Baseball players have been bunting for a hundred years, but it was only recently that there has been growing criticism of the strategy. Let’s look at why.
Like on-base percentage, “never bunt” has been one of the most quoted adages to come out of the Moneyball book, era, revolution, whatever you want to call it. And like many other things from the book, it has been largely misunderstood and inaccurately used. Things are a bit more nuanced than “never bunt,” so let’s look at two stats that show what the numbers really say and when it’s OK to ignore them.
Every point of a baseball game exists in one of 24 base states. Before you run away, it’s not as complicated as it sounds. Base states are comprised of two things: The number of outs and which bases there are runners on.
When the inning starts, we have no runners on base and zero outs. That’s our first base state. The leadoff guy singles, meaning there is a runner on first base and nobody out. A second base state. You can probably see where this is going, here are all 24 base states:
From there, people smarter than me have gone back through every game* and found the average number of runs scored from each base state to the end of that inning. Our Run Expectancy Chart for every game from 1993 to 2010 looks like this:
* You can find base state charts based on different ballparks, years, and run environments. It’ll be different in Colorado than San Francisco, for example.
Note our typical bunting situation: With a runner on first base and nobody out, an average of .941 runs will score in that inning. A sac bunt moves the runner to second base with one out, meaning the team will now score .721 runs on average.
So this is one place we can see why bunting is discouraged. This is not some kind of complicated mathematical formula with arbitrary benchmarks that few can understand, it’s simply counting what has actually happened. And it’s telling us that an out is almost always more valuable than a runner moving up a base.
One thing that Run Expectancy does not take into account is the inning or score. When guys are bunting late in the game they are often not concerned with maximizing the number of runs scored; they are simply trying to score one run to tie the game or go ahead. That’s where this next chart is helpful, the chance that a single run will score in the inning based on each base state:
Again we see that while we have a 44% chance to score a run with a runner on first and no outs, we have a 41% chance to score with a runner on second base and one out. It’s not a huge difference, but it’s not helping.
One change we do see is that bunting a runner from second to third base decreases the average number of runs scored on average, but it increases the chances of scoring one run (63% to 67%), only if nobody is out though.
Our second tool is something you should incorporate into your viewing experience once or twice. This is Win Expectancy. It is similar to Run Expectancy, but it also incorporates the score, the inning, and the run environment. The result is the chance that each team has to win the game based on their current situation and how teams in the past have fared in that same situation.
This is the 2014 AL Wild Card game between the A’s and the Royals:
You can see that Kansas City had somewhere around a 3% chance to win at one point in the eighth inning, a 70% chance in the 11th, and a 10% chance in the 12th. People often criticize numbers for taking emotion out of the game, but those numbers clearly show how amazing the game was, objectively.
Fangraphs’ Win Expectancy graphs update in real-time, so pull one up while you are watching a game sometime.
So when can we bunt?
Neither RE or WE, take into account the quality of the hitter at the plate or pitcher on the mound, both are based on the averages. Kansas City bunted four times in the aforementioned game. All four had a negative impact on their Win Expectancy, but that does not tell the whole story.
Two of the bunts were from Alcides Escobar, who is one of the weakest hitters on the team. When you have a poor hitter, especially a pitcher, at the plate laying down a bunt is often the right call; as it is going to be more likely he will get a bunt down than get a hit. Another of Kansas City’s bunts came from Jarrod Dyson who is also a below-average hitter, but also has a knack for bunting for hits.
“Never bunt” is only the case if every batter and pitcher in the game is average, and even then there are a few exceptions. Real games don’t happen in Average Land, however. There are plenty of situations where the numbers are fine with bunting, it just takes some time to figure them out (and even if you make the wrong call there is always some chance you will score anyway). It may be no surprise that the A’s, Red Sox, and Astros were three of the least likely teams to sacrifice bunt in 2014, but bunts have been on the decline over the past few seasons league-wide. Like it or not though, small ball and bunting will always play some part in baseball.