Wherein I estimate the likelihood of each team winning the tournament, consider how this has been changed by the new tournament format, and investigate the case of apparently mis-seeded teams.
On Friday we used Log5 to estimate the likelihood of each team winning the MAC women's tournament, and also to compare the new tournament format to the old format. This morning we'll take a look at the men's tournament, but with an added twist.
In addition to using actual winning percentages, we'll also take a look at Pythagorean win percentages -- another Bill James invention -- which use scoring margins (or some derivative, such as tempo-free scoring margins) over the season to create "expected" win percentages. In this case, I've used the Pythagorean win percentages calculated by Ken Pomeroy, using 10.25 as the exponent.
Key: P=Pythagorean win percentage; A=Actual win percentage; O=Old tournament format; N=New tournament format
Several of the same effects we saw in the women's tournament are also visible here. The new format obviously boosts the chances of the top two seeds, at the expense of the other ten teams. Even the #3 and #4 seeds have lower chances, since if they win their quarterfinal games, the new format guarantees that they will face the #1 and #2 seeds in the semifinals, while under the old format they might also have faced a team seeded #7 to #10 in the semifinal round. We also see that, no matter how you slice it, Central Michigan and Northern Illinois are not going to win, and all the teams seeded #5 or worse are long shots.
One difference here from what we saw in the women's tournament is the parity among the top three seeds. It's also worth noting that, whether you take the Pythagorean or the actual winning percentages, #2 Buffalo seems to be the third-most-likely champion. The reason is obvious if you take a moment to think about it. Either winning percentage is based on their performance over the entire season, and Buffalo started rough, but then came on strong once MAC play started. Another way to look at this is that teams that seem to be seeded too high, like Buffalo and EMU, exceeded expectations in MAC play relative to their non-conference games.