where they ain't II

some followup re yesterday's discussion of BABIP and luck:

first, jeff has another post up on the subject at Brew Crew Ball, this one showing which players undershot their career BABIP by the widest margins in 2005 -- the year's unluckiest hitters, if you will. none of the 29 guys on that list has hometown implications, altho a few (jacque jones, jose vidro, bret boone, eric byrnes) were real or hoped-for acquisition candidates.

once again jeff adds some empirical context to his post by looking back to 2004, listing that year's unluckiest hitters and tracking their performance into 2005 to see if their luck turned around. those results were not quite as one-sided as the outcomes for 2004's luckiest hitters (who, you'll recall, ran out of luck in 2005), but jeff offers a plausible explanation for that. i'll let you head over there to read it for yourself, check the data jeff presents, and form your own conclusions.

jeff also was generous enough to e-mail me a complete list of BABIP figures for the 2005 cardinals, with the team's offseason acquisitions tacked on as a bonus. the table includes not only each player's 2005 and career BABIPs, but also the percentage of at-bats that result in a BIP. it's a long list, so i'm not going to roll it into the post; click here if you want to check it out.

in the comments to yesterday's post, a number of readers questioned the notion that changes in a player's BABIP are necessarily the result of luck. why couldn't a higher BABIP reflect increased batting skill? a guy learns to take the slider the other way, and instead of grounding out to short he slaps singles to the opposite field; there's your increased BABIP. the guy didn't get lucky; he became a better hitter.

these kinds of adjustments take place all the time, right? even veteran hitters are constantly tinkering -- trying new stances, moving up on the plate, dropping their hands, shortening or lengthening their swings. and at the same time, pitchers and hitters are constantly adjusting to each other. the pitchers find they can jam a guy on hard stuff inside and pop him up, and they keep throwing it until the hitter adjusts and yanks a couple out of the park; then they change their approach, set him up for breaking balls low and away, and the hitter starts reaching for those and dribbling weak grounders; he eventually learns to look for those pitches and poke them down the line, which occasions yet another round of adjustments. this stuff never ends.

so couldn't variations in BABIP result from this natural ebb and flow, rather than blind luck? and my answer to that would be: absolutely. there are no doubt periods of time -- a week, a month, a whole season -- in which a given batter manages to stay a step ahead of the pitchers, make better and quicker adjustments, recognize pitches and patterns with more clarity than usual, and consequently hit the ball harder and better than he typically does. maybe that's what womack did in 2004; maybe it's what nunez did in 2005. or maybe they really were just lucky, and the ball had eyes for 'em one summer.

the question is whether such variations -- whether attributed to luck, adjustment, or god's will -- can produce sustained improvement in a batter's performance. and more to the point: can we find a way to distinguish sustainable improvement from the ephemeral kind? it's a more or less eternal question; get it wrong and you end up overpaying for tony womack or adrian beltre or (cough) tino martinez, or you hang onto a player a year too long or get rid of him a year too soon. nobody can get these things right with 100 percent certainty, but is there a way to improve the odds?

that's what jeff is asking with his BABIP stuff: does it give us a means to more reliably distinguish sustainable improvement from fleeting spikes in performance? maybe it does; maybe it's one of a range of tools (along with k rates, batted-ball data, jc bradbury's prOPS) that can help teams make better decisions in this regard. keep in mind that jeff is only looking at the margins of the spectrum -- the top and bottom 5 percent. BABIP might not be a useful benchmark for 90 to 95 percent of the hitters in a given year. it will be interesting to see how the ~60 players on jeff's luckiest/unluckiest lists for 2005 fare in 2006.