Bunts are bad. They hurt your pitchers, they destroy your offense, and they drive your fanbase insane when Ryne Sandberg calls for them every time a man gets on base. But what if bunts are good? The Phillies certainly think so, at least in a different way this spring, as they work on the squeeze play. The play has joined the Phillies’ arsenal of tricks on how to compensate without a lineup of competent hitters. Small ball gets a bad name, but so does watching Freddy Galvis flail away at three pitches nowhere near the strike zone. At the core, small ball is about trying to increase the percentage of scoring a single run at the cost of scoring multiple runs. More runs lead to more chance of winning the game, but it also requires hitting talent and some luck to chain together the events in order to cause that to occur. The best possible spot here is to find a place where you greatly increase the chance of scoring a run while not severely cutting your overall offensive potential, enter the squeeze.
The squeeze play involves a runner on third base and the batter laying down a sacrifice bunt. In ideal circumstances the runner at third is not subject to a force play and there are less than two outs (so the defense doesn’t just throw the ball to first and laugh at you). The squeeze itself comes in two different flavors, the safety squeeze where the runner does not break for home until contact is made and the suicide squeeze where the runner breaks for him with the pitch. The safety squeeze makes for a closer play at home, but it lessons the chance of the runner at third making an out due to the batter’s inability to make contact. The suicide squeeze has an almost certain chance of the runner scoring if good contact is made, and he is almost certainly out if contact is not made.
Let’s play around with run expectancy to see whether this actually all makes sense to do. I am going to use Baseball Prospectus’ 2015 Run Expectancy numbers for this exercise.
I am going to narrow down my scenarios to those plausible, so here are the scenarios to run through:
- Scenario A: Runner on 3rd, no outs
- Scenario B: Runner on 3rd, 1 out
- Scenario C: Runner on 1st and 3rd, no outs
- Scenario D: Runner on 1st and 3rd, 1 out
- Scenario E: Runner on 2nd and 3rd, 1 out (there is no reason to squeeze with no outs and 2nd+3rd)
Since we have our scenarios, let’s focus on the potential outcomes:
- Outcome A: Run scores, batter is out (all other runners move up a base)
- Outcome B: Runner is out, batter is safe
- Outcome C: Runner is out, batter makes no contact
- Outcome D: Runner and batter is out (bunt popped in the air)
- Outcome E: Run scores, batter is safe (all other runners move up a base)
Here is the outcome in terms of RE24 or the change in run expectancy with 0 being that the expected outcome stayed constant.
Scenario A: Runner on 3rd, no outs (003)
| Outcome | Starting Expectancy | End State | End Expectancy | Runs Scored | RE24 |
| A | 1.3 | 000 1 out | 0.26 | 1 | -0.04 |
| B | 1.3 | 100 1 out | 0.5 | 0 | -0.8 |
| C | 1.3 | 000 1 out | 0.26 | 0 | -1.04 |
| D | 1.3 | 000 2 out | 0.1 | 0 | -1.2 |
| E | 1.3 | 100 0 out | 0.84 | 1 | 0.54 |
Scenario B: Runner on 3rd, 1 out (003)
| Outcome | Starting Expectancy | End State | End Expectancy | Runs Scored | RE24 |
| A | 0.89 | 000 2 out | 0.1 | 1 | 0.21 |
| B | 0.89 | 100 2 out | 0.22 | 0 | -0.67 |
| C | 0.89 | 000 2 out | 0.1 | 0 | -0.79 |
| D | 0.89 | 000 3 out | 0 | 0 | -0.89 |
| E | 0.89 | 100 1 out | 0.84 | 1 | 0.95 |
Scenario C: Runner on 1st and 3rd, no outs (103)
| Outcome | Starting Expectancy | End State | End Expectancy | Runs Scored | RE24 |
| A | 1.67 | 020 1 out | 0.65 | 1 | -0.02 |
| B | 1.67 | 120 1 out | 0.89 | 0 | -0.78 |
| C | 1.67 | 100 1 out | 0.5 | 0 | -1.17 |
| D | 1.67 | 100 2 out | 0.22 | 0 | -1.45 |
| E | 1.67 | 120 0 out | 1.44 | 1 | 0.77 |
Scenario D: Runner on 1st and 3rd, 1 out (103)
| Outcome | Starting Expectancy | End State | End Expectancy | Runs Scored | RE24 |
| A | 1.13 | 020 2 out | 0.32 | 1 | 0.19 |
| B | 1.13 | 120 2 out | 0.44 | 0 | -0.69 |
| C | 1.13 | 100 2 out | 0.22 | 0 | -0.91 |
| D | 1.13 | 000 3 out | 0 | 0 | -1.13 |
| E | 1.13 | 120 1 out | 0.89 | 1 | 0.76 |
Scenario E: Runner on 2nd and 3rd, 1 out (023)
| Outcome | Starting Expectancy | End State | End Expectancy | Runs Scored | RE24 |
| A | 1.28 | 003 2 out | 0.36 | 1 | 0.08 |
| B | 1.28 | 103 2 out | 0.48 | 0 | -0.8 |
| C | 1.28 | 020 2 out | 0.32 | 0 | -0.96 |
| D | 1.28 | 000 3 out | 0 | 0 | -1.28 |
| E | 1.28 | 103 1 out | 1.13 | 1 | 0.85 |
Before even continuing forward we can throw out any scenario with no outs as there is no positive change to run expectancy unless the defense makes an error. I am going to remove scenario E and focus purely on the other two 1 out scenarios. The goal is to then find the break even percentage of success to make the strategy work. To simplify I am going to remove the disaster outcomes of an error or misplay by the defense or a pop up bunt by the batter. I am then going to further separate into safety (Outcome B) and suicide (Outcome C) squeezes. The breakeven percentage would be RE24(success)*x+RE24(failure)*(1-x)=0 or breakeven=RE24(fail)/(RE24(fail)-RE24(success)).
| Scenario | Type of Play | RE24 Success | RE24 Failure | Breakeven Percentage |
|
Runner on 3rd
|
Safety | 0.21 | -0.67 | 76.1% |
| Suicide | 0.21 | -0.79 | 79.0% | |
|
Runner on 1st and 3rd
|
Safety | 0.19 | -0.69 | 78.4% |
| Suicide | 0.19 | -0.91 | 82.7% |
Those are large percentages, but they are not unreasonable. If the Phillies can exceed those levels during the major league season they could add some runs to what should be a sputtering offense. That bring this to another point, which is the run expectancy numbers are based on league wide data. The Phillies have a well below average offense which means their run expectancy will be lower than the league. As the RE24 of success is mostly determined by the value of the run scored, any changes to run expectancy will have more effect on the negative value of the RE24 of failure, which will only serve to lower the breakeven percentage. The Phillies are not going to win a bunch more games by employing the squeeze play, but they can prove that we should not laugh at all bunts.
Does the league-wide run expectancy include the sac fly? In that circumstance, a run scores but you’ve increased your outs as well, similar to the bunt scenarios.
Good stuff though.
The league wide should include that. I believe it is calculated based on how many runs scored in an inning where the plate appearance started in that state. Then the results are averaged for each of the beginning states.
If you were to approach that problem in the same way you find that hitting a sac fly with no outs is right about the expected run outcome.
Great analysis, Matt. I’ve wondered about the impact of a safety squeeze bunted to a pitcher’s follow through side vs. his opposite side. In other words, the effectiveness of the bunt itself. So far in ST, the Phils seem to be doing a pretty good job of forcing the first or third baseman to field the bunt.