https://github.com/llimllib/ncaa-bracket-randomizer

Uses Ken Pomeroy's College Basketball rankings plus randomness to generate a bracket for you
https://github.com/llimllib/ncaa-bracket-randomizer

Last synced: about 1 month ago
JSON representation

Uses Ken Pomeroy's College Basketball rankings plus randomness to generate a bracket for you

• Host: GitHub
• URL: https://github.com/llimllib/ncaa-bracket-randomizer
• Owner: llimllib
• Created: 2009-03-18T19:10:14.000Z (almost 16 years ago)
• Default Branch: main
• Last Pushed: 2024-06-19T11:30:54.000Z (7 months ago)
• Last Synced: 2024-11-30T02:50:44.691Z (about 1 month ago)
• Language: Jupyter Notebook
• Homepage: https://llimllib.github.io/ncaa-bracket-randomizer/
• Size: 957 KB
• Stars: 21
• Watchers: 8
• Forks: 6
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# NCAA Bracket Randomizer

Generates a reasonable bracket for you, based on Ken Pomeroy's

college basketball ratings.

This code was a hack when I first wrote it 15 years ago, another hack when I

re-wrote it 9 years ago, and even worse today. I use it once a year though, and

it works just fine.

For each matchup, it compares the two teams in the game, calculates the odds of the favorite

winning the game, and chooses the winner. There are three settings that determine how it

does so:

#### No randomness

Always picks the favorite

#### Some

Picks the underdog if they win twice in a row.

If team A has a 75% chance of beating team B, the program will generate two random numbers

between 0 and 100; if both of them are greater than 75, team B will advance. Otherwise,

team A will advance.

#### Lots of randomness

Picks the underdog if they win once

If team A has a 75% chance of beating team B, the program will generate a random number

between 0 and 100; if it's greater than 75, team B will advance. Otherwise,

team A will advance.

## Calculating the Odds

[This ipython notebook](https://github.com/llimllib/ncaa-bracket-randomizer/blob/master/fitting_kenpom/fitting%20kenpom.ipynb) shows how I fit a very simple exponential curve to kenpom's predictions to reverse-engineer his win percentages. I then use the curve that fitted his predctions to decide how likely one team is to beat another, based on the difference between their kenpom ratings.