Two algorithms for predicting the result of an NCAA tournamet:

NCAA Ball Lightning

A purposfully silly algorithm for filling out an NCAA Basketball Tournament bracket

• Conststruct a directed graph of all teams
• Parse the season records to get win/loss records
• An edge from Node A to Node B means team A beat team B in a game
• The weight of an egde reflects
  • how recent the win was
  • how big the score difference was
  • how different the two teams win records are for the season
• To predict a winner:
  • randomly find paths from A to B with n tries
  • randomly find paths from B to A with n tries
  • longest most improbable path wins!

usage: python3 ball_lightning.py seed season_file.txt bracket_file.txt

example season and bracket files in test_data

seed = your favorite integer

NCAA Simulator

An extremely serious algorithm for imputing and simulating games and predicting a March Madness winner

• Read in the win/loss records for all teams
• Do simulated matchups (ignoring conferences and other real world conventions)
• Teams win or lose based on their score records and the score records of their opponents
• Winning teams are awarded wins with scores based on their actual winning games
• Run the tournament!

usage: python3 ncaa_simulator.py seed season_file.txt bracket_file.txt

example season and bracket files in test_data

seed = your favorite integer

Input data from here

https://github.com/lbenz730/NCAA_Hoops/blob/master/3.0_Files/Results/