Bill James, the great baseball historian, raconteur, and statistician, long ago pointed out a few counterintuitive things about baseball statistics. A couple of them are borne out by some snooping I did the other day.
Let's take them one by one. Hitting DOUBLES, for example. Now, it is a very good thing to hit a double. You put yourself in scoring position. You drive in runners from second or third, and sometimes from first -- sometimes, especially with two outs, the double clears the bases and does almost all of the work of a triple. Some of the greatest hitters in history were doubles machines. Stan Musial always was, because he left the batter's box on a dead run. Pete Rose was, for the same reason. Keith Hernandez -- who ought to be in the Hall of Fame -- was. Tris Speaker was. And yet, if you ask, "How much more likely does leading the league in doubles make it that your team will reach the playoffs?", the answer is, "Not as much as you'd think." Since 1995, when baseball instituted the Wild Card, 20 teams that led their league in doubles made the playoffs, while 30 teams did not. Random chance would have predicted 14.4 to make the playoffs. So leading the league in doubles does give you a boost. (I should have looked at doubles + triples, now that I think of it.) It's just not that big a boost, not like leading the league in batting average, or slugging, or home runs.
What's the reason? Well, we can think of a double in two ways: it's a hit that otherwise would have been a single, or a hit that otherwise would have been a home run. If you hit a lot of doubles because you don't hit home runs, that's a negative; if you hit them instead of hitting singles, that's a positive. And it turns out that there are two kinds of ballparks that are favorable for doubles: really good hitters' parks (Colorado) and bad hitters' parks (Arizona). The park in Denver is a great park for doubles because it is a great park for hitting of all sorts. The park in Phoenix is a park for doubles because it is a lousy park for home runs.
Another matter: DOUBLE PLAYS. The double play is a BIG deal. Some DP combinations deserve to go down in history for their excellence: Gene Alley and Bill Mazeroski; Alan Trammell and Lou Whitaker; Ozzie Smith and Tommie Herr. Third basemen and first basemen who start double plays save their teams a lot of runs: Brooks Robinson, Keith Hernandez, Albert Pujols in his prime, Mike Schmidt. The DOUBLE PLAY is not just two outs. It's two outs and the obliteration of a man on base. And yet -- teams that LEAD THE LEAGUE in DOUBLE PLAYS are more likely to be POOR teams rather than great teams. Since 1995, teams leading the league in double plays are a total of 399 games UNDER .500: they lost 399 more games than they won. True, some very good teams, even great teams, led the league in DP's: SIX Cardinal teams were among those. But maybe the worst team in that stretch, the 43-119 Detroit Tigers, LED the league in DP's. That did not mean that they had a good defense. DP totals are not just a function of defensive excellence. They are also a function of HOW MANY RUNNERS ARE ON FIRST BASE with no outs or one out. That is, if your pitching is lousy and your defense is lousy, you will have a lot more opportunities to make a DP than will the team with great defense or great pitching or both.
A third matter: ASSISTS. That's when you throw the ball or (rarely) deflect the ball to somebody else who makes the putout. Obviously, an ASSIST, considered alone, is a good thing. It means an OUT. But how good is it? It depends. Bill James pointed out that Bill Buckner, he of the croquet wickets for legs, made huge numbers of assists at first base, and that led to his being hugely overrated as a defensive first baseman. In point of fact, though, Buckner was not even a good first baseman. He was a LOUSY first baseman (as more sophisticated analysis shows). So why did he make all those assists? Well, it's because he always made the pitcher cover the base unless he was sitting on top of it. Plays that other first basemen made by themselves, Buckner made more complicated by having the pitcher go over even when it should not have been necessary. So the assists were meaningless. The REAL thing to look at, said James, is when the 1B makes an assist by THROWING TO SECOND, THIRD, or HOME -- and those are all difficult plays, and important plays, too. But that means that you have to disaggregate the assist totals.... When you do that, Buckner comes in as lousy.
The lesson of course is that numbers of themselves have no meaning. You have to look at the Big Everything. A home run at Coors Field in 1998 is not worth the same to your team as was a home run at Chavez Ravine in 1966....