I'm probably going to reveal my math ineptitude here, but oh well... Doesn't the mean have to be .500? If each game always results in a win and a loss, and there are X number of games, then there will always be just as many wins as losses each season leaguewide. Now---in each division it's not necessarily the case of course. No. The mean is the average. The median is the middle and the mode is the number that comes up at the highest frequency. When people talk about regression toward the mean it means that the performance is likely to revert to an historical average. For example if St. Louis regress toward the mean of the last 5 years, this year they will likely win 96 games (100,105,85,97,93/5). Not so good for us Cub fans. Ouch. Let's find a string of seasons where St. Louis was terrible and call that their historical average instead! :D