Tuesday Nats Stats: What's up with Dunn's batting splits?

Does Adam Dunn really hit worse when he plays RF than when he plays LF?  Here's a quick-and-dirty look based on a thread comment that was getting waaaay too long.

Okay, here are Dunn's overall batting lines:

Career- 248/383/520, 131 OPS+, 13.85 HR/AB, 16.9% walk rate, 1.6 K/BB

He's a slugger. Low average, hits dingers, walks a lot, and strikes out even more.  This we know.  What's he done, season-to-date?

2009- 286/425/579, 157 OPS+, 12.09 HR/AB, 19.8% walk rate, 1.2 K/BB

Looks like Dunn's having a career season.  He's added 40 points of BA, slightly increased his walk rate, while holding down his Ks and still hitting the long balls.  Frankly, it's more than I ever expected we'd get from Dunn.  There might be some regression to the mean over the rest of the season, but that's a heck of a start.  Of course, keep in mind that those numbers only cover 167 PA.

Alright, what happens when we split out his time in RF vs time in LF?  Our 2009 sample size is even smaller here, but the raw-numbers-to-be-taken-with-a-grain-of-salt are:

LF- (91 PA) 333/495/725, 198 OPS+, 8.6 AB/HR, 24% BB, 0.9 K/BB

RF- (57 PA) 235/298/412, 81 OPS+, 25.5 AB/HR, 8.8% BB, 3.2 K/BB

1B- (18 PA) 250/500/500, 140 OPS+, 12 AB/HR, 50% BB, 0.7 K/BB

Dunn goes from league-killer to league-lagger if you put him in RF instead of 1B/LF.  WTH?  That makes no sense.  Is it the sample size?  Let's take a look at his career splits:

LF- (4009 PA) 250/381/527, 132 OPS+, 13.6 AB/HR, 16.5% BB, 1.6 K/BB

RF- (376 PA) 245/383/448, 113 OPS+, 21.9 AB/HR, 17.3% BB, 1.4 K/BB

1B- (455 PA) 224/396/501, 131 OPS+, 13.1 AB/HR, 20.6% BB, 1.2 K/BB

Bizarrely, his career numbers seem to support worse hitting in RF vs LF/1B.  What could explain this?  Let's take a look at his career platoon splits:

vs RHP: 253/394/542, 140 OPS+, 13.0 AB/HR, 18.1% BB, 1.4 K/BB

vs LHP: 237/359/475, 111 OPS+, 16.0 AB/HR, 14.2% BB, 2.1 K/BB

Well, hmm... Dunn doesn't hit the lefties as well, and that performance looks slightly worse vs LHP than it does in RF, but not by much.  Could it be platoon advantage, if he only ends up in RF when there's a lefty on the hill (or at least, more often when there's a lefty pitching)?

After a bit of fiddling with baseball-reference's event finder, I found that his PA vs LHP by position were 1241(LF), 175(1B), and 106 (RF).

That means as a LF, 30.9% of his PA were vs LHP as a 1B, 38.5% of his PA were vs LHP, and in RF, 28.2% of his PA were vs LHP.

WTH?

That's right, he faces more righties (whom he crushes, by his career splits) as a RF, but he still bats about 15% below his career average!  I'll stay that again: on average, Dunn gets more platoon advantage when he's played in RF, but his hitting falls off!  Having to remember to hit a different cutoff man seems to adversely effect his pitch selection, or something.  (And before you say, "what about home/away splits," I took a quick look, and they're too small to account for the difference, even if he only played RF on the road.)

Now, it's possible I made a stupid arithmetic error, but this result just seems... goofy.  It's also possible that's it's an artifact of random chance and a relatively small sample--even 376 PA isn't that many (a bit over half of a season).  What could explain this?  I welcome your speculation in the comments.