It is the pinnacle of achievement for a pitcher. To pitch the whole game without the number under the “H” on the scoreboard ever changing is the dream of everyone who takes the mound. The perfect game may be even better, but it is so rare (the last few seasons excepted) that it rubs little of the gloss off of the no-hitter.
But what might strip that gloss away is how the pitcher fares in his next start.
No-hitters have not been kind to pitchers lately. Johan Santana, in a comeback season from injuries that cost him all of 2011, pitched the New York Mets’ first no-hitter in franchise history on June 1, 2012. He gave up six runs in five innings in his next start, and within two months was seriously ailing again, a situation many blamed on the 134 pitches he threw to complete his no-hitter.
Philip Humber did little better. He followed up his 2012 perfect game with a nine-run horror, and by season’s end was out of the starting rotation. Francisco Liriano‘s 2011 no-no had a three-inning, four-run stinker as an encore. The D-Backs gave Edwin Jackson extra rest after his 2010 no-hitter, in consideration of his 149 pitches, but he still struggled to four runs over five frames in his next game.
It’s not that there’s a curse that comes with no-hitters—or Nolan Ryan would have been struck by lightning three or four times by now—but could there be a negative effect on the pitcher’s next start? Could the all-out effort it takes to hold an opponent hitless burn a pitcher out, leave him weakened the next time he goes to the hill?
Recent anecdotes suggest it could be so. Conventional wisdom about the fatiguing effects of long starts and high pitch counts raises its own warning flag. But if I were satisfied with anecdotes and conventional wisdom, I wouldn’t be writing here, and you wouldn’t be reading here if you were satisfied with them yourself. So what do the records really say?
In the shadow of glory
My study period covers the years 1950 through 2012. This encompasses 158 no-hitters, but I am using only 135 of them for various reasons I will sketch out.
Nine no-hitters involved multiple pitchers, so there’s no one player to study. Two came in the postseason (Don Larsen and Roy Halladay), circumstances that really don’t equate to the regular season. Seven came in a pitcher’s last appearance in a regular season, so looking at their next game in six months time is meaningless. (Or if they went to the playoffs, they were facing a guaranteed good team rather than a random selection, which skews the numbers.) And five times, a pitcher followed up his no-hitter with a relief appearance rather than a start.
(None of the relievers followed his no-hitter with a save. However, Clay Buchholz of the Boston Red Sox did earn a relief win with three scoreless innings against Baltimore in the game after his Sept. 1, 2007 no-hitter.)
For the remaining 135, I collected basic data for the pitchers’ full seasons, the no-hitters, and the games after their no-hitters. My objective was to compare the cumulative performances in post-no-hit games with the cumulative performances for the seasons, minus the no-hitters themselves. No-hitters being such extraordinary events, it’s not really fair to compare a normal game to them, even with their excellence heavily diluted by all the other games in the year. (Johnny Vander Meer might quibble, but his double no-no came too early for this survey.)
For pitchers with two no-hitters in a season—in this case, Virgil Trucks and Nolan Ryan—both their following games get counted, and compared against the remainder of the season which, in these cases, does contain the other no-hitter the man pitched. Allie Reynolds also threw two no-hitters in 1951, but his second came in his last game before the World Series, so it did not make the cut. The game after his first no-hitter is compared to the remaining season that includes the second.
Here are the numbers in question, with innings pitched thrown in so you can see how big the sample sizes are.
IP R/9 ER/9 H/9 BB/9 K/9 No-hitter season 26623.2 3.889 3.489 8.204 3.077 6.526 Game after no-hitter 925 3.853 3.444 7.881 2.724 6.645
The tale the numbers tell is pretty clear. Not only is there no letdown effect, but the game after a no-hitter will, on average, be a little better in all the primary categories than the average for all the other games the pitcher throws that season.
One intriguing pattern does emerge. The hit rate for the afterglow games is about four percent below average, and walk rate is down by a good tenth. Yet the rates for runs and earned runs dip only by about one percent each. Extra-base hits may explain the discrepancy. At least, it’s a sounder hypothesis than the pitchers having forgotten during their no-hitters how to avoid bunching their hits.
For an alternative view of the numbers, I compared the rate stats directly by individual pitcher, to see whether he did better or worse than average in his post-no-hitter effort. The results fall mostly in line with the mean numbers above. Out of 135 games counted:
R/9 ER/9 H/9 BB/9 K/9 Better 71 70 73 78 62 Worse 64 65 62 57 72 (Same: 1)
The runs numbers are a little better; the hits and walks numbers are significantly better. As for strikeouts flipping the expectations, I have a theory. Normal strikeout numbers are far closer to the theoretical minimum (zero) than the theoretical maximum (27, not counting dropped third strikes). There’s more room for the better results to go far beyond the norm than the worse ones, thus pulling the mean back past the mark for the rest of the season.
