Slate Strikes Out

Writing in Slate, Tim Marchman attempts to explain why low-strikeout pitchers who allow batters to hit the ball into play are less likely to throw no-hitters than high-strikeout pitchers. He fails miserably, which I’ll explain at length, mostly because it’s a pretty good illustration of how people who have no idea what they’re talking about are inexplicably given space by major news organizations to write absolutely nonsensical articles.

If you don’t care about the details, consider this: Marchman’s piece purports to explain why low-strikeout pitchers are less likely to through no-hitters, but he doesn’t offer a single shred of evidence that they are, in fact, less likely to do so. And, indeed, that has not been the case in recent years. That tells you pretty much all you need to know: Apparently neither Marchman nor his editors at Slate thought it mattered whether the basic assumption Marchman’s piece is based upon is true.

Ok. Now, the details.

Marchman’s first mistake is thinking Nolan Ryan, Randy Johnson and Roger Clemens all belong together in a group of pitchers who walk a lot of batters:

Great pitchers divide, broadly, into two classes. The first includes what you might call the snorting bulls—grimacing maniacs with huge fastballs and merciless attitudes: Roger Clemens, Randy Johnson, Nolan Ryan. These pitchers will typically strike out lots of batters and walk lots as well, glowering all the while and occasionally coming close to killing an opponent. For these men, every at-bat is a fight and every hit a failure, and what makes them so good is that they never give in at all. Nolan Ryan, who had years where he gave up as many walks as hits, was almost a caricature of the type, seemingly more interested in overpowering hitters than in winning.

Over the course of his career, Nolan Ryan walked 4.7 batters per 9 innings pitched. His lowest annual BB/9 was 3.3 in 1990.

Over the course of his career, Roger Clemens walked 2.9 batters per 9 innings pitched. His lowest annual BB/9 was 2.0 in 1984; he had 15 seasons with a lower BB/9 than Ryan ever had, and he never had a season as bad as Ryan’s career average.

Over the course of his career, Randy Johnson walked 3.3 batters per 9 innings pitched. That’s significantly lower than Ryan’s career average, and it’s badly skewed by Johnson’s first four seasons, in which he was not yet Randy Johnson. After those first four years, he walked 2.7 batters per 9 innings.* His career best was 1.6, and he had 14 seasons better than Ryan’s career-best.

Johnson had four seasons in which he was among his league’s top ten in lowest BB/9, Clemens had 6 such seasons, and Ryan -- obviously -- had none.

More broadly, Marchman’s suggestion that the three pitchers were comparable in greatness, or that Clemens and Johnson were, like Ryan, “more interested in overpowering hitters than in winning,” is ludicrous. Clemens is arguably the greatest pitcher who ever stepped on a mound, and Johnson not far behind. Ryan, on the other hand, was a good pitcher, and a very freakish one, but several rungs behind the others. (Again, let’s just take a quick-and-dirty measure: ERA+ compares a pitcher’s Earned Run Average to the league average, and adjusts for home ballparks. 100 is average. Roger Clemens’ career ERA+ is 143. Nolan Ryan had only one individual season as high as 143, and a career total of only 112.)

So anyway: Roger Clemens and Nolan Ryan: Very different. Tim Marchman, a professional sports writer, should know this.

Marchman moves on to what he thinks is the other type of great pitcher:

The other kind of great pitcher takes a more modest approach. For him the point is not to master a given hitter but to master the game—think of Greg Maddux, Mike Mussina or Robin Roberts. These kinds of pitchers frequently lead the league in fewest walks allowed per game, throwing strike after strike, daring the batter to put the ball in play and let the fielders do their work. They accept that outcomes in baseball are random or nearly so, trust percentages and concentrate on what they can control.

Oh, boy.

Outcomes in baseball are not random or nearly so. If they were, lists of league leaders in home runs and walks and strikeouts would not be littered with the same names every year, and the Pirates would not stink every year. What Marchman means -- assuming he has any idea what he’s talking about -- is that the outcome of a ball hit into play is random, or nearly random. That’s a bit of an overstatement, but what it basically means is that a hitter’s (or pitcher’s) home run, walk, and strikeout rates are relatively consistent, but the rate at which ground balls, line drives, and (non-home run) fly balls become hits or outs fluctuate more wildly.

