Archive for May 17th, 2010
Editor’s Note: If you care at all about what you’re about to read (and can handle a little bit of advanced math), I strongly encourage you to read “The Book,” written by Tom Tango, Mitchel Lichtman, and Andrew Dolphin.
I was watching Saturday’s Angels-A’s game, when, in the fifth inning with a 3-2 lead, the Angels started with a Juan Rivera double, Mike Napoli single, and a Brandon Wood single. Now up 4-2 with runners on first and second, having knocked starter Tyson Ross out of the game, and facing left-handed pitcher Jerry Blevins, Erick Aybar, batting right-handed, laid down a sacrifice bunt (presumably at the discretion of Mike Scioscia). The FanGraphs play log demonstrates just how poor of a decision this was (please click the last link; I’m about to spend the next two paragraphs talking about it).
The first two columns shows the pitcher and hitter in every plate appearance throughout the game. The next four columns show the game state: inning, outs, runners, score. The play is then described in full detail. The “RE” and “WE” columns are the columns I want to focus on. “RE” is run expectancy, or additional runs the team batting can expect to score based on the current game state. The “WE” is win expectancy, or chance of winning the team chosen has based on the current situation. You’ll notice that at the beginning of every inning, each team is expected to score 0.58 runs. After each play, the game state changes and both the run and win expectancy change with it. One more important thing to note is that run expectancy does not account for runs previously scored. Therefore, if a player hits a grand slam, he’ll decrease the run expectancy (because the next hitter doesn’t get to hit with the bases loaded) but raise the win expectancy.
If you scroll down to the bottom of the fifth, you can see the part of the game I described earlier. After Wood’s single, the Angels, on average, would score about 1.61 runs and have an 85.8% chance of winning. Lucky for the Angels, 3 of the next 4 players got hits, tacked on 3 more runs, and finished the inning with a 94.3% chance of winning. The result of the bunt itself decreased the Angels chances of scoring; their run expectancy decreased to 1.50 runs and their win expectancy dropped slightly to 85.7% (the last two columns show the changes in the expectancies for each action). The bunt here only served a purpose if the next hitter hit a fly ball on which the runner at third could now tag up and score. In other words, Scioscia gave up an out in order to play for one more run in the fifth inning of a game he led 4-2. He took the situation with the fourth-highest run expectancy and decreased it. Call me crazy, but I’m going keep my outs and play for a bunch of runs, especially with the lead, and especially at home.
I don’t like the sacrifice bunt because of several reasons, chief among them is that the batting team is spending its sole resource (outs) to gain one base. There are times and situations to attempt a sacrifice, but in the fifth inning of a game the batting team is winning is not among them. You also give the pitcher and the defense something they would have had to earn otherwise. The second reason is that most of the time, it’s a boring play; the bunt goes down and the throw goes to first and all that happened is that we’re closer to the end of the game. Now, exceptions occur when the bunter has speed or lays down a particularly good bunt or the defender makes a bad play; let’s just say I didn’t show up to the ballpark to see a hitter give himself up by hitting the ball 20 feet when it doesn’t help his team win. The third reason is that it sends a bad message to the hitter; in the middle of a rally (as the Angels were), the leadoff hitter bunts? Maybe he shouldn’t be hitting leadoff.
Like I said, there are situations to bunt (you’re going to want to click that link). For instance, I 100% support National League teams having the pitcher bunt with a man on base (or for that matter, when any really poor hitter is at the plate with men on base). Also, late in games when one run can literally win you a game (like in the bottom of the 14th or the top of the 19th in this monster). It’s also important to do it occasionally to keep the defense honest; if the defense knows there is no threat of a bunt, it can play back in presumably better defensive position. Early in games, in normal run scoring environments, there’s not enough reason to take the bat out of the hands of a decent major league hitter. As this situation showed, just because the desired result was achieved doesn’t mean the decision behind it was sound.