Sportemind's College Basketball

In order to make the tournament, you need to do better against your schedule than the worst at-large team would do. That is, you need to win more games than the theoretical 46th best team would win against your schedule. It's easy to calculate how many games you have won. To calculate how many games the 46th best team would win, an expected winning percentage is calculated for each game. This is done utilizing the Pythagorean Expectation. All of these are then summed to find a total expected number of wins. The opponent's value (for an input for the Pythagorean Expectation) is a blend of its Kenpom and NET rankings. This methodology is not just logical, but useful. For one, the cut line is at zero wins. Also, it is easy to calculate how many games you need(ed) to win to make the tournament.

​

This is really useful around the bubble, but isn't great for the top seed lines. Particularly, it underrepresents teams with weak schedules (read: mid-majors) as the calculation is not linear. If that isn't enough, a thought experiment; a bubble team would be expected to go 28-2 against a very easy schedule. Thus, even a perfect season would be just two wins away from the bubble, and a perfect record is what would be expected of the best team in the country given a very easy schedule. The expected seed calculation uses the same math as above, however, instead of comparing to the 46th best team, it finds the nth best team that would have your record given your schedule. So, if the best team would be expected to produce your record, then you get a 1.0 xSeed. Likewise, if the 10th best team would likely have your record, then you would get a 3.25 xSeed.