When BPers talk about regression toward the mean as it relates to baseball players, we’re usually trying to suggest that a given player’s performance seems unsustainable, in that some aspect of his production to date has represented an extreme observation that’s likely to become less extreme with repeated trials. But the evidence that a player might be playing over his head can take more than one form. In some cases, a player has seemingly fluked into much of the success he’s experienced; for instance, a pitcher could have served up hittable pitches, only to be bailed out by his defense and an assist from Lady Luck, or a hitter could have watched seemingly every bloop and bleeder off his bat fall in for a hit.

But even if a player has really, truly, earned his superlative statistics, as far as we can tell, we still shouldn’t necessarily believe that he can maintain his pace. Sure, it could be that he’s reached a higher plane of performance, but if his current stats over a relatively limited sample seem out of line with what we’ve seen him do in the past, we’d still be wise to exercise caution.

That multi-faceted nature of regression to the mean came to mind while I was watching Marcus Markley Thames give Mariano Rivera a lead that he and the Yankees wouldn’t relinquish last Wednesday, by poking a two-run single through the left side of the infield. By any estimation, Thames’ has been a remarkably successful signing: inked to a minor-league contract in February, Thames is costing the Yankees only $.9 million (plus incentives) this season, and has already contributed a win (1.0 WARP) to the cause in abbreviated, offense-first action. Although recent Yankees rosters have been criticized for their lack of bench depth, Thames’ presence recalls dynasty-era role players in the Chili Davis/Tim Raines/Glenallen Hill mold, making an outsize impact in a limited role (Hill, acquired for the pauper’s ransom of Ben Ford and Oswaldo Mairena, exploded for 16 homers in 132 at-bats in pinstripes down the stretch in 2000, then hit only one more in 2001 before calling it a career).

After 139 pinstriped plate appearances (not counting his brief 2002 tour of duty in New York), Thames’ line stands at .316/.403/.462. His career rates rest at .248/.313/.489, and PECOTA projected a .243/.308/.462 performance prior to this season. That degree of outperformance (at the age of 33, no less) makes Thames a regression candidate, but does he fall into the first category (the fortune-favored) or the second (those who’ve legitimately managed to raise their game, if only for a limited time)?

Knowing what we know about Thames’ season, there’s one good reason to have expected him to exceed his typical seasonal stats. Thames has always been a good fit for a platoon role, but the Yankees are the first team to restrict him to one consistently. Before this season, the right-handed Thames had spent 2006-2009 making 812 plate appearances against righties and 498 plate appearances against lefties, despite boasting a .256/.327/.534 line against southpaws and a much less useful .243/.297/.493 against same-handed hurlers during that period. For the first time in his career, Thames has seen more lefties than righties this season, making 80 plate appearances against the former, and 59 against the latter.

If Thames were merely coming closer to matching his career rates against lefties, his platoon role would explain how he’s managed to exceed expectations. However, Thames is not only blowing away that benchmark, but also defying his past platoon performance: he’s hit .343/.413/.443 against lefties, and .277/.390/.489 against righties meaning that he’s been almost equally productive against pitchers of each handedness.

So has Thames suddenly learned to rake against righties, while taking his true talent against lefties to an even higher level? Not likely. Thames is running a .418 BABIP this season, miles above his .278 career mark, despite the fact that his line-drive rate is languishing below his career rate. That .343 average against southpaws has been buoyed by a .460 BABIP, and his BABIP figures from both sides have been inflated by a sky-high overall 20.8% infield-hit percentage; that game-winning grounder that I watched narrowly elude opposing fielders last week may have been the rule for 2010, rather than the exception.

2006 was likely Thames’ peak performance; that season, the outfielder posted a .284 TAv that doesn’t appear to have been particularly luck-inflated. That’s likely the best-case scenario for the Yankees going forward, but considering Thames’ price tag and the team's judicious usage of its asset thus far, even a fall from grace is likely to leave behind a productive player in its aftermath, especially if he continues to keep his glove in storage.