February 13, 2015

Prospectus Feature

The Golden Age of Immaculate Innings

by Ian Frazer

Baseball is no stranger to misnomers. A splitter stays in one piece; a drag bunt contacts the bat no longer than any other hit; a knuckleball is usually thrown with the fingernails. This point is even true concerning some of the game’s most notable events. A perfect game is never quite so: Spots are missed, mistakes are made, pitches are often hit extremely hard. But they result in outs all the same.

A truly perfect game, theoretically, would be composed entirely of strikeouts, with no balls thrown. Nobody has done this outside the world of video games (and probably within that world, as well). But if we shrink the scope down to a single inning, baseball history offers us a number of examples of perhaps the closest thing to true perfection that baseball offers.

Per baseball-almanac.com, the game has seen 79 instances of these “immaculate innings.” They are more than three times as rare as no-hitters and about as much so as a cycle. But they lack the narrative momentum of those events, the build-up and climax. With a perfect game, we have 26 batters to see it coming. With a cycle, we have three (or more) at-bats. But with an immaculate inning, it’s a single inning, its significance clouded by that of the game as a whole.

Still: What, if any, information can we glean from examining the history and patterns of the immaculate inning? And can that help predict or guess when and where the next one’s going to come?

The frequency of immaculate innings has had a similar trajectory to the game’s recent increase in strikeouts. There were 18 in the 1990s, and while that number shrunk to 15 in the 2000s, there have been 14 in the current decade, including a whopping seven in 2014. The highest yearly total before that was five, in 1998, where both Randy Johnson (Hall of Famer) and Jesús Sánchez (5.32 career ERA, a career that in Taiwan) managed the feat. Twelve of the 79 immaculate innings came from Hall of Famers. That’s 15 percent. Of the 287 no-hitters that have been thrown, 34 have been by current Hall of Famers, just 12 percent.

But random variation and regression render that slim advantage useless. What could be much more telling, then, is to take a cross-section of a selection of immaculate innings. The MLB.tv archives date back to 2012, and there have been 11 since then, so let’s go with that.

I’ll start with a table. These are the pitchers who accomplished the feat in that set timetable and a selection of their stats from the year when they threw their immaculate inning.

K% and BB% are the best of the selected stats at giving us a basic impression of a pitcher’s “game.” In 2014, major-league pitchers recorded a strikeout in 20.4 percent of plate appearances, and a walk 7.6 percent of the time. These ratios vary no more than 0.5 percent in 2012 and 2013, so it’s not like the increase in immaculate innings in 2014 went along with a precipitous increase in strikeouts.

Looking at the ratios of our sample, a fairly inclusive picture of the immaculate inning club emerges: Yes, there are some veritable strikeout machines there, in Boxberger and Delabar, but the group generally tends to hover around the league average, even dipping below it with regularity. The same is true for walk rate: You can be at twice the league average or half of it, but the immaculate inning fairy can still sprinkle her magic dust in your glove. (Or maybe that’s sandpaper residue.) WARP is a bit more conclusive, with Buchholz, Brothers and Contreras being the only ones to record negative figures in the years where they immaculate inning'd, and the ERA figures were more often under 3.50 than not, though only by one.

So, advice for aspiring immaculate inning throwers: Be not bad, generally! But it’s not imperative. After all, you only need one inning to do it. Also: Strike out a decent amount of people! But it doesn’t have to be, like, everyone. Other than that, you probably have a chance! The immaculate inning club is just as diverse and random as that of perfect games and no-hitters.

That’s not to say that, statistically, a better pitcher isn’t more likely to accomplish this feat. Let’s use an example that mixes projection and observation: On baseballsavant.com, I searched first for pitch outcomes on 0-0 and 0-1 that would result in a strike. This included foul balls. Then, on 0-2, I eliminated those fouls from the search criteria, because we’re looking specifically for three-pitch strikeouts. I multiplied those three percentages to figure out the likelihood of a pitcher recording a three-pitch strikeout and multiplied that number by itself twice. (Three three-pitch strikeouts in a row.)

Let’s use a couple of real-world examples: For 2012 Clay Buchholz, who recorded an immaculate inning, the previous mentioned methodology gives us a .000073 percent chance of that happening, taking his performance from that year into account. Clay Buchholz was bad in 2012. Corey Kluber, in 2014, was fantastic. Still: With 2014 Corey Kluber’s outcomes, he would’ve had a .0005 percent chance of throwing an immaculate inning. Technically, 2014 Corey Kluber would be 6.85 times more likely than 2012 Clay Buchholz to throw an immaculate inning. But with odds so astonishingly small, those ratios are effectively meaningless except in enormous sample sizes.

So is it surprising that a bad (at the time) pitcher has accomplished this feat, while many very good ones haven’t? In a very broad, almost cosmic sense, yeah. But in context, it’s surprising that people have thrown immaculate innings at all! The astronomical improbability of the event even occurring means randomness wreaks complete havoc on the results. They’re noisier than a drum factory.

