Welcome to the online home of the #KPI. This day has been 10 years in the making, and I’ve finally decided to share a formula and data that has helped shape how I’ve thought about sports for a long time.
Over time, I will prove through analytics how precise and accurate the formula is. Numbers, interpretations and rankings will be updated regularly. I’ll look to have some fun, too, both here and with the debut of the site’s Twitter @KPIsports.
The details of the formula can be found here. I’ve also compiled data from the 2012-13 college football and college basketball seasons to allow readers to sample how the math works. The story of how the formula came to be can be found here.
The #KPI measures teams by assigning a value for each team as part of each game (that’s 11,344 values for 5,672 college basketball games in 2012-13 as of Selection Sunday). A team’s season-long #KPI is determined by the sum of each game’s #KPI, divided by games played. Each game is worth exactly the same. In the weighted-RPI, a game can be worth 0.6, 1.0 or 1.4 games depending on its location. The #KPI values the numerator of the equation only, and does not mess with the denominator as the RPI does.
The #KPI also incorporates games against Non-Division I (in basketball) or FCS (in football) competition. It is data and it is a game played, so it counts. Every game is good data to measure a team’s overall performance. Nothing is being left out.
After a decade’s worth of compiling data, I am convinced this is the most accurate metric in judging a team’s performance in a season (Notice I wrote it judges performance, not which team is necessarily better). Stay tuned this Fall and Winter as the formula proves itself. Welcome aboard!
This is the #KPI.