The Sacrifice Bunt: Not the Way to Go

My reader(s) know by now that I am sabermetrically-inclined. I like to know the numbers and see the game at an objective level. I’d like to start dropping some words about things that I don’t think are very important or useful that teams, fans, and the mainstream media will often hang onto and use as talking points when they shouldn’t in fact be discussed. Earlier in Blogservations I talked about the intentional walk and its potential. While there are indeed times to use it, managers most likely use it more often than they should.

Another “tool” managers, especially those of the National League, will often employ is the sacrifice bunt. I grew up a fan of “small ball,” of moving runners over and manufacturing runs. The truth of the matter is that those days are really gone in baseball. People have begun to realize that outs are so important that wasting them for things like moving runners from first to second is almost entirely a wasted effort. Today I ran into a particularly interesting situation regarding last night’s game versus the Padres. There’s the link on FanGraphs, but here’s what I’m looking at.

As you can see, it was the seventh, we were down one run with Cody Ross and Jeremy Hermida on first and second respectively. Dan Uggla came to the plate, he of the .229 batting average despite good peripheral numbers that people continuously bash. On a FanPost by one FishFan218, he called for Fredi to get Danny bunting to move the runners over. I disagreed, citing the general idea that bunting and giving away outs isn’t useful unless the run expectancy jumps severely. He respectfully argued his point (with a bit of attitude regarding Danny’s .217 BA with runners in scoring position, to which I say small sample sizes yield funny numbers). Here was my response, copy-pasted in full with all the fun WPA math, directed in regards to this response, not posted for brevity’s sake. Of course, being the narcissist that I am, I went ahead and posted my response. Hey, it’s my blog!

—-

Lucky happens, it’s part of the game. The bottom line is that up until the very late stages of the game, when WPA jumps significantly and guaranteed wins can be picked up directly from runs scored, bunting is a sure fire way to kill a big inning. I was going to make my point before checking out that the Padres had only a 53.6% chance of winning the game when Uggla stepped to the plate. His single brought it to 38.6%, while a sac bunt would have likely brought it back up to around 55-57%. Since I don’t have the exact numbers from Tango, I used a Win Expectancy Finder created by Christopher Shea to illustrate this point. Game data is taken fron 1977 only up to 2006, so it’s not completely up to date, but bear with me. Win% in the context of the home team, WPA in context to our hitting team.

Current Game State: 12_, 0 outs (first and second, 0 outs); Win%: .566

Consider 90% success rate for Uggla bunting runners over:

Game State: 23, 1 out; Win%: .583
Weighted WPA: -.015 rounding

10% chance Uggla fails to move runners:

Game State: 12, 1 out; Win%: .631
Weighted WPA: -.007

Total WPA of move: -.022

Consider 33% chance Uggla reaches base (around his OBP this year, if I used better numbers, like his career total, it would be 34%):

Game State: 123, 0 outs; Win%: .527 (this one had the smallest sample size)
Weighted WPA: .013

67% chance he fails to move runners

Game State: 12_, 1 out; Win%: .631
Weighted WPA: -.044

Total WPA: -.037

After that undertaking, it looks bad for my argument. But that only considers instances where he simply gets on base. What if he singles, or gets a hit in general?

23% chance of single and driven run

Game State: 12_, 0 outs, tied; Win%: .356
Weighted WPA: .048
10% chance he walks, doesn’t drive in run
Game State: 123, 0 outs; Win%: .527
Weighted WPA: .004
67% chance he fails to move runners

Game State: 12_, 1 out; Win%: .631
Weighted WPA: -.044

Total WPA: .008

And this still doesn’t account for extra base hits, which I don’t want to get into. In addition, the bases loaded situation, which Uggla eventually worked himself into, probably yields a better Win% for the hitting team than .473, so I think that might have been a small sample size issue for this range of study (only 96 games recorded with that base/out state). In the actual game, it yielded a WPA of .146 to Uggla. Still, the little study should show that bunts can be the incorrect move, even this late in the ball game. When you’re down a run, I’m pretty sure the right move is almost always to swing away, because you don’t want to risk losing the big inning due by giving the pitcher a free out.

Just to make sure people didn’t miss that:

Uggla bunts (assuming 90% success rate): WPA: -.022
Uggla swings (assuming current BA and OBP, no extra base hits): WPA: .008

That’s a two percent win difference when weighted, which I think is decently significant.

—-

What do you guys think? Agree with my sentiment or feel a bunt was still in order? I thought my premises were fair and indeed could have gone further and used Uggla’s better BA/OBP.