July 6, 2010
The Replacement-Level All-Stars
Each year’s announcement of All-Star Game rosters inspires a round of internet remonstrations, focused on exclusions and inclusions alike. Rather than lend my keyboard to that chorus, which seems to have hit all the right notes already, I thought I’d take a quick look at the initial All-Star rosters with the aid of replacement level. Even the worst All-Star selected easily surpasses major-league replacement level—yes, even Omar Infante, and with plenty of room to spare—but one wonders (okay, so maybe I’m the only one wondering) how each pick would stack up were the replacement-level baseline adjusted for All-Star talent. In other words, let’s pretend that the members of the AL and NL All-Star teams seceded from their circuits and formed their own league, with the rest of the majors serving as a talent pool for potential stand-ins. Would any replacement-level players lurk among those initially chosen to travel to Anaheim? And if so, which would lay claim to the title of Replacement-Level All Star?
For the purposes of this post, I’m going to use a version of VORP with a few adjustments—by which I mean, a version to which I’ve neglected to make most of VORP’s adjustments. Instead of using the same name for this bastardized version of the original stat (which might make Keith Woolner roll over in his Progressive Field office), or even rechristening it VORAS, I’m going to go ahead and call it FAUXRP, in recognition of its status as a pale imitation. If you say it fast, it rhymes with VORP, but it lacks the positional and park adjustments of its antecedent. Combine those omissions with the fact that FAUXRP, like VORP, ignores defensive performance, and it’s clear that what we have here is the statistical equivalent of an evolutionary dead end, which may never again rear its mutant head on these pages.
I divided “True Runs” by outs for hitters, and runs allowed by innings pitched for pitchers, and treated 80% of the weighted average of each group’s performance as All-Star replacement level (regarding starting pitchers and relievers as separate groups). Then I took the difference between the number of runs scored (or allowed) that each player chipped in and the number of runs scored (or allowed) that a “replacement-level All Star” would have contributed, and called it FAUXRP. Justin Morneau leads the majors with 45.5 VORP, and leads the All-Stars with 33.2 FAUXRP (Morneau's lower FAUXRP total stems from the fact that the All-Star-only replacement level is significantly higher than the all-inclusive first-base replacement level used to determine his VORP).
The average RA of the starting pitchers selected for the All-Star staff is approximately 3.03, which makes their replacement level 3.78. By that benchmark, two starters on the staff would qualify as sub-replacement. The average All-Star reliever RA is 2.28, so let’s call reliever replacement level 2.85. Three firemen would fall below that threshold. Here are FAUXRP’s least-favorite five:
It’s no coincidence that three of the five pitchers featured here hold the distinction of being named their teams' only All-Stars. A fourth hails from the AL manager’s club, but we can't insinuate any favoritism in his case, since Hughes was selected via the player vote.
The following batters failed to post a positive FAUXRP:
Five of these six players were either chosen as their teams' lone representatives, ushered in under the mysterious Selig-inspired “multi-position rule,” or (in Derek Jeter's case) recognized as one of the members of the famous athletic foursome in those Gillette commercials. In fairness, all of these players have played a premium defensive position, so they’re getting dinged unfairly here by the absence of a positional adjustment. For instance, Jeter leads AL shortstops in VORP, despite a down year, and seems to have held his own on defense; since the team needs a shortstop, one would be hard-pressed to argue that he doesn’t belong. Given that Molina has amassed -4.3 VORP in the National League, however, no degree of positional adjustment could’ve saved him from posting a negative FAUXRP. The lowest-FAUXRP (4.99) player who doesn’t hail from one of the most difficult positions on the defensive spectrum is Alex Rodriguez, a Joe Girardi pick.
Take all of that for what it’s worth (roughly two minutes of your time, which is all, I hope, that it took you to read), but remember to appreciate the skill level on display come July 13th, even in the absence of a few noteworthy names.
While I have what remains of your attention: I’ll be chatting live today beginning at 2:00 PM ET, so if you have any pressing questions that require my attempts at answers, feel free to drop by.