Yesterday, Adam Jones walked to get on to first base. Michael Bourn, the Orioles September wRC+ king, came up to bat. He put down a sacrifice bunt. It was a miserable bunt, clean and right back to the pitcher. However, Adam Jones had a decent jump and the pitcher was not aggressive with the bunt, so it was "successful". Bourn was out at first and Jones was now at second base with one out. He would remain there for the entire inning.
The score was 3-1 and perhaps the small ball draw of getting one more run across was quite tempting to the Orioles. We can get a rough look at whether that was the right call to make. We have something called a run expectancy table. This table informs you how many runs you should expect to score given a certain baserunner and out scenario. The caveat is that this table is for league average scenarios. In the situation above with a fast-ish runner on first and the heart of the lineup coming up, run expectancy might be higher, but we will go with league average for the moment.
Man on 1st, no outs: 0.859 runs expected in inning (42% chance)
Man on 2nd, one out: 0.664 runs expected in inning (40% chance)
Whether we are looking to total expected runs or the chance of one run crossing the plate, the decision to sacrifice bunt range from being a poor to inconsequential move. Below is the entire list of sacrifice bunts.
| Date | Batter | Score | InnSc | xR dif | RoB | Out | |
|---|---|---|---|---|---|---|---|
| 2016-04-30 | Chris Davis | tied 0-0 | 1 | Out | -0.22 | 1-- | 1 |
| 2016-05-04 | Joey Rickard | tied 0-0 | 0 | 1B | 0.62 | 1-- | 0 |
| 2016-05-10 | Joey Rickard | down 1-0 | -1 | Out/SH | -0.18 | -2- | 0 |
| 2016-05-18 | Joey Rickard | ahead 2-3 | 2 | RoE | 0.62 | 1-- | 0 |
| 2016-06-06 | Paul Janish | tied 0-0 | 0 | 1B | 0.62 | 1-- | 0 |
| 2016-06-06 | Adam Jones | tied 0-0 | 0 | Out/SH | -0.10 | 12- | 0 |
| 2016-06-26 | Francisco Pena | ahead 1-4 | 4 | Out/SH | -0.20 | -2- | 0 |
| 2016-06-26 | Adam Jones | ahead 1-4 | 4 | 1B | 0.59 | --3 | 1 |
| 2016-07-04 | Yovani Gallardo | ahead 2-5 | 3 | Out/SH | -0.19 | 1-- | 0 |
| 2016-07-05 | Chris Tillman | ahead 1-4 | 3 | Out/SH | -0.19 | 1-- | 1 |
| 2016-07-08 | Hyun Soo Kim | down 6-1 | -5 | Out | -0.31 | 1-- | 1 |
| 2016-07-09 | Joey Rickard | down 2-1 | 0 | Out/SH | -0.10 | 12- | 0 |
| 2016-07-10 | Joey Rickard | ahead 1-2 | 1 | Out/SH | -0.20 | -2- | 0 |
| 2016-07-21 | Julio Borbon | ahead 1-4 | 3 | Out/SH | -0.22 | 1-- | 0 |
| 2016-07-22 | Julio Borbon | ahead 0-3 | 4 | Out/SH | -0.10 | 12- | 0 |
| 2016-07-24 | Ryan Flaherty | tied 3-3 | 2 | Out/SH | -0.22 | 1-- | 0 |
| 2016-08-14 | Wade Miley | down 3-0 | -2 | Out/SH | -0.20 | 1-- | 1 |
| 2016-09-07 | Caleb Joseph | down 7-6 | -1 | Out/SH | -0.20 | 1-- | 0 |
| 2016-09-11 | Michael Bourn | ahead 1-3 | 2 | Out/SH | -0.20 | 1-- | 0 |
From the table above, the most successful sacrifice bunt attempts according to base state run expectancy were failed sacrifice bunts. Three of the six failed attempts were singles and one was a reached on error. The others resulted in a lower expectation of runs. Some of those events are certainly arguable. Is Chris Tillman bunting really worse than him trying to swing away? Ditto for Yovani Gallardo, regardless of his past hitting performances? Same for Wade Miley. Paul Janish, Julio Borbon, and maybe Caleb Joseph fit into those arguable calls as well.
If we consider all sacrifice bunt situations, the Orioles scored eight times in their 19 attempts (42%). If we consider the percent likelihood of scoring with run expectancy tables, we get 8.76 times out of those 19 attempts (46%). That looks pretty even and gives some credence that, well, maybe the right calls are being made in these situations. If we negate the pitchers, then the numbers are seven out of 16 attempts (44%). The expected scoring times would be 7.8 out of 16 (49%). If we ignore all of the poor hitters (i.e., pitchers, Paul Janish, Francisco Pena, Julio Borbon, Ryan Flaherty, and Celeb Joseph), we have 10 attempts. The Orioles wound up scoring after a sacrifice bunt in four of those ten occasions (40%). Expectation would be for the team to score in 4.9 of those ten scenarios (49%).
All in all, it does not appear to hurt the club much to sacrifice bunt in the small sample sizes that we have for 2016. It also does not appear to help either. As is, they probably should just do what they are doing without any additional front office intervention. For me to consider there being a problem would be if we were observing a situation where the club was making the wrong decision more than twenty percent of the time. You might look at that and think that I am valuing a run too much in these situations, but typically these situations are close scores in later (not always) innings. I think you increase the value here, but maybe only twice as valuable.
In order to do this comparison, you would have to make one adjustment to the league average percent chance of scoring for:
ReplyDeleteMan on 1st, no outs: 0.859 runs expected in inning (42% chance)
because it would include those many times that the next play would be team bunted with a man on first and no outs.
(This would mean they are overlapped, that is, the second group is a subset of the first and this is not what you want when comparing these situations to see specifically if the bunt helped)
Then after removing this information you would be able to compare the two percentages.
Looking at it, likely the percentage was lowered, maybe 45% or higher, down to 40% for the situation with man on 2nd and one out.