How does BABIP tell you any more about that than batting average? You just eliminate the home runs and strikeouts from the equation. How does that tell you whether it was luck or that they hit the ball hard?

here's a link to a related study that seems to support the notion that hitting line drives is higher on the skill-to-luck spectrum than plain batting average:

The article seems to assume that, but presents no evidence for it. In fact it seeme to present the opposite case. Palmeiro, in fact, struggled with the same "bad luck" the following year and is out of baseball. So did Hidalgo.

But I doubt anyone would be surprised to learn that good hitters tend to hit line drives more often than bad hitters. Or, perhaps more to the point, those that hit a lot of soft line drives don't hang around the major leagues very long.

And, not surprisingly, the article finds a lot of speedy players are "lucky" and a lot of slow players are "unlucky".

But hitting infield singles is obviously not a "luck" issue either since guys who have speed are more likely to beat out a ball than guys who aren't. In fact, players with speed often consciously work at making use of that tool by hitting the ball on the ground. And its not really surprising that guys who are slow and hit a lot of groundballs don't tend to hang around the major leagues very long either.

The problem here is that in an effort, I suppose, to simplify, the idea is that you can isolate one aspect of the game and pretend the other aspects don't effect its value. Moreover there is a pretense that players are not consciously adapting their games to get the complimentary benefits of their skills. I think either one is rarely, if ever, true.

the thing is, you could compute the same stat for both pitchers and hitters, and see who's been "hit-lucky" or "hit-unlucky" (where the working assumption is that the ability to produce a certain distribution of batted ball types constitutes "skill", and "luck" encompasses everything else).

this might turn out to be valuable for player projection systems. or it might not--i really don't know whether it's true that producing batted-ball types is a repeatable skill (though i strongly suspect it is), or that it is the only relevant repeatable hitting/pitching skill other than K/BB/HR (i sort of suspect that it isn't, but i really don't know). the most surprising thing to me is that nobody has yet done this on BP or HBT or anywhere else that has come to my attention.

i'll second your comment that this is a fascinating discussion.