Without knowing the answer or ever seeing if anyone has studied such a thing, I wonder if Pythag might have a tough time measuring teams that are truly elite -- teams that are at or nearing +300 or so RD. When it measures teams that are all fairly close to .500, it should work fine. But, when you get to the extremes -- a team that scores way more runs or allows way less runs -- then there is more room for variance. If all you have to do is lose 50 games to be below your Pythag, then just a few slip-ups here or there can throw you way off. So variance becomes a bigger factor. Yes this is my suspicion as well - there was a similar problem last year with the Blue Jays who had an elite run differential (they had a +220 run differential, but underplayed the pythagorean by 9 games). I think luck has a factor to play, but there is something to outliers not following the same pattern. There's also something to that buzzword everyone is itchy about, cluster luck, in that the Cubs maybe "should" have scored even more runs than they have. Looking at their 2nd and 3rd order winning percentage at baseball prospectus they're underplaying it be 9 and 7 games respectively. http://www.baseballprospectus.com/standings/index.php?dispgroup=div&standings_sort=pct1 This soothes the inner meatball inside of me that feels like we don't score enough when we have runners in scoring position and no outs. And I wont look those numbers up because its total emotional meatballery and I don't care how real or unreal it is.