Moneyball by Lewis, Michael (mobile ebook reader txt) 📕

Obviously some pitchers give up fewer hits than others, but that might be because some pitchers had more strikeouts than others, and, therefore, allowed fewer balls to be put into play. But it was generally assumed that pitchers could affect the way in which a ball was put into play. It was generally assumed that a great pitcher, like Randy Johnson or Greg Maddux, coaxed hitters into hitting the ball in a way that was less likely to become a hit. The trouble with that general assumption is that it didn’t square with the record books. There were years when Maddux and Johnson were among the worst in baseball in this regard.

If Voros McCracken was right, then what had heretofore been attributed to the skill of the pitcher was in fact caused by defense, or ballpark, or luck. But the examples of Greg Maddux and Kevin Millwood suggested that defense and ballparks might be of secondary importance. They pitched in front of the same group of fielders, and, usually, in the same ballparks. That led Voros to a second radical thought:

What if what was heretofore regarded as the pitcher’s responsibility is simply luck?

For a century and a half pitchers have been evaluated, in part, on their ability to prevent hits once the ball is put in play. A pitcher who suffered a lot of balls falling for hits gave up more earned runs and lost more games than one who didn’t. He was thought less well of than a pitcher against whom balls in play were caught by fielders. A soon-to-be unemployed young man, soon to be back living with his parents in Phoenix, Arizona, begged to differ. He was coming to the conclusion that pitchers had no ability to prevent hits, once the ball was put into play. They could prevent home runs, prevent walks, and prevent balls from ever being put into play by striking out batters. And that, in essence, is all they could do.

Voros McCracken had a radical theory. And he was staring at a lot of hard evidence supporting it.

What happened next bolsters one’s faith in the American educational system: Voros McCracken set out to prove himself wrong. He wrote a computer program that paired major league pitchers who had very similar walks, strikeouts, and home runs—but had given up very different number of hits. He located ninety such “pairs” from the 1999 season. If hits per ball in play were indeed something a pitcher could control, Voros reasoned, then the pitchers who had given up the fewer number of hits in 1999 would proceed to give up fewer hits in 2000. They didn’t. There was, in fact, no correlation from one year to the next in any given pitcher’s ability to prevent hits per balls in play.

Instead, baseball kept thrusting before Voros beguiling situations explained by his theory. A couple of months into the 2000 season, for instance, the newspapers were full of stuff about how heretofore mediocre White Sox pitcher James Baldwin, off to a great start, had somehow become the next Pedro Martinez. Voros looked more deeply at the numbers and saw that Baldwin had extremely low number of hits for the number of balls he’d put in play. His earned run average was sensational—because he’d been lucky. Sure enough, the hits started falling and Baldwin regressed to mediocrity and people stopped putting his name in the same sentence with Pedro Martinez.

For pretty much the whole of 2000 Voros McCracken, as he put it, “went looking for a reason Maddux got hit in 1999, and to this day I’m still looking for it.” At length, he penned an article revealing his findings for baseballprospectus.com. Its conclusion: “There is little if any difference among major league pitchers in their ability to prevent hits on balls hit into the field of play.” ESPN columnist Rob Neyer saw Voros’s piece and, stunned by both the quality of the thought and the force of the argument, wrote an article about Voros’s article. Several thousand amateur baseball analysts wrote in to say that Voros’s argument, on the face of it, sounded nuts. Several suggested that “Voros McCracken” might be a pseudonym for Aaron Sele, a well-hit pitcher then playing for the Seattle Mariners.

Bill James also read Rob Neyer’s article. James wrote in, and said Voros McCracken’s theory, if true, was obviously important, but that he couldn’t believe it was true. He—and about three thousand other people—then went off to disprove it himself. He couldn’t do it, and the three thousand other guys couldn’t either. About the most they could suggest was that there was a slight tendency for knuckleballers to control hits per balls in play. Nine months later, on page 885 of his mammoth Bill James Historical Baseball Abstract, James laid out Voros McCracken’s argument, noted that Voros McCracken “is NekcarCcM Sorov spelled backwards,” then went on to make four points:

Like most things, McCracken’s argument can be taken too literally. A pitcher does have some input into the hits/innings ratio behind him, other than that which is reflected in the home run and strikeout column.

With that qualification, I am quite certain that McCracken is correct.

This knowledge is significant, very useful.

I feel stupid for not having realized it 30 years ago.

One of the minor consequences of Voros McCracken’s analysis of pitching was to lead him to Chad Bradford, Triple-A pitcher for the Chicago White Sox. Voros had developed a statistic he could trust—what he called DIPS, for defense independent pitching statistic. It might also have been called LIPS, for luck independent pitching statistic, because the luck it stripped out of a pitcher’s bottom line had, at times, a more warping effect than defense on the perception of a pitcher’s true merits. At any rate, Chad Bradford’s Triple-A defense independent stats were even better than his astonishingly impressive defense dependent ones. (Chad pitched a total of 202% innings in Triple-A, with an earned run average of 1.64.) And