Testing for Bias and Manipulation in the National Basketball Association Playoffs

The following paper examines potential sources of bias in National Basketball Association (NBA) playoffs. Employing NBA playoff data from 2003 through 2008, we investigate the expected margin of victory by the favored team. This method, by incorporating potentially mitigating variables, can also be used to test for potential bias and manipulation in tournament results. We find that the regular season winning percentage, used as the method for seeding teams for the NBA playoffs, is a good predictor of post season play. In addition to regular season play, we find confirming evidence of the home court advantage hypothesis suggested by previous studies. However, the analysis does indicate potential bias in playoff results in favor of large market teams. A team from a larger market is expected to win by a larger margin in the case of being the better team (stronger seed) or lose by a smaller margin in the case of being the worse team (weaker seed). Bias is also detected which would indicate closer margins of victory in early playoff series games which increases the probability of a longer series. While financial motivation for playoff manipulation is discussed, the findings only suggest bias and do not indicate its source.