The Setup: Last night, the Reds and Pirates played a thrilling baseball game. Pittsburgh kept going up by one, and Cincinnati kept coming back to tie it. In the bottom of the ninth, with the Pirates up 4-3, Ryan Ludwick hit a solo homer (his second of the game) off Pirates closer Joel Hanrahan to send the game into extra innings. Then in the top of the tenth, Aroldis Chapman did something he hadn’t done all year: give up an earned run. He was done in by the unlikely duo of Clint Barmes and Michael McKenry, who led off the inning with back-to-back doubles. Pirates lead 5-4.
The Situation: Joey Votto led off the bottom of the tenth with a double. But the Reds couldn’t bring him home. Brandon Phillips grounded out to short. The Pirates intentionally walked Jay Bruce. Todd Frazier fouled out to the catcher. And the game ended when Ludwick struck out looking.
The Second Guess: After the game, talk radio hosts slammed Dusty Baker for not having Phillips bunt Votto from second to third with nobody out. You have to play for one run there, they said. Extend the game. Play the percentages. The callers, some of whom sounded vaguely drunk, agreed. One even made the specious claim that Dusty is too old-school to play small ball. (As you know, small ball is old-school.)
The Analysis: My guess is that when those radio personalities said “play the percentages,” they didn’t really have any idea what the percentages actually say. So let’s look at the facts.
According to these really cool run expectancy charts (we’ll use the 1992–2010 data), a team with a runner on second and nobody out can expect to score an average of 1.170 runs. A team with a runner on third and one out will score an average of 0.989 runs. So the number of runs that the Reds would be likely to score goes down if you have Phillips bunt Votto over to third (assuming his bunt attempt would have been successful).
But that’s only part of the equation. Yes, the bunt decreases the number of runs you are likely to score, but does it increase the chance of scoring at least one run? If Baker wanted to simply play for the tie, scratch one across and head to the eleventh, in that case does the bunt make sense, as the radio talent suggested? With a runner on second and nobody out, a team has a 63.7 percent chance of scoring one run. With a runner on third and one out, that goes up to 67.4 percent. So yes, the bunt does increase your chances of scoring, but only by the slimmest of margins, a few percentage points.
Is it really worth giving up one of your three outs for that tiny increase? Well, you have to compare it to the likely outcomes of letting Phillips swing away. That is where it gets a little more tricky. If Phillips hits a home run, you win. If he hits a double or a triple, the runs scores. You’ve tied the game and still have a rally going. If he hits a single, even if the run doesn’t score, you’re in a much better position. If he makes an out and the runner advances anyway, it’s the same as the bunt. But if he makes an out and the runner stays at second (which is what really happened), your chances of scoring go down to 41.8 percent.
The math required to combine all of those possible outcomes of Phillips swinging away is above my pay grade. Certainly, it’s a gamble. I don’t like bunting, but the numbers show that in this situation, if you decide to go for the tie instead of the win, it’s not a bad move.
Personally, I’d rather see Phillips swing. He’s your cleanup hitter. (Whether he should be is a debate for another day.) You’re paying him to hit in these situations. And as I’ve said before, outs are the most precious commodity in baseball. I don’t like to give them away for free for any reason. The Reds brought three hitters to the plate with a chance to drive Votto in from second. They all failed. But the chances of at least one of them getting a hit are pretty good. You certainly don’t need to bunt in that situation.
But I’m not saying I’m absolutely right. Intelligent minds can disagree. My complaint is simply that this isn’t how most pundits frame the conversation. Instead of looking at the evidence and then discussing it, they jump straight to the conclusion. But while just saying whatever pops into your head is easier than seeking out the truth, passing off conjecture and fact does a disservice to the people who rely on your for information.