README.md

In bbolker/cricketr: Evaluation of cricket stopping rules

cricketr

When limited-overs cricket games are stopped prematurely for weather-related or other reasons, stopping rules such as the Duckworth-Lewis method are used to determine which team won, i.e. how many runs the team that batted second should have scored (its target), considering the number of overs played and wickets lost, in order to be considered the winner.

While the Duckworth-Lewis method is "an attempt to set a statistically fair target for the second team's innings", its actual statistical properties have rarely been evaluated objectively and independently. This repository includes/will include:

• code for scraping relevant match data from ESPN's web site (as far as we can tell, such scraping is within the terms of service provided the results are not redistributed)
• code for tabulating resource tables (% of resources remaining as a function of number of overs played and wickets lost)
• different methods of estimation of resource surfaces (essentially stoppage-rule tables) from raw resource tables
• tools for evaluating methods (RMSE, kappa, bias, etc.) and applying cross-validation

bbolker/cricketr documentation built on May 11, 2019, 9:28 p.m.