Why You Shouldn't Pay Attention to Baseball Statistics in May
Just took a look at this season's combined American League statistics. I looked at every team's runs scored and runs allowed to date, and discovered something I thought was interesting. First of all, the simple calculation of Runs Scored/Runs Allowed correlates very well to winning percentage--88.5% correlation so far this year. Using the Pythagorean Theorum of Bill James, we can calculate an expected winning percentage for every team. For the season, this correlates slightly better with the actual won/lost percentage, at 89.9%.
So then I subtracted the actual wins from the expected wins via the Pythagorean Theorum. As might be expected, the best teams in the league have won a few more than their run ratio would indicate, while the worst teams have done poorer. The luckiest team in the American League are the California Angels, with 2.5 more wins than they would be expected to have, while the unluckiest team are the Royals, with 3.3 fewer wins than would be anticipated. Not by coincidence, the Angels have the best record in the AL, while the Royals have the worst.
For awhile I labored to figure out a good name for the extra wins/losses, and I figured why not call it lucky wins--wins above expected is rather dry. Then I got the brilliant idea of comparing luck to the team's won/lost percentage so far this year. The correlation was 78.7%. In other words, there is a strong and positive correlation between "lucky wins" and overall winning percentage at this point in the season.
Team R RA Exp W% W L Act W% Lucky W
Angels 170 141 58.5% 20 10 66.7% 2.5
Athletics 137 146 47.1% 14 15 48.3% 0.3
Blue Jays 137 147 46.8% 11 18 37.9% (2.6)
Devil Rays 94 136 33.7% 9 19 32.1% (0.4)
Indians 142 163 43.7% 12 16 42.9% (0.2)
Mariners 120 141 42.7% 12 17 41.4% (0.4)
Orioles 144 127 55.7% 14 12 53.8% (0.5)
Rangers 167 123 63.6% 18 11 62.1% (0.5)
Red Sox 140 112 60.1% 18 11 62.1% 0.6
Royals 135 162 41.7% 8 19 29.6% (3.3)
Tigers 167 174 48.1% 14 15 48.3% 0.0
Twins 152 143 52.8% 16 12 57.1% 1.2
White Sox 145 139 51.9% 17 11 60.7% 2.5
Yankees 140 136 51.3% 16 13 55.2% 1.1
Correl W/Act W% 68.7% -51.6% 89.9% 98.9% -98.9% 100.0% 78.3%
