`nother note about last night's game . . . .
you'll recall that the astros got a man to third base with one out in the bottom of the 9th last night. they intentionally walked biggio, a hot hitter and a veteran, to bring up the rookie willy taveras, who while not a DP threat was extremely likely to hit the ball on the ground. or strike out, which he did for the 2d out. that brought up berkman, and the white sox put him on base too to load the bases and bring up ensberg.
the move upheld the prevailing conventional wisdom among mlb managers -- ie, don't let the other team's best hitter beat you. i've always hated that type of thinking, not only because it is spineless but also because it usually runs counter to the percentages. and so it did in this case. berkman needed a base hit to end the game, all ensberg had to do was get on base -- walk hit hbp whatever. and since berkman's batting avg (.293) is roughly 100 points lower than ensberg's on-base pct (.388), the walk to berkman put the white sox in a far more perilous position. if you go by the season-long percentages, that is.
but if you narrow your data set to the postseason, then a berkman base hit -- he's hitting .340 in october -- is in fact far more likely than an ensberg hit/walk/hbp -- morgan's postseason obp is only .288. so within that narrow context, walking berkman is the percentage play. i normally think a 162-game data set is more meaningful and more reliable than a 15-game october data set, but in this particular case i would have played it as guillen did. ensberg looks lost at the plate and is an easy out; berkman on the other hand is making hard contact every time up and using the whole field. the season-long sample size means less at this point than the evidence of one's eyes. these guys are not strat-o-matic cards, they are human beings -- and ensberg is the human i would have wanted to pitch to if i were the white sox.