Friday, January 20, 2012

Regarding Player A vs Player B

Esteemed colleague Sean,

I have parsed the statistics from your post on the relative merits of Player A vs. Player B (SPOILER ALERT: Glavine vs. Schilling), and I'm forced to concede that I think you're right, just not completely for the reasons that you laid out. Turns out, you couldn't really have picked two better pitchers to illuminate the flaws in advanced statistics. I initially preferred Schilling because he put up the same WAR as Glavine in about 75% of the time, but that's more a reflection on WAR than on the merits of each pitcher.

When looking through their peak stats, a few things pop out. Glavine pitched a lot more innings with about the same ERA and ERA+ as Schilling, yet in that time Schilling was more valuable by 6 wins, according to WAR (70-64.2). Off the bat, that looks impossible. How can two pitchers have almost the same ERA (even adjusted for park) yet one throws 600 more innings and has considerably less value?

The answer lies in the problems with WAR, especially when it comes to determining outcomes of skill versus outcomes of luck.

Glavine, for all the moxie and intangibles he was said to possess, was an overachiever in a different sense. For his career he vastly outpitched his statistical profile. Glavine struck out just 1.74 batters for every walk, which puts him behind such greats as Gil Meche (1.76) and knuckleballer Tim Wakefield (1.79). He never looked like he had the stuff to be an ace, but he also never really put up ace-like peripheral numbers. Glavine's WHIP of 1.31 puts him in the company of Derek Lowe and Bronson Arroyo, both good pitchers but not nearly in Glavine's league. Yet, year in and year out he performed better than those numbers would predict.

WAR is an attempt to equalize the advantages of luck (be it ballpark, defense, etc), but in using numbers like WHIP and K/BB it undervalues pitchers like Glavine, who always pitched better than those other numbers would suggest.

Schilling, on the other hand, was a classic underachiever. His strikeout to walk ratio (4.38) is second all time, and his career WHIP of 1.137 is actually better than that of Greg Maddux. So his WAR is inflated, even as the actual numbers don't look as great. Again Schilling's peak ERA+ was 130 while Glavine's was 128. Glavine pitched 20% more innings in that time period and posted 10% lower WAR. That's a flawed statistic.

That Glavine hung on for 300 wins by performing at league average is not that relevant. He certainly got that opportunity because he was a veteran, and he was closing in on 300 wins, and opportunity, as I covered yesterday, cannot be overstated. Should we really reward him for failing to outpitch washed up Jeff Suppan in 2008?* Furthermore, Glavine's peak was great enough that hanging on as an average starter to rack up milestones should have little bearing on his hall of fame candidacy.

*Sean, you made an interesting point that the Suppans and Randy Wolfs of the world have made great careers out of being league average. I would counter that just because general managers vastly over-value these players doesn't mean they're actually adding value to their team.*

Schilling is a little trickier. His entire persona is built on being a guy who stepped up, who was lights out in the playoffs, the bloody sock, World Series co-MVP. Yet, he was injured A LOT throughout his career (making over 32 starts only 5 times to Glavine's 14), and he was statistically an underachiever who never won a Cy Young.

Playoff resumes are just as arbitrary as wins, but he did have those successes. Whether or not one believes in clutch players, Schilling put up incredible numbers whenever he pitched in the playoffs. People may still point to the Cy Young argument, but that's flawed as well. In 1992, Schilling out-pitched  Cy Young runner up Glavine in just about every category, even innings, but Schilling won 6 fewer games and received no support. That's not Schilling's fault, but the fault of voters valuing the wrong numbers. To keep him out of the hall for lack of Cy Youngs is just perpetuating inexplicable voting patterns, but not an accurate reflection of the pitcher Schilling was at his peak. And while he wasn't the workhorse Glavine was, Schilling did pitch over 3,000 innings and won 216 games. He certainly had the longevity of a hall of famer as well.

Schilling was great when he was healthy, though not as great as WAR would have you believe, and he was clutch when it mattered, and in my opinion a hall of famer. He just might not have been better than Glavine.

No comments:

Post a Comment