This year, Roger Bernadina has been all over the Nationals lineup--where should he be? So far this season, he's compiled this stat line:
That's almost exactly league-average (99 OPS+), and his BABIP is also right around league-average, so he hasn't been especially lucky or unlucky. He has an okay batting average, acceptable patience, and a bit of pop (150 isolated power). That OBP isn't quite high enough for a leadoff spot (you'd like something more like 340-350), meaning he'd need to increase his contact rate or take more walks if you wanted to bat him 1st/2nd. He has swiped a few bases this season, and was successful in 14 of 16 tries, which is nice for the top of the order. League-average hitting, while nice from a SS or CF, isn't really what you'd like to see in the middle of the order--you want to be striking fear into opposing pitchers with high OPS guys who will punish mistake pitches and not hack themselves into a strikeout if they get pitched around. So we've ruled out 1-5; it seems a shame to put a decent bat somewhere in 6-8, although there's something to be said for being able to "turn over the order." Having a reasonable threat to drive in Zim/Dunn/Hammer after they hit an RBI double, or to restart a rally as "top of the order 2.0" is a good thing. After the jump, a quick look at how Bernie's done this season by where he's batted.
Here's how Bernie has done at different places in the order. Keep in mind that he only has 20-40 PA outside of the 5-6-7 spots, so take the splits with a grain of salt (and he has less than 100 in each of 5-6-7!). There are two OPS+ columns: The first one (tOPS+) is how Bernie's done relative to his overall numbers (so >100 means he bats better than average for him in that spot, and <100 means he bats worse than average for him in that spot). The second one (sOPS+) is how Bernie's done relative to the league in that spot (so >100 means he's above league average for players hitting in that spot, and <100 means below league average). This includes PH appearances, which I couldn't filter out (except for times he batted 9th, which I figured were all PH appearances, so I deleted them).
Interestingly enough, Bernadina has put up the best numbers in the 6- and 8-spots. He has few enough PA at 8th that it's hard to say that result is meaningful, but he seems legitimately league-average or better hitting 6/7. Although it's only in 45 PA, his OBP as a leadoff hitter is actually right about where you'd want it--.341! Batting 2nd or 5th, he's struggling. Batting 3rd he seems okay, but we can't say much about performance in only 27 PA.
Take a look at his BABIP in each spot: it's an eye-watering .397 as a 6-hitter, which means he's been pretty lucky hitting there. Likewise, he's got a pretty unlucky .235 hitting 5th--is his apparent success in the 6-spot a factor of random chance? Well, let's adjust his stats a bit: we can estimate an "expected" BABIP based on things like how often he strikes out, his speed, and his distribution of hit types. Roger "should" have a BABIP of around .325. If we drop his 6-spot stats by an even 20% to offset his luck with balls finding grass instead of gloves, his slash stats drop to about 270/325/443--look familiar? That's pretty close to his overall batting line, but with a bit more SLG--solidly league average or a bit better. Doing the same trick to put a 40% bonus on his unlucky 5-spot stats gives us 263/340/388--below league average, and definitely stretching the limits of this sort of statistical game (since he only has 50 PA batting 5th). Still, he's hit better in 6th than in 5th, even trying to normalize his BABIP to even out luck. Just for reference, applying a 325 BABIP to his 7-spot stats gives a 280/335/430 line--right around his overall numbers again, and solidly above average.
Overall, Bernie's stats look well-suited to hitting 6th or 7th, and his performance (normalized for luck) has been best there this season. Note well, that he seemed to alter his approach at the top of the order to up his OBP, although the small number of PA there makes it hard to say whether the effect is for real.