solvingbasketball.libsyn.com
Good podcast with Ken Pomeroy. A little discussion near the beginning about Wes taking the timeout technical foul at the end of the X game.
Near the end Ken talks about why plus minus isn't a useful stat. Basically there are too many variables involved, or if you eliminate variables by only using 5 man lineups, there isn't a big enough sample size to compare with other lineups.
Another interesting discussion was how preseason ratings still have predictive value even at the end of the season. Compared to baseball, college basketball has a relatively short 30 game season. The sample is small enough that noise in the season data is significant and many teams will not have reverted to their long term average performance after 30 games. Nate Silver has been using preseason rankings in his NCAA tournament projections because it works better than leaving them out.