from Amazon
Monday, February 08, 2010
Fangraphs now has some Splits data
Personally, I’d like to see each split data have its own subtab under the Career/Year tab. If I’m interested in L/R splits, I’d like to see the STandard / Advanced / Batted Ball all together, not navigate through the other splits I don’t care about (at that moment). Also, as more splits are added in (like men on base, role, park, etc), it just makes this more cloudy. Basically, splits should extend horizontally (add more tabs) rather than vertically.
David’s taking suggestions, so post it here or there.
***
I linked to Granderson’s splits, who has a .270 wOBA against LHP (685 PA) and .380 against RHP (2211 PA). (If David wants to wow us, when he does his leaderboards, give us the “differential”, and also shows us the batter handedness on that table.) Andy said that for LHH, that you would regress the observed split (110 points in this case) 50% toward the mean if you had 1000 PA against LHP. In this case, with 685 PA, you would regress 1000/(1000+685) or 60% toward the league observed splits for LHH, which I think is like 27 points.
So, you regress 110 60% toward 27, or 110*.4 + 27*.6 = 60. So, our estimate of Granderson’s handedness split is .060 in wOBA. He’s a career .358 wOBA, with 24% PA against LHP. We can reconfigure his observed split to be:
.358 = .24(x) + .76 (x+.060)
So, his LHP adjusted-observed split is .312 and his RHP split is 60 points higher or .372. This is his observed baseline. If he’s a true .350, as opposed to his observed career .358, you bring him down by 8 points on both sides.
In any case, with his excellent fielding, even his poor hitting against LHP means he’s an average player. When you are an average player against the platoon, this is not really the kind of guy you need to be platooning. (Unless of course you already have a good platoon partner, so you might as well take advantage when you can.)
***
Here is Andre Ethier in high-leverage situations. Dave in the THT Annual noted how Ethier had back-to-back years of great high-leverage performance. We see that this can extend even to 2007. Here are his wOBA in hi-lev, 2007-09: .403, .451, .454. (Hi-Lev must be based on LI of at least 2.0, since we have about 10% of Ethier’s PA. This is different from b-r.com’s hi-lev that uses LI of 1.5 or higher, which is about 20% of PAs.) Anyway, compare that to his overall career wOBA of .363. His incredibly sucky hi-lev in his rookie year really drags down his career split in hi-lev.
Anyway, love this stuff.