By Keith Glab 1/30/04
How a bullpen is used rates nearly as important as how good the bullpen is, as far as expected wins and losses
are concerned.
Usually, there are three members of the bullpen who pitch in important situations: the closer, and the right and
left handed setup men. Then there are 3-4 "other" members who are normally used in long relief or blowout
situations.
Basically, teams who have a "front three" significantly better than their "back three" should do
better in close games. This might explain why the 2002 Cubs managed a whopping -7 games against the
Pythagorean estimation for winning percentage. Joe Borowski, far and away their best relief pitcher that year,
was only used in blowouts. So the positive impact he had on the team's Runs Allowed isn't reflected in the
standings.
This might also partially explain why there is very little correlation between a team's Pythagorean differential
from year to year; a smart team will recognize the quality reliever and move him into a more prominent role, just
as the Cubs did with Borowski in 2003 (producing a +2 differential).
Unfortunately, there's no way that I can think of to prove this scientifically, because you really have
to follow a team closely to determine how their bullpen was used, exactly. We must rely on anecdotal evidence
here.
Agree or disagree with Keith's theories? Let him know: Keith@BaseballEvolution.com