How about the exact procession of the innings? As is the case with a sport as endlessly varied as baseball, there are quite a few ways for an inning, even one as straightforward as nine strikes, three strikeouts, to proceed.

Oftentimes, repetition played some sort of role. Buchholz started Adam Jones with two curveballs and Matt Wieters with two fastballs, but he ended the at-bats by freezing both on two-seamers low in the zone. Brandon McCarthy fanned Wil Myers, Nick Franklin and Matt Joyce with practically identical 93 mph fastballs on the outside black, all swung through. Brad Boxberger entered in the highest-leverage situation of those surveyed — bases loaded, no outs with the Rays up 3-1 on the Orioles. He threw early-count changeups that Steve Pearce and Jonathan Schoop each laced foul to left, then ran fastballs up by the letters for swinging strikeouts. He started his final batter with a changeup and went inside-outside with two fastballs to finish off the at-bat.

Iván Nova fanned Ike Davis, Mike Baxter and Rubén Tejada on curveballs low and out of the zone. The first strike of Davis’ at-bat was a check swing mistakenly marked as a ball on the YES broadcast’s chyron, so while the inning was immaculate for Nova, it was less so for the folks on television.

But are there any patterns that apply to pitch usage for the group in general, rather than a single pitcher? The broadest way of meaningfully separating put-away pitches is fastball/not-fastball. In 2014, 17,356 of the 37,441 strikeouts recorded in all of major-league baseball came on fastballs, defined in the query as two-seamers, four-seamers and cutters. That’s 46.4 percent.

Of the 33 strikeouts our sampled pitches recorded, 18 of those came on some sort of fastballs, which is 54.5 percent. So—putting aside for a moment the small sample factor—would relying on fastballs as your putaway pitch make you a better immaculate inning candidate? Well, among individual pitchers, the relative frequency of pitches used to record strikeouts varies wildly. In 2013, the year of his immaculate inning, just 19.8 percent of Nova’s strikeouts came from fastballs, as he preferred to use his plus-plus curveball to finish hitters off. Boxberger, on the other hand, got 83.7 of his whiffs in 2014 from fastballs. And when I averaged out the yearly fastball-for-strikeout percentages of the surveyed group, it was 53.6 percent, effectively the same proportion of fastballs that finished off whiffs in the immaculate innings.

The closest thing to pitch selection guidance that these results can provide is: Be yourself! How nice. If you’ve got a frisbee slider, use it, as Justin Masterson did for all three of his strikeouts. If you’ve got a fastball whose late life makes it deadly when changing velocities, use it, like Boxberger did on all three of his whiffs. And if you have an upper-90s fastball, a positively mammoth curveball and a high-80s slider/short curve, like Garrett Richards does, well, use all three of them!

Missing one’s spot can also be strangely conducive to completing an immaculate inning. In Steve Delabar’s innings, twice did J.P. Arencibia call for a fastball low in the zone and twice did Delabar’s heater miss up but result in a swinging strike. Carlos Contreras missed high and away on a fastball that Jordy Mercer swung through and got Gregory Polanco to swing through an out-of-reach changeup and a fastball in the dirt.

And sometimes, a pitcher is just so overwhelming that, if he hits his spots and mixes well enough, any other outcome is inevitable. Richards started Jon Singleton — who was playing in just his second major-league game, Mike Trout bless him — with a 96 mph fastball low and outside. Then he threw a spike curve (at 88 mph!) that Singleton swung and missed on, then a 79 mph curve that broke from just about from Singleton’s shoulders to his ankles. Singleton’s swing actually did a decent job of following the last pitch, but it just broke about an inch too far. The at-bat serves as a demonstration of just how far the proverbial pendulum has swung in favor of pitchers: The power and control of an at-bat is in their hands, far more often than not. Hitters’ only chances, oftentimes, are mistakes.

Now, to address the questions of what patterns we can see with immaculate innings and whether we can really predict anything about them: There aren’t really any — other than the pitchers usually being better than bad — and, well, we really can’t, at least on any meaningful scale.

Nonetheless, I think immaculate innings deserve more coverage and excitement than they actually get. For me, it’s their symmetry and indisputable dominance — it is very much possible to throw a no-hitter and be pretty far from dominant (cough, Tim Lincecum, cough, 147 pitches), but each immaculate inning is a player simply in complete control of the situation.

It’s so unexpected and so serendipitous when one comes around that even announcers tend to miss the significance. The more broadly based feat of striking out the side often overshadows the manner in which it was done. Michael Kay called McCarthy’s feat a “perfect inning.” Both broadcast crews had high praise for Delabar in general, but neither recognized the exact accomplishment.

The immaculate inning, somewhat paradoxically, is an exceedingly rare feat that can be overshadowed. During Miley’s, there was no mention in the Diamondbacks booth of anything even related to baseball: sideline reporter Amanda Pflugrad was interviewing the new broadcast team for the Phoenix Suns. I mean, it was Diamondbacks-Rockies, two days away from the end of the season. Some situations are impossible to spice up.