I have updated my rankings to included an expected margin number alongside the original rating, which is the result of a cumulative distribution function of the ratings vector. The old ratings are displayed under the **CDF Rating** heading and the expected margins are displayed under the **EM **header.

The expected margin value is **the expected margin of loss/victory for a given team against an average team** **over 100 possessions. **“Average team,” here, is a hypothetical team with a rating that is the average of all Division I team ratings (Note: This is not the same as a *median* team. My ratings are skewed in the direction of higher ranking teams, which means that the average team is a lot better than the median team.) And 100 possessions were chosen simply because it is a nice number, and the numbers are easier to read than a 1-possession EM.

Expected margin is usually calculated as the difference between offensive efficiency and defensive efficiency. However, my expected margin numbers are based on a model of my ratings.

Expected margin has the benefit of being a linear value. For example, the difference between a +14 EM team and a +10 EM team is the same as the difference between a 0 EM team and a -4 EM team. This makes EM a more useful rating tool than the CDF ratings.