Or maybe it’s just some odd random fluctuation. Those happen, too.
The equal game under K/9 merits a quick look. Sandy Koufax pitched his second no-hitter on May 11, 1963 against the Giants. For all his other games that season, he pitched 306 innings and notched 306 strikeouts, a perfect 9 K/IP ratio. So obviously, he threw the same number of strikeouts as innings the game after his no-hitter.
What isn’t so obvious is the number: 12. Koufax followed up his bit of history with a long, long effort against Philadelphia, the longest in the survey. His Dodgers finally pushed one across in the 12th to win 3-2. The two runs he gave up in those 12 innings raised his ERA for that young season, from 1.15 to 1.22.
For some added irony, he punched out a mere four batters in his no-hitter, while walking two. The FIP formula tells us that, with no walks in his subsequent 12-K outing, Koufax’s no-hitter was inferior to the game that followed, by nearly two runs. If you think this shows how meaningless no-hitters are, I cannot stop you. If instead you think this shows how meaningless FIP is, I can’t stop you there, either.
The Koufax matter does highlight a potential flaw in the numbers: they are rate statistics, not counting the raw amount of pitching done in the games after the no-hitters. If the pitchers lack the stamina to go deep into their games, they are costing their teams something after all.
I checked this idea, using a further subset of the no-hitter pitchers. I took every pitcher in the survey who pitched only as a starter, excluding anyone who had one or more relief appearances, thus making it a simple matter to calculate how many innings they pitched per start. I then made the same comparison as before between the post-no-hitter game and the modified remainder of the season.
Starts IP IP/G No-hitter season 2032 13644.1 6.715 Game after no-hitter 67 449 6.701
The pitchers do miss their average length, by a margin that adds up to less than one inning spread over all of their post-no-hitter starts. By any reasonable standard, it’s dead even. Pitchers show no more fatigue after a no-hitter than in any other random start.
Out of the 135 no-hitters fully included in the survey, 11 were what I would call “disaster” starts, lasting three innings or fewer. On the other hand, 29 of the 135 lasted a full nine innings, and four others went beyond nine. Of those, only Koufax won his. Dean Chance lost the 10-inning effort after his 1967 no-no; Nolan Ryan lost after 10.1 innings following his second no-hitter of 1973; Dave Righetti got no decision for 10.1 frames after his Fourth of July 1983 no-hitter, in a game his Yankees lost in 12.
It is interesting to look at some of the peripherals for no-hitters, and compare them to regular performance. Out of 140 no-hitters—I’m throwing back the five games followed up with a relief appearance here—pitchers averaged 2.676 walks per nine innings, and 7.687 strikeouts per nine. (It’s not exactly per game, because Jim Maloney had a 10-inning no-hitter in 1965.) Look back up-page, and you’ll see that for the rest of the season, the no-hitter pitchers averaged 3.077 BB/9 and 6.526 K/9.
The no-hitter numbers are better, but not dominantly so. Strikeout figures have the bigger edge, not surprisingly: the fewer balls in play, the fewer chances for one to sneak through. Walks are virtually a null event in a potential no-hitter, though run up enough and you could get pulled before you complete it, either for pitch count or because a run or two got across. This doesn’t happen often, unless your manager is Preston Gomez.
Look at the post-no-hitter stats, and the margins shrink even more, especially for walks. It’s 2.676 per nine for the no-no, and 2.724 for the game after, not quite a two percent difference. No-hitters still have the definite edge on strikeouts, 7.687 versus 6.645, but it’s still not overwhelming.
Seen that way, a no-hitter isn’t really all that different from a regular game. It just has no hits, plus an extra strikeout on average. Tom Tango and his FIP may have been onto something after all.
The wrap-up, or the unraveling
Another baseball myth melts under the impassive heat of the numbers. There’s no evidence for a letdown the game after a pitcher hurls a no-hitter. If anything, he marginally improves, supporting the idea of hot streaks more than of exhaustion or the instant balancing of the books.
Then again, I did not include 2013 results in my survey. Homer Bailey had a no-hitter on July 2 (incidentally, the first time I have ever seen a no-hitter from the first inning to the last: thanks again, Homer), but stumbled through 5.2 innings of 10-hit, three-walk, one-HBP, four-run ball his next start. Just 11 days later, Tim Lincecum pitched a no-no of his own, and after eight full days of rest still crashed to eight earned runs in 3.2 frames, including three home runs.
Add those stats to the 1950-2012 study set, and suddenly the R/9 and ER/9 figures for the game after end up worse than the season averages. At least, assuming Bailey and Lincecum don’t have such bad seasons that they drag the mean down enough to offset those rocky follow-ups. (Bailey ought to prevent that.)
Baseball keeps trying to persuade me that the letdown is there. Where’s Johnny Vander Meer when we need him?