NCAA Tournament - Bracketology

It's that time of the year when everyone will be filling out their NCAA basketball brackets. I have calculated  some interesting and amazing statistics about the total number of possible brackets in a 64 team tournament. (There are 63 games total in all rounds of the tournament. 32 games  in the first round, 16 games in the 2nd round, 8 in the 3rd, 4 in the 4th, 2 in the 5th, and then the championship game. 32+16+8+4+2+1=63.) Never mind those pesky "buy-in" games. I am ignoring those.

  • The total number of different brackets that are possible is around 9 quintillion. To be more exact, that's  
    9,223,372,036,854,775,808
  • If you could fill out 2 brackets per minute, it would take you 8.7 Trillion Years to complete all the possible combinations of brackets.

So, maybe you need a computer to fill out all of the possible brackets for you? Consider this:

  • If a computer could fill out 1 million brackets per second, it would take over 250,000 years to complete all the possible combinations of brackets.
  • If a computer could fill out 1 billion brackets per second, it would take over 250 years to complete all the possible combinations of brackets.
  • Even if a super fast computer could fill out 1 trillion brackets per second, it would still take over 3 months to complete all the possible combinations of brackets.

The good news is that if you know something about basketball, you know that each game is not a 50/50 proposition. For example, in the first round, a #16 seed has never beaten a #1 seed, and a #15 seed rarely beats a #2 seed.

Depaul University professor Jeff Bergen suggests that by taking a more nuanced approach to the tournament along these lines, we can improve our odds to around one in 128 billion.

That's still a very long shot. If every man, woman and child in the United States completed one bracket using the more nuanced approach of "knowing something about basketball", there would still be less than 1/4 of a percent chance of anyone completing a perfect bracket.