What’s With All The Bunting?

Sacrifices by manager this season:

Dave Trembley: 5
Juan Samuel: 14
Buck Showalter: 6

On a per game basis:

Dave Trembley: 0.09
Juan Samuel: 0.27
Buck Showalter: 0.38

Buck – who was supposed to be a big inning type guy – has had his guys lay down the sac bunt more in 16 games than Dave Trembley did in his 54. It’s quite the pace of bunting; it would put the team at 61 over a full season, which is close to the most in the last decade (65 by the Tigers in ’03, 63 by three other teams – hat tip to Toirtap/Patriot for looking that up for me.)

Here’s a look at the bunts:

8/03 – Bottom of the 8th, O’s up 5-3, 0 outs,

Adam Jones on second; Felix Pie bunts him to third

Win probability went from 95.3% to 95.1%

Matt Wieters singled in Jones to make it 6-3

8/04 – Bottom of the 2nd, O’s up 2-0, 0 outs

Wieters on second, Josh Bell on first; Cesar Izturis bunts them both over

Win probability went from 80.0% to 79.8%

A Brian Roberts groundout scores a run

8/05 – Bottom of the 9th, score tied 4-4, 0 outs

Julio Lugo on second, Corey Patterson bunts him to third

Win probability went from 82.1% to 83.5%

Iztruis singles him in for the walk-off win

8/08 – Bottom of the 3rd, O’s up 1-0, 0 outs

Roberts on first; Patterson bunts him to second

Win probability went from 69.6% to 68.0%

Two straight outs leaves Roberts stranded

8/16 – Bottom of the 10th, O’s down 4-3, 0 outs

Patterson on second, Jones on first; Pie bunts them both over

Win probability went from 54.1% to 56.1%

Wieters was intentionally walked. Izturis made an out at home. Lugo walked to force in the tying run. Roberts grounded out to end the inning (at 4-4).

8/18 – Bottom of the 9th, O’s down 6-5, 0 outs

Izturis on first; Lugo (hitting for Bell) bunts him to second

Win probability went from 35.9% to 29.9%

Roberts K’s. Markakis grounds out sharply to the pitcher. O’s lose.

Not counting last night’s bunt, the overall win probability actually went up by 1.4%. Of course, this just looks at the successful bunts. There was at least one instance in which an attempted sacrifice was unsuccessful – on the 16th, Roberts tried to bunt a runner from first to second in the 9th and the lead runner was forced. On the other hand, a couple of the sacrifices were almost hits – Pie and Patterson (on one occasion), were close to beating their bunts out. Ideally you’d want to do a break-down of all the probabilities involved before each play, but that would be a lengthy exercise. Here’s a quick one of the 8/04 bunt:

Baseball Prospectus has Izzy’s double play rate at 14.3% this year (and about the same last year), while his OBP is going to be around double that, at least.

Breaking it out (very quickly, with various simplifications)

20% chance of runners on first and second, no outs, with a run in (Izzy singles the guy home) + 10% chance of bases loaded, no outs (his OBP – the 20%) + 15% chance of runner on third, two outs (double play) + (let’s say) 30% chance of runners on first and second, one out (”unproductive out”) + 15% chance of runners on first and third, one out (”productive out” moves one) + 10% chance of runners on second and third, one out (”productive out” moves both)

>, =, < (?)

90% chance of runners on second and third, one out + (let’s say) 5% chance bases loaded (might beat it out) + 5% chance runners on first and second, one out (might not get it down)

Now we have the same batters coming up in either case, so let’s use the run expectancy tables.

0.20 * (1 run in + 1.6 runs) + 0.10 * 2.4 runs + 0.15 * 0.4 runs + 0.30 * 0.9 runs + 0.15 * 1.2 runs + 0.10 * 1.4 runs= ~ 1.4 run

0.90 * 1.4 runs + 0.05 * 2.4 + 0.05 * 0.9 runs = ~ 1.4 runs

Obviously if you change things a little here and there it could move the results around (frankly, I just made most of the percentages up – you can say this is useless if you’d like). It’s pretty close to even.

So when it’s close like that, I wouldn’t take it as a big deal. Especially since the Orioles aren’t the highest scoring team, and they’ve got some less than stellar batters at the bottom of their line-up.

Some findings on the subject from The Book:

“If the opposing manager is thinking about attempting a sacrifice (with a runner on first and no outs and a non-pitcher at the plate), tell him that you will gladly give the runner second base in exchange for an out. In fact, tell him that he has that option – in advance – any time there is a runner on first and no outs!”

Wow.

“Early in the game in a low run-scoring environment, it is correct to often sacrifice bunt with a runner on first and no outs.”

This is largely because the defense is often not playing for the bunt, as they are towards the end of games. This increase the batters chances of successfully moving the runners over and getting himself on-base as well. If the defense is expecting it though – such as due to frequent use, – then that advantage goes away.

“Late in a close game, in a low runs-scoring environment, it is correct to often sacrifice bunt with a runner on first and no outs. In an average run-scoring environment, you should sometimes sacrifice to keep the defense honest.”

So the same kind of thing. The defense is looking for it more, but it’s a close game so the run is more important (so to speak).

“When deciding whether to bunt or not, anticipating or observing the position of the infielders is critical.”

If they know you’re bunting then the odds of getting on base (ie, not making an out) go down.

“If the batter has an expected wOBA of less than .300, 44 points less than the league average, he can attempt a sacrifice bunt, even if the defense is expecting it.”

That’s a hitter about as bad as Izzy, or worse.

The gist is that giving up an out to move a runner from first to second is terrible, but the chances of getting on base from a bunt attempt can sometimes make it worthwhile. It is very important to note how the defense is playing; if they know you’re bunting and adjust accordingly then that makes it a worse idea. Bunting every 2-3 games (like the O’s have been doing recently) instead of every 4-5 (like they had been doing early in the season) isn’t a mistake in and of itself. My instinct was to hate it out of hand, but that isn’t the right thing to do. It’s a very complicated issue that depends on a lot of situational factors.

The bunt in last night’s game probably was bad though, since the M’s knew the bunt was coming (and Aardsma has control issue sometimes). That 6% drop in generic win expectancy was a doozy, so I’m definitely not happy with Buck about that. I’ll try to keep all this in mind though, and maybe give him more of a break about the bunting in the future.

A lot more from MGL on bunting here. Enjoy.