Underachieving Red Sox

I was just wandering along amid the ESPN expanded standings trying to figure out if the Braves were really that good when I noticed something unusual. While most teams are a game or two within their expected win-loss records, the Boston Red Sox are a full 7 games worse than expected. They are trailing the Yankess by a hefty four games when they should be four games up on New York. I found that a little hard to believe so I thought I would check it out.

Sure enough, the Red Sox are way behind what would be expected of them. By the way, I use the Pythogorean Win Percentage with the values raised to 1.83, and the ESPN site uses to the power of two but we get similar results. Witness:

            Actual                           Pythagorean      ESPN
            W-L   PCT GB  HOME ROAD   RS  RA  W-L   PCT GB    W-L   PCT GB
NY Yankees 69-41 .627  - 33-17 36-24 650 506 67-43 .613  3   68-42 .623  4
Boston     65-45 .591  4 28-24 37-21 610 445 70-40 .640  -   72-38 .653  -

That's one game difference, but still the Red Sox should have a comfortable margin over the Yanks. Wha'appened?

I noticed that the Red Sox home record is not all that spectacular (28-24). Also, their record in one-run games is 11-16 and in extra inning games is 1-4. Then I looked at their schedule and noticed that had had a large number of games decided by large margins. There were 8 decided by 10 or more runs, seven of which the Red Sox won. Their runs for and against in those games is 104 to 31 for a Pythogorean winning percentage of .902. This translates into 7 expected wins, which is how many the Sox won. However, if you subtract the totals from these 8 games from the Boston record you get 506 runs for to 414 runs against for a Pythagorean win percentage of .591. Or a 60-42 record just two better than their actual record (minus the eight high scoring games) of 58-44. The difference, I would assume, was made up for by the issues I mentioned above and a few other high-scoring affairs (though not with double-digit differentials). Eight high-scoring games can really affect the expected win-loss totals for a team.