What does a test for a rating system even look like? Can we get a slice of historical data from somewhere and run it against our rating system, and check that the output is approximately right? Some sort of crude visualization tool to see how player ratings react to an event would be useful. (e.g. a numpy script that plots a line diagram of peoples' ratings over time.)
What does a test for a rating system even look like? Can we get a slice of historical data from somewhere and run it against our rating system, and check that the output is approximately right? Some sort of crude visualization tool to see how player ratings react to an event would be useful. (e.g. a numpy script that plots a line diagram of peoples' ratings over time.)