In the bottom of the 10th, Torre intentionally walked the bases loaded. There had previously been a runner on 1st and 3rd with two outs. In other words, one run will win the game for TB. There's no concern over creating a double play opportunity or anything. It's quite simple: record an out before that guy on third comes home.

He had two choices. Pitch to Huff or walk Huff and pitch to Gomes. There are 4 ways that the runner can score:

1) passed ball or wild pitch. The chances of this happening are unknown to me, but are the same regardless of whether Huff or Gomes is at bat (roughly). Thus, we can ignore it.

2) the batter hits a ball that is either misplayed (error), or a fielder's choice that doesn't record and out. An example would be a play deep in the SS hole where Jeter is forced to go to 2nd instead of making the long throw to first. Perhaps if there was a ton of speed on first, and the hitter was fast himself, then it might make sense to load the bases just on that rare chance that such a perfect ground ball is hit. But, Perez was on first and doesn't have a stolen base all season. Huff has 8 sbs in a full season, and Gomes has 4 since beign called up. They are not exceptionally fast, and this scenerio thus isn't applicable. And the chances of a fielder commiting an error are also probably the same regardless of the hitter.

3) They pitch to huff and huff gets a hit. Everyone knows how to measure the chance huff is going to get a hit right? While there are many things to consider (pitcher he's facing, past performance, etc etc), let's keep this nice and simple and look at 2 stats. His average for the season is .265. Thus, 26.5% of his "at bats" turn out to be a hit. His carreer average is better: .290. So, if you pitch to "average Huff", he'll get a hit slightly less than 30% of the time.

4) They walk Huff, and then pitch to Gomes. Now, even though Gomes is batting .279, one could argue that it's less stable than Huff's .290 b/c he's basically a rookie and thus it's a small sample space. I wouldn't argue that Huff is a better hitter than Gomes....he probably is. But, Gomes doesn't need to get a hit. he needs to either get a hit OR walk. So, looking at his OBP yields a number of .373. Over 37% of the time this season, Gomes has gotten on base either by walk or a hit.

And while Gomes' numbers are unstable b/c they are a small sample size...there's no doubt that his OBP will be higher than Huff's average. Let's throw in the fact that Proctor has control problems and thus is even more likely to walk someone.

If you were manager of the Yankees, would you rather put yourself in a situation where you lose 29% of the time, or 37% of the time?

Torre's decision is ponderous. What's more confusing is why he's allowed to continue to make these decisions.