Three of the four baseball playoff series have been decided already, and in all three cases, the losing team lost in part because a late-inning pitching specialist failed to do his job.

For the Minnesota Twins, it was Joe Nathan relinquishing a lead to the Yankees; for the Boston Red Sox it was Jonathan Papelbon collapsing against the Angels; and for the Cardinals it was Ryan Franklin (in part because of a teammate’s error) failing to hold off the Dodgers.

Nathan’s case is most interesting. As Muhammad Cohen writes from Singapore, Minnesota manager Ron Gardenhire put Nathan into the game even though at least two Yankees — Mark Texeira and Alex Rodriguez — hit him extraordinarily well. Sure enough, Rodriguez hit a game-tying, two-run homer off Nathan.

It reminds you of the story a week or so ago in which a high school football coach says he is unlikely to ever punt again. He says that the statistical probabilities argue against punting, and there is at least one study that backs him up.

So whether the question is whether to punt or to bring in your late-inning specialist, you can find yourself in a debate: Statistics are on one side and conventional wisdom on the other.

Conventional wisdom lets managers make their moves without becoming vulnerable to the second-guess. As Cohen points out, baseball has evolved to a point where managers manage by formula. It is conventional wisdom to bring in a late-inning specialist to protect a lead in the ninth inning.

But conventional wisdom can also contradict statistics, just as they did when Nathan faced the Yankees.

Unfortunately, using conventional wisdom — doing something simply because that is the way it is done — is not confined to sports. People, groups and entire countries do things that are accepted but not necessarily effective.

If you know of any cases in which evidence seems to defy conventional wisdom, I would love to hear about it.