clock menu more-arrow no yes mobile

Filed under:

Checking in on the Early Season Regression Candidates

Travis Hafner was so upset with my prediction that his early season performance was due to randomness that he's remained on offense beast just to spite me.
Travis Hafner was so upset with my prediction that his early season performance was due to randomness that he's remained on offense beast just to spite me.

Back in April, I looked at the most and least likely candidates for in-season regression based on their 3-year average BABIP and HR/FB ratio. The idea here was that the first month of the season can lead to some very skewed performances. By regressing player's performance on those two metrics we might be able to gain some idea as to which players were displaying their true talent for the year and which players were simply riding a wave of randomness (good or bad).

We are now past the halfway point, so how do some of those players look today?

Some "Lucky" calls I got right

Russell Martin: At the end of April, Martin had a wOBA of .454, largely due to his hitting homers early and often. The regression said he should be more of a .306 wOBA guy and he's declined sharply since. He's currently at .315.

Jeff Francoeur: One of the perrenial teases in all of baseball, Frenchy was hitting at a .407 wOBA clip through the end of April. Despite all the talk about a changed approach at the plate, Francoeur currently sports a .333 wOBA which is a bit higher than the .309 the regression predicted.

Some "Lucky" calls I got wrong

Travis Hafner: The regression had Hafner and his .418 wOBA taking a big dive to .298. Hafner's wOBA has declined, but only to .405. He's having a great season and at looks to be a legitimate slugger once again.

Lance Berkman: I had my doubts about Lance Berkman still being an offense force, and the regression agreed predicting his .491 wOBA should be closer to .365. Berkman has stayed hot all year, sporting a .417 wOBA and leading the National League in home runs. He's come down, sure, but he's still raking.

Some "Unlucky" calls I got right

Mark Reynolds: Reynolds had an atrocisou start to the seasons, posting a measley .271 wOBA. The regression thought he was more of a .401 wOBA player, which seemed high. He hasn't bounced back that much, but he's sporting a .361 wOBA, which represents a .090 point swing in the right direction.

Nick Swisher: Swisher had a great 2010, but got off to terrible start in 2011 (.286). Since April, he's managed to boost his wOBA .060 points to .346. That's less than the regression predicted (.408), but a solid movement in the predicted direction.

Some "Unlucky" calls I got wrong

Angel Pagan: Maybe my Met fandom influenced my regression, because it said that Pagan's early .249 wOBA should have been more like .392. Not so much, it turns out. Pagan has played a little better of late, but he's still only at .307, well below average.

Hanley Ramirez: Like most people, the regression thought Hanley's .257 wOBA was about .142 points too low. Alas, Ramirez has managed to pull himself to the vaunted level of league average (.334). Technically speaking, he has raised his wOBA by .077 points. But I wouldn't say a prediction that one of the best young hitters in the game wasn't really a .257 wOBA guy is all that impressive. 

The lesson? Prediction is damn hard, doubley so in baseball.