I think I stopped paying much attention to pitchouts around 1987. I know they happen; if you’d asked me to guess, I’d have guessed there was one every two games, enough that I know they happen but don't really notice them. I know they work sometimes; if you’d asked me to guess, I’d have guessed the pitchout was timed correctly about half the time, and if you’d asked me to guess, I’d have guessed that in such cases the baserunner was out around three times out of four. If my guesses were correct, it would make the pitchout a tremendously valuable strategy, but one that, for obvious reasons, could be deployed only occasionally.
As it turns out, my guesses are way off. Pitchouts are rarer than I thought, correctly executed less often than I thought, and successful at gunning down the baserunner less often than I thought. They are, it turns out, sort of dumb.
Like any strategy, different managers have different philosophies about it. Santangelo was correct about the Nationals’ pitchouts. That one was Washington’s fourth and final pitchout of the year, meaning Santangelo could count the times on one hand, even somebody else’s hand. Two of those four were attempts to snuff out a squeeze play, and for our purposes today we’re going to ignore those types of pitchouts, or at least the ones we can identify easily*; whenever you see “pitchout” henceforth, it is limited to pitchouts that were probably not squeeze-thwarters. So Davey Johnson called just two pitchouts in the 2012 season, the fewest (probably) in baseball. Meanwhile, another manager called a whopping 34 pitchouts. And so which of those managers, the two or the 34, was better?
Let’s talk about pitchouts!
There were 415 pitchouts in major-league baseball in 2012, roughly one every six games (or one in 12 for each individual manager.)
Davey Johnson’s opposite is, unsurprisingly, Mike Scioscia. My suspicion is that the more likely a manager is to play smallball, the more likely he is to suspect his opponent will play smallball. As a former catcher, Scioscia probably feels good about his ability to anticipate the running game.
Actually, let’s check this. Are former catchers more likely to call pitchouts? There are 10 former catchers who managed full time this season, and their tendencies weren’t consistent. Joe Maddon called three pitchouts, compared to Scioscia’s 34. On average, each former catcher called 15.2 pitchouts, while the rest of the league’s managers called 13.2 on average. In Houston, former catcher Tony DeFrancesco took over as interim manager on Aug. 19. The Astros had thrown eight pitchouts in nearly five months to that point, but DeFrancesco called for 14 in the final six weeks. So, yes. Catchers are more likely to call pitchouts. Mike Scioscia is more likely to call pitchouts. Tony DeFrancesco calls pitchouts during pitchers’ warm-up throws.
One of the conclusions of this piece is ultimately going to be that pitchouts aren’t all that effective and usually cost the defense runs. But that doesn’t mean I’m just going to mock Scioscia’s pitchout tendencies. Scioscia’s actually better than most, perhaps all, managers when it comes to guessing correctly. Fifteen times Scioscia called for a pitchout when the runner was actually going; that’s twice as many as the next-best manager, and 18 percent of all the correct guesses across baseball. If you’re doing the math in your head, you’ve just realized that managers are really terrible at deciding when to pitchout.
Of the 415 pitchouts we identified, just 85 came when the runner was actually going. That’s a little better than 20 percent, which means that managers had their pitchers throw five balls for every successful pitchout, where “successful” means that the runner was going, not necessarily thrown out. Five balls for one success is a problem that we’ll talk about in a bit.
Scioscia guessed more than any other manager, but he also guessed better than almost any other manager.
|Chicago White Sox||15%||13|
|New York Mets||14%||7|
|New York Yankees||10%||10|
Surprisingly, managers who pick their spots more aren’t more accurate; there’s actually a pretty good correlation between frequency of pitchouts and success rate of pitchouts.
The former catchers, incidentally, are now clearly ahead of their non-catching peers, with a 26 percent success rate to the rest of the league’s 17 percent. The outsized role of Scioscia undoubtedly skews these numbers, and non-Scioscia catchers guess right 20 percent of the time. But Scioscia was a catcher, so why would we exclude him?
Catchers' Success Rate.
Alas, guessing correctly doesn’t guarantee an out. Of the 85 correct guesses, 84 were with runners going to second base. Half of those baserunners were safe anyway, as was the lone basestealer attempting to take third against a pitchout. So now we know that managers are ordering five balls thrown for a 50-50 chance of catching a baserunner who, if leaguewide rates are a reliable indicator, would be thrown out more than a quarter of the time anyway.
There are three quick math problems here, all to determine how much value this season's 415 pitchouts added. Feel free to skip to the end of the section if you’d like to avoid the worst of it.
