In 2011, there were 2,346 sacrifice bunts attempted. The results were as follows:
- 1667 sacrifices
- 235 hits
- 72 reaches on error
- 44 doubles plays
- The rest were pop-outs, fielder's choices, or other ineffectual outcomes.
If we add the Win Probability Added for all 2,346 attempts, the result is -16.4. That's more than 16 entire games lost by virtue of the sac bunt. On a per-attempt basis, it's -0.007, a minuscule (and yet negative) number that clearly shows that, on average, the sacrifice bunt is a losing play.
Now, sometimes, a sac bunt attempt turns out well. There were 437 times when the sac bunt attempt resulted in a positive WPA. The majority of those were large gains when the result was either a hit or a reach on error, or a small gain on a small number of successful bunt-outs. However, there were 1,599 times when the attempt resulted in negative WPA. Some of these were terribly negative plays resulting in a double play, but most of them were simply the normal desired outcome of the play. (Yes, apparently the manager is trying to do something that is actually hurting his team's chances of winning.)
All of this being said, there are still some times when a sac bunt makes sense:
- WPA doesn't take into account who is batting. When the pitcher is up with a runner on first, bunting probably makes a lot of sense in most cases. Many pitchers are more likely to hit a weak ground ball double play and therefore might as well register one out instead of two.
- If the batter is a good bunter and a fast runner, the chances are increased of a base hit or a reach on error. The problem with this is that batters who are good bunters and fast are usually good hitter (often leadoff hitters) and it's better to have these guys swing away.
- With a good bunter up, it's sometimes good to attempt a sacrifice when the defense does not expect it, such as if the 3rd baseman is playing deep.
- There are a few other select cases when it might be a good idea, but then are limited.