|Wednesday, July 3
Teams have most success in 'hits prevented'
By Rob Neyer
Over the last couple of years, two people have made a big splash in the world of objective baseball analysis: Bill James and Voros McCracken.
Bill James, you know about. This spring he published Win Shares, and while some sabermetricians love it and some can't seem to stop finding things to complain about, the fact is that everybody's talking about it. As for Voros McCracken ... why, even Bill James is talking about him. As James wrote in last fall's New Historical Baseball Abstract, "McCracken argues that, other than getting strikeouts and allowing home runs, there is little that a pitcher can do to cause his hits allowed to be higher or lower. Therefore, he argues, if the pitcher's hits allowed are higher or lower than we might expect (in view of his strikeouts), this reflects not skill, but pitching in good or bad luck."
McCracken's findings represented a breakthrough in the field, and I reported on them in this space back in January of 2000. James used McCracken to inform his latest work on Win Shares, and recently Dick Cramer -- like James, a pioneer of sabermetrics -- took McCracken's research in a different direction. Last weekend at the annual convention of the Society for American Baseball Research, Cramer presented a paper titled "Evidence that fielders are much more important than pitchers in preventing opponents' base hits."
Let me stress once more -- as Cramer did -- that nobody is suggesting that pitchers don't influence how many hits they allow. If the pitcher strikes out the batter, there won't be a hit. Similarly, if the pitcher allows a home run, it's his responsibility and his alone (aside from those rare occasions when the fielder has a chance for a catch).
The point is that if the ball is put in play, whether or not it becomes a hit or an out has virtually nothing to do with who's pitching (unless the pitcher throws knuckleballs, as there's some evidence that knuckleballers are different).
How do we know this is true? As McCracken demonstrated and Cramer confirmed, there simply isn't any season-to-season consistency for batting average on balls in play (BABIP) against pitchers. None. It's a basic principle of baseball statistics that if something's an ability, it will repeat. A player with the ability to hit a lot of home runs will hit a lot of home runs year after year. A player with the ability to steal a lot of bases will steal a lot of bases year after year. A pitcher with the ability to strike out a lot of hitters will strike out a lot of hitters year after year.
But knuckleballers aside, pitchers do not show any consistency in the batting average they allow on balls in play. There simply isn't anything there.
However, what Cramer discovered is that there is some season-to-season consistency on the team level. There's not a lot of consistency, but there's more than enough to suggest that it's not random, that while pitchers may not affect the batting average on balls in play -- or to use Cramer's measure, the "hits per batted ball in play" (HBBP) -- teams do.
This makes sense if you think about it, because not all defenses are created equal. We would expect teams with good fielders to turn batted balls into outs more often than teams with poor fielders, right? Apparently the differences between defenses aren't significant enough to show up when analyzing individual pitchers, but they are significant enough to show up when you look at teams.
How do we know it's due to the defense rather than, say, the ballpark? Because Cramer found a correlation between HBBP and a metric he labels "Conventional Fielding Skill," which is an amalgam of errors, double plays, and passed balls. Of course, those aren't perfect measures of fielding skill, but I think most of us can agree that good defensive teams tend to fare better in those traditional measures than poor defensive teams.
From there, Cramer considered "individually some of the largest, less-likely-to-be-chance, season-to-season changes in team hits prevented," relative to what we would expect given the general tendency of all pitchers to allow hits on approximately 30 percent of the balls in play.
From 1990 through 2001, the second-best "Hits Prevented team" was ... the 2001 Seattle Mariners, who prevented so many hits that they saved 141 runs. Here are the top five and bottom five since 1990:
Runs 1998 Reds +145 2001 Mariners +141 1991 White Sox +135 1990 Athletics +112 1998 Yankees + 94
Cramer also found a huge season-to-season improvement for the 1991 Atlanta Braves, who featured virtually the same pitching staff from the previous season, but turned over the defense at every position except catcher.
Why does all of this matter? Because until Voros McCracken came along, nobody had anything better than a wild guess as to how much of defense is the pitchers and how much is the fielders. Now, with the work that McCracken has done, and the follow-up work by James and Cramer and others, we're getting very close to the answer. And for people like me, that's pretty exciting.