Last night, in the bottom of the ninth, Arizona got the runner to third with one out in a tie game. Joe Torre elected to walk the bases full and pitch to Mark Reynolds, who walked to force in the winning run.
I found a table in a Baseball Analysts article for probability of scoring at least one run for all situations of runners and outs. It is from data of 1977-1992, so it's a little old, but if it's good enough for a 2006 TBA article, it's good enough for me. Because the Dodger's were facing the heart of the D'backs order, I'll use the AL chart (no pitcher) as more representative of the situation Joe Torre faced.
Runner on 3rd, 1 out, 67.0% of the time, a run is scored.
Bases loaded, 1 out, 67.9% of the time, a run is scored.
1st & 3rd, 1 out, 65.6% of the time, a run is scored.
Maybe the best strategy is to IBB one guy and play the infield at DP depth, while hoping for a K, a pop up, or a GIDP?
Torre, as all managers do, played the infield in to cut off the run at the plate. The negatives include making it easier for a grounder to get through, making it easier for a bloop to clear the IF (Luis Gonzales to beat the Yankees in the 2001WS anyone?), and removing "standard" GIDPs from getting you out of the inning. (Thus, walking one guy and playing the IF in would seem pretty stupid.)
I've always wondered about the strategy of playing the IF in. How much does it skew the run probability? My gut feeling is that playing in hurts more than it helps the defense but it is hard to find stats based on general infield positioning, in, back, DP depth. Certainly conventional wisdom thinks it helps overall in those situations where you cannot afford to allow the run to score. Anyone ever seen an article on outcomes based on general IF positioning?
I'm guessing that playing the infield in is a much bigger factor on the outcome than whether you IBB two guys or no guys. Ultimately, Torre decided to face Reynolds and Montero (if no DP), but he could have chosen Drew, then his choice of Upton or Reynolds.