15%*1/3 = 5%.

The math depends on your pool’s scoring, and the right way to do it is a full Monte Carlo simulation that you probably don’t have time for.

The March Madness Fan_Simulator widget might help. It will simulate the top 5 teams you should have picked as Champion given pools of different sizes. This is a scaled down model focused only on the Championship game, but for most pools it that’s good enough. Enter your pool size a few times to watch how the recommended teams change (or not).

 Posted by on March 19, 2014 at 5:45 pm