Win probability in cricket

An attempt to calculate the win probability for every ball in a one-day international (ODI) cricket match.

This project began after I watched India play Zimbabwe in the 2015 Cricket World Cup. India won with 8 balls to spare, despite my belief while watching that Zimbabwe had the game in hand.

Sources:

• https://saidee27.wordpress.com/2014/09/29/in-game-win-probability-twenty20-cricket/
• http://thespread.us/

To get started:

$ python get_scripts/01_download.py
$ python get_scripts/02_makemaster.py
$ python get_scripts/03_calcs.py

This will do the following (and take a few minutes to complete):

1. Grab the latest .zip of ODI data from Cricsheet.org and unpack it to a folder named all_odis.
2. Create two files in master_data folder:

• odi_info.csv (about 0.1 MB) -- information about the game such as location, teams, toss, etc.
• odi_ballbyball.csv (about 50 MB) -- every ball in every ODI match

3. Process the master files into odi_ballinfotrim.csv that merges some ball and match info and has extra columns with basic calculation about the game state.