In other words: the things pitchers (and hitters) “can control” are strikeouts, walks, and home runs. That’s pretty much the opposite of Marchman’s implication that pitchers like Randy Johnson and Roger Clemens don’t focus on what they can control and don’t try to “master the game.” Of course they do. In his desire to draw value judgments about supposedly-cerebral pitchers who induce contact and dumb brutes who care more about strikeouts than helping their teams, Marchman makes the nonsensical argument that pitchers who rely on random outcomes are the ones seeking to “master the game” by focusing on what they can control. Weird.

Marchman then puts Halladay in that second group of great pitchers, the ones who encourage hitters to “put the ball in play” and who “let the fielders do their work.” This is a good place to point out that Roy Halladay has struck out at least 200 batters for three straight seasons, has been among the league leaders in strikeouts per 9 innings pitched three times, and strikes hitters out at a better-than-average rate. But, Marchman argues:

Halladay has an arm like Ryan's and is capable of throwing like him—never relenting or giving batters anything to touch. Instead, he almost always gives a batter something to hit. In fact he gives up a lot of hits. He's twice led the league in hits allowed, and among pitchers with at least 2,000 innings and a .650 winning percentage, only Lefty Grove gave up more per inning.

Marchman doesn’t mention this, but the two seasons in which Halladay led his league in hits allowed were seasons in which he led his league in innings pitched, which is not a coincidence. Halladay has three times been among the top ten in his league in fewest hits allowed per 9 innings pitched. He does not give up a lot of hits.

(Oh, and that bit at the end is just stupid. The .650 winning percentage is an irrelevant screen that narrows the total number of pitchers in question dramatically and unnecessarily. You know how many post-WWII pitchers have thrown at least 2,000 innings with a .650 winning percentage? Six. Halladay also has the lowest rate of hits allowed among active pitchers named Roy who have played for the Blue Jays! Who cares? Whenever you see a sportswriter employ multiple screens in order to rank or group players, beware. It’s one of their favorite tricks, and it’s usually dishonest, dumb, or both. Ask yourself why the screens are there, and whether they’re necessary, and whether they illuminate more than they obscure. The answers are usually “to get the desired outcome,” “no,” and “hell no.”)

Still, Halladay’s K-rates aren’t extraordinary, so let’s stipulate to Marchman’s categorization of him as a low-K, ball-in-play pitcher and see where he’s going with this.

It’s unusual for such pitchers to throw even one no-hitter, for the simple reason that they don't particularly care whether they give up hits. A pitcher trying above all else to avoid contact will generally not throw first-pitch strikes to 25 of 28 hitters or fail to run a single 2-0 count, as Halladay did Wednesday.

It’s unusual for all pitchers to throw even one no-hitter, or to throw first-pitch strikes to 25 of 28 hitters. Tim Marchman does know this, doesn’t he? And that brings us to Marchman’s greatest failing. Incredibly, in a column dedicated to explaining why low-K pitchers who allow balls to be put into play are less likely than high-strikeout pitchers to throw no-hitters, Marchman never gets around to establishing that they are less likely to do so.

So, are they? I don’t know. I’d sure like to. I do know that Roger Clemens, one of Marchman’s key examples of high-K pitchers who are supposedly more likely to throw no-nos, never threw one, despite being perhaps the greatest pitcher who ever lived. Interesting. I also know that Mark Buehrle, who couldn’t strike you out, has thrown two in the last four years. And of course, Hallady has thrown two this year.

Let’s take a quick look at the last ten no-hitters in Major League Baseball.

Of the last ten no-hitters, guess how many were thrown by pitchers with higher strikeout rates than Roy Halladay? Two.

Only four of the last ten no-hitters were thrown by pitchers with K-rates that were even league average -- and two of those were thrown by Halladay. Let me say that again: 6 of the past 10 no-hitters were thrown by pitchers who had lower-than-average strikeout rates in the season in question.

The last ten no-hitters is obviously a small sample-size, but it’s considerably better evidence against the premise that low-K pitchers are less likely to throw no-hitters than the evidence Marchman provided in support of that premise. Which was, remember, absolutely no evidence at all.

* Ryan’s first few years don’t skew his totals nearly as much; he’s still at 4.5 walks per 9 if you exclude his first six seasons.