The Pittsburgh Pirates have been accused of aggressive base running. Never was this more apparent than in the recent series against the Cincinnati Reds. Sometimes it is the right thing to do, sometimes not. Here are two examples to illustrate the difference.
In the second game of the series, the Pirates were ahead 3-2 in the eighth inning when Andrew McCutchen was caught stealing. In the first game of the series, the Pirates were behind 2-1 in the bottom of the fourth, and Garrett Jones was cut down at home plate trying to "leg out" a double into a three-bagger from first base. One of these moves was (statistically) right, and one was wrong.
McCutchen's attempted steal was wrong. First of all, stealing is barely a break-even proposition under "baseline" circumstances (e.g., a tied game). Then, stealers have to succeed about 70 percent of the time before it contributes to a team's success. McCutchen has tried to steal seven times and has been caught twice for a success rate of 71 percent, which is just above break even.
But there was one additional factor against such a move. The Pirates were ahead. In such situations, the value of the extra base you try to get is less than what would be the case if you were tied or behind.
Since it was the bottom of the eighth, Pittsburgh's chances of winning with one man on, one out and with a one-run lead were about 88 percent. (These, and similar statistics below, are courtesy of FanGraphs). If he had swiped the second base, the Bucs' winning chances would have increased by about one percentage point. With the second out, their "win probability" dropped two percentage points to 86 percent, and then to 85 percent when the third out was recorded.
With the benefit of hindsight, we know that McCutchen would have scored (by hit or walk) in the inning, giving the Pirates a two-run lead. From the end of the eighth, the chances of the home team winning with such a lead are, well, 95 percent (compared to the 85 percent above). Ultimately, he had sacrificed over 10 percentage points of "win probability" to try to get one.
On the other hand, trying to send Jones home in the previous game was a good move. With Jones on first, and two out in the bottom of the fourth, the Pirates' chances of winning the game were 38 percent. If Jones had succeeded, the game would have been tied 2-2, and the Pirates would have been better than even money going forward (they would have had 5.1 more innings to bat versus 5.0 for the Reds).
When Jones made the out, Pittsburgh's chances of winning dropped only three percentage points to 35 percent. They had risked three points (starting from 38 percent) to try to gain more than 12. When you're behind, it usually pays to take risks.
From time to time in the past, the Pirates have tried to do "aggressive" things—baserunning, "swinging for the fences," etc.—for their own sake. It's time that they learn to guide their actions using "sabermetric" (or other quantitative) principles to maximize their chances of winning games.