Sunday, March 16, 2008

In Which Mr. Boswell Discovers Baseball Reference

Boz writes a fun little column using Baseball Reference's similarity scores to compare today's players to old guys who are probably dead.

The team indicates that they futz around with similarities from time to time.
Last winter, Manager Manny Acta consulted with a stat-fanatic friend who suggested that inconsistent starter Joel Hanrahan should be switched to the bullpen because other power pitchers with similar minor league statistical progressions had made the switch. Last week, Hanrahan struck out eight Braves in three innings. "He was throwing 95 to 98 [mph] and made some good hitters look real stupid," Bowden said. "It's too early to tell with Joel. But when you find career parallels, pay attention."

The key is that the BBRef similarity scores are toys more than hardcore examples of analysis. The big flaw is evident:
In the case of the Nats, some of the answers are shocking. It's no surprise that Zimmerman's offensive numbers after two seasons are virtually identical to hitters such as Cal Ripken, Ron Santo, Eric Chavez and Greg Luzinski. However, who would think Pe¿a's career, through age 25, compared plausibly with Albert Belle, Willie Stargell, George Bell and Jermaine Dye?

These sim scores don't adjust for era. Given the offensive explosion of the last 10-15 years, today's above-average player is going to look like a great player from the '60s.

I love Zimmerman, but he's not Ron Santo -- who had the misfortune of starting his career in the low-scoring '60s. Same thing with Lopez. Further in the column, Boz compares him to Bobby Grich, a border-line Hall-of-Famer. Nope!

About a week ago, I looked at the similar players from Baseball Prospectus' PECOTA projections. Those are adjusted for era so that a great player in the 60s is the same as a great player in the 90s.

There, Ryan Zimmerman is compared to players like David Wright and.... Ron Santo.... So forget what I said!


Post a Comment

<< Home