1. I’m going to follow Tangotiger’s math and say that a stolen base is worth .19 runs, and that a caught stealing costs the offense .44 runs. Excluding attempts that coincided with a pitchout, the league was successful in 74.47 percent of its attempts in 2012 (and unsuccessful in 25.52 percent). So from the defensive team’s perspective:
(.2552*.44)-(.7447*.19) = run value of a stolen base attempt, or -.029 runs.
So, the moment a baserunner (presumably acting rationally) takes off for second, the defensive team has lost .029 runs. In the 85 instances where the manager called a pitchout, the defensive team had lost about 2.5 runs before the pitch was thrown. Independent of the manager’s decision, then, this is our expected outcome: -2.5 runs for the defense.
2. Throwing the pitchout, though, made it more likely that the runner would be thrown out. The 42 cases in which the runner was caught saved the defense 18.5 runs. The 43 cases in which the runner was safe cost the defense 8.2 runs. The defense comes out 10.3 runs ahead, instead of an expected negative 2.5 runs before. Factoring in the baserunner alone, then, the pitchouts were a 13-run swing in favor of the defense.
3. Dan Turkenkopf has calculated the value of changing a ball to a strike (and vice versa), including in individual counts. In the counts in which pitchouts were thrown in 2012, the value of each pitch was about .118 runs. In 2012, 63.7 percent of pitches were strikes. Had no pitchouts been thrown, then, those 415 pitches would have likely produced 264 strikes. Switching those 264 strikes to balls cost managers 31 runs. (The other 151 pitches would have been balls anyway, so there was no loss to the defense.)
So factoring in the batter alone, there was a 31-run swing in favor of the offense.
In total, then, managers called 415 pitchouts in 2012 and actually gave the opposing offense 18 extra runs. The average pitchout costs a team .04 runs.
The Managers’ Value.
If the average pitchout is bad, then the simplest thing to say would be that managers who call more of them are producing more bad. But, of course, a pitchout at the right moment is a net gain. The 84 pitchouts that came when a runner actually was stealing second base produced a roughly 6.5-run swing in favor of the defense. So if a manager could just guess correctly every time, he’d be wise to pitchout.
Here is how each manager did, in actual results.
Those numbers are misleading, because they reflect not only the managers’ decisions but the catchers’ throws. If a manager correctly called eight pitchouts but the catcher doesn’t throw any of those runners out, it’s not really the manager’s fault, is it? (Maybe somewhat, for not factoring in how likely the catcher was to make a strong throw, but on average at least.) Ned Yost, for instance, benefits from his catchers throwing out five of seven baserunners on pitchouts; Ozzie Guillen suffers for his catchers' 0-for-3. So let’s do this again, but assign a league-average 50-percent gundown rate for each pitchout correctly called. This will tell us how many runs we expect a manager to save (or give away) based on his rate of calling pitchouts, and accuracy at such.
As you can see, the stakes are incredibly low, about a run and a half separating the best and the worst. That’s about $750,000 or so on the free agent market, which isn’t nothing—it’s about as much as the lowest-paid managers’ entire salaries—but falls into rounding error territory. As always, one or two runs can matter a great deal if they’re the right one or two runs, or nothing at all in the context of a long season.
But we see two philosophies at the extreme in Washington and Anaheim. Mike Scioscia called 17 times as many pitchouts as Davey Johnson did in 2012. Scioscia showed an impressive ability to sense a basestealing attempt, and Johnson showed good discipline, yet both managers ultimately failed to add to their teams’ chances. That’s true of every manager in between, except for the Phillies’ Charlie Manuel, whose 50 percent guess rate made his calls slightly profitable. There's context we didn't look at that might sway these numbers a bit more in one direction—the leverage of the specific stolen bases; the stolen-base success rates of the specific baserunners involved, or the specific catchers involved; the stolen-base attempts that a frequent pitchout-caller might discourage simply by threat of pitchout; etc. But based on everything we just looked at, we have reason to assume that this is relatively pointless managerial theater. Managers, it seems, just aren’t very good at sniffing out a stolen base attempt. And when they think they know that the other team is going to run, they should maybe just ignore that feeling.
When Davey Johnson called for his second pitchout of the season, the one that caused F.P. Santangelo to comment on the specialness of it, the baserunner wasn't attempting to steal. He went, instead, on the next pitch. There was no pitchout. He was thrown out.
*Pitchouts were identified using PITCHf/x and game-event logs. They were excluded from our tally if they came with the bases loaded; runners on second and third; or on third. Pitchouts with runners on first and third were included, which means a small number of squeeze-related pitchouts were likely inadvertently included.