From page 289 of "Baseball for Dummies", by Joe Morgan (with Richard Lally):
"The Lowdown on Statistics" -- Everyone believes that a .300 hitter is a good player and that a pitcher with a low ERA is a good pitcher. That belief is not necessarily the case. . . A .300 hitter makes seven outs for every ten at-bats, and if his seven outs come with men on base and his three hits come with no one on base, these hits are not very productive. . . Likewise, many pitchers pitch just good enough to lose. . .Run production is how you measure hitters. Wins and losses are how you measure pitchers. Batting averages and ERAs are personal stats."
>>Again, where to begin?
Let's take the case of the .300 hitter who makes all his outs with runners on base, and all his hits with men on. Either he's talking about (a) just ten at-bats or (b) a guy who does this sort of thing over the long haul.
If we're talking about 10 at-bats...who cares? Remember, the law of small numbers is: there is no law of small numbers. If we're talking about a guy who does this over the course of the season, well, I'd like to see that guy. I'd like Joe Morgan to show me anybody who -- over the course of his career -- had a markedly different average with runners on base as opposed to with the bases empty. Maybe they exist. But let's be reasonable: over the course of time, most players are about the same with bases empty or with dudes on base. On top of all this, of course, Morgan's chosen a terrible metric to measure players (batting average).
Now for the really crazy stuff. "Run production is how you measure hitters." He's talking about runs and runs batted in (or so he says in another part of this little sidebar). Two of the most team-dependent stats you could pick. Wins and losses, even more so. It's the year 2005, and we're still measuring how good a pitcher is by his won-loss record? Tell that to Roger Clemens. Tell that to Ken Tremendous, and he'll go to your house and murder your dog.
The best line, of course: "Batting averages and ERAs are personal stats."
I'm sorry. What? You mean personal, like, they're just for that hitter or pitcher -- like a sentimental photo? Or you mean personal, as in, they are the opposite of team-dependent, and therefore much better at measuring players' abilities than fucking wins and losses and run production (again, ignoring that batting average is a terrible hitters' metric)?
Imagine using this standard for any other line of work. Let's say I'm the manager of the factory where they manufacture and ship Joe Morgan Punching Bags (they come with a picture of Joe Morgan on them). I want to know how good John Kruk is at packaging up these bad Larries. Kruk works down on assembly line C. If I want to know how good he is at packaging up JMPBs, should I count how many boxes he "personally" packages, or how many everybody on assembly line C puts together?
Joe Morgan says: the bottom line is, if you want to know how good a player is, forget any information about the player alone. Ignore all data that tell you what the guy does, holding the team he plays for as a variable. Look at the team-related numbers. See how many wins a pitcher has, without even looking at his run support. See how many runs a guy scores, without looking at the guys hitting below him, or the player's on base percentage.
I'm just saying, if Joe Morgan's dog is dead tomorrow, you know who did it. Sorry, Ken.
