October 20, 2005
Lucky Teams
The Chicago White Sox won almost seven more games than predicted by Bill James Pythagorean formula, runs^2/(runs^2 + ra^2). This makes them the 6th luckiest team to reach the World Series:
| Team | Wins | Pred. Wins | Difference | Won Series? |
| 1970 Reds | 102 | 91.4 | 10.6 | No |
| 1961 Reds | 93 | 83.4 | 9.6 | No |
| 1931 Athletics | 107 | 99.2 | 7.8 | No |
| 1930 Athleitcs | 102 | 94.8 | 7.2 | Yes |
| 1959 White Sox | 94 | 86.9 | 7.1 | No |
| 2005 White Sox | 99 | 92.2 | 6.8 | ?? |
| 1960 Yankees | 97 | 90.2 | 6.8 | No |
Not very good news for the White Sox as only one of these lucky teams managed to win the World Series. It seems lucks runs out late in the season.
BP has the Sox at 91.1. The question is what were the opponents expected wins of those teams? Even at 91.1, the Sox expected wins is greater than the Astros 90.
The two teams look pretty evenly matched on any system that uses run diff. That's not surprising because the teams' differentials were close neighbors: +96 for the Sox, +84 for the Astros.
If the Padres had somehow snuck into the Series, you'd see somewhat different evaluations based on run diff (wink).
BP puts run diff through seven million contortions, but comes out pretty close to the simple James formula, whether you use the 2.00 exponent or 1.85 (as some have suggested).