bracket busters

Show hard words

I'm not doing a March Madness bracket this year.

one can only predict the future so much. best not to waste the gift on popular frivolities.

but: some notes from last year.

----

When filling out a bracket, it is important to realize one's goals. Is it:

  • to win the group
  • to get the best individual score
  • to make picks that allow one to gloat to one's friends

----

To win a bracket, the key is to correctly pick the national champion. If it is a team that nobody else in the group has picked, you will probably win even if the rest of your bracket is average.

----

for other picks, the rule of "take the better seed (that is, the 1 seed, or then the 2 seed, etc.)" is a good starting point.

  • it's no fun to take all the favorites
  • very roughly: games with up to 3 seed difference (aka 7-10) are coin-flips; games with up to 8 seed differential (aka 5-12 or 1-8) are 80% favorites for the higher seed
  • there will probably be one or two top-4 seeds that lose in the first round. trying to pick these is largely futile.
  • the computers are generally pretty good at knowing which team is better. for the men, https://kenpom.com/ is a good baseline rating.

----

other notes:

  • actually having watched the teams play can help to know how good they are
  • if there is a "good story" that isn't too unlikely, pick into it

----

for the women's pool:

  • seed differences mean more here. only a 1-seed difference is a pick-em, and a 4-seed difference is enough for 75-25 odds.
  • picking all of the 1 and 2 seeds to reach the round of 8 is far more likely to be successful for the women's bracket