Doing the math on Mike Napoli in left field

If you have been following the Rangers at all over the last week or two (and if you haven't been, I'm not sure why you'd be here, now that I think about it), you know that a big topic of discussion of late has been Jeff Banister's decision to start using Mike Napoli in left field, particularly against lefthanded pitchers.  Napoli, of course, is a former catcher who has moved to first base, and who had, prior to this past week, never played the outfield as a professional.  The Rangers have struggled against lefthanders, and in particular, have struggled to find a viable left field option against lefties, and so Banister turned to Napoli.

(And if you are asking, why not play Mitch Moreland in left field and Napoli at first base, congratulations!  You are the 1 millionth person to ask that question.  The answer, apparently, is that Moreland's ankles make playing the outfield unworkable).

My reaction when I heard that Banister was doing this can be summed up in this clip:

That being said, fifteen years ago, I would have been thrilled with this move, and credit Banister with being an outside-the-box thinker who wasn't afraid to try new things.  I was much less concerned, then, with defense and the losses there than I am now.  Have I gone too far the other direction?  Am I just being too conservative?

So let's look at what the calculations and projections suggest the impact of this move will be.

First of all, let's be clear...some people have said that this is to get Napoli's bat in the lineup against lefthanders.  That's false.  Napoli has been starting against lefties since he got here, in a platoon at first base with Mitch Moreland.  Putting Napoli in left field is to get Moreland's bat in the lineup against lefties, not Napoli's.

So, what would we expect from Moreland against a lefthanded pitcher?  For his career, Moreland is a .232/.289/.358 hitter against lefties, in 516 plate appearances, good for a .286 wOBA.  This season, Moreland is hitting .246/.289/.389 against lefties in 135 plate appearances, for a .295 wOBA.  2015 is largely in line with his historical averages -- its a little better, actually -- so in the interest of giving Banister the benefit of the doubt for this move, let's assume Moreland is reasonably expected to put up a .295 wOBA the rest of the way.

How about Napoli's defense?  The floor defensively is usually put at a -20 runs over 150 games in left field -- again, that's conservative, because there's a self-selection at work, and players who are worse than that usually don't get sent out there.  On the other hand, Hanley Ramirez is at a -32.7 UZR/150 this year, so it can be worse.  Still, in an effort to be conservative, I think putting Napoli at -20 runs over 150 games is a reasonable estimate.

So what we have to figure out is if the number of runs you'd be expected to lose with Napoli in left field at -20 per 150 games is offset by the gains in Moreland's .295 wOBA against lefties compared to whoever would be in the lineup instead of him -- most likely Drew Stubbs, Ryan Rua or Ryan Strausborger, although possibly Will Venable.

But how do you convert difference in wOBA into runs?  Fangraphs explains:

A good rule of thumb is that 20 points of wOBA is worth about 10 runs above average per 600 PA. This is not a precise measurement and specific calculations are always better, but if you’re looking for an approximate rule of thumb, this may be useful.

Because I don't want to do a bunch of separate calculations, I'm just going to use the 20 points = 10 runs for our purposes here, but if you want to do more specific ones, feel free to share in the comments.

So, what do we see when we compare our other left field options to Mitch Moreland, when it comes to offense?

ZiPS projects Ryan Rua to have a .287 wOBA.  That's against everyone, not just lefties, but regardless, that suggests that there's basically no advantage to having Moreland hit against lefthanded pitchers instead of Rua, even without regards to the defensive advantage Rua has over Napoli in left field.  In his major league career, Rua has a .328 wOBA against lefties -- a significant edge over Moreland, even without taking into account defense.

If you prefer looking at just this season, because you think, for whatever reason, that's more indicative that projections (even though there's just 53 plate appearances), Rua has a .280 wOBA against lefties -- in other words, somewhere around 5 runs per 600 PAs worse than Moreland has been this year against lefties.  Rua, defensively in left field, is probably a -5 to -10 defender, so even using that methodology, the math suggests you are better with Rua than with Moreland.

Ryan Strausborger is projected by ZiPS to have a .270 wOBA going forward this year -- like Rua, this is against everyone, not just lefties, so its going to be lower than if we just looked at lefthanders.  If we just use the projected wOBA without splits and compare it to Moreland against lefties, Moreland has about a 12.5 run advantage over 600 PAs.

However...Strausborger is considered to be a very good defensive left fielder, and a decent defensive center fielder.  You are probably look at him being a +5 defender in left field, conservatively speaking, which means that you still have a significant net edge when you use Strausborger.  And that edge would be greater if you adjusted the projected .270 wOBA upwards to account for him playing against just lefties.

If you prefer to use this season's stats, and assume that's what he will do going forward -- and against, we are talking about 29 plate appearances, or a figure small enough to be pretty close to useless -- Strausborger has just a .217 wOBA.  That's roughly 80 points lower than Moreland, translating to -40 runs, which would give the advantage to going with the Moreland/Napoli pairing instead.

Then there is Drew Stubbs.  ZiPS projects Stubbs to have a .286 wOBA the rest of the way -- again, without splits, so you would have him as roughly equivalent to Moreland in terms of hitting against lefthanders.  Stubbs has a big defensive advantage over Napoli in the field, so that would seem to be a slam-dunk in favor of using Stubbs.

If you want to look at just this season, Stubbs has a .250 wOBA against lefties.  That translates into about a 20 run deficit to Moreland offensively, compared to what he has done this year...I suspect that Stubbs is more than 20 runs better than Napoli in the field, not to mention his edge in the basepaths, so I think Stubbs still gets the nod, although its much closer.

So looking at the numbers, I think a lot of this comes down to, what do you trust to be more indicative of what you think will happen over the final three weeks?  What ZiPS says, or what has happened so far this year?  If you go with ZiPS, going with Napoli in left field instead of Rua, Stubbs or Strausborger seems to be a clear mistake.  If you put more weight on 2015 performance, the answer isn't so clear.

And again, this analysis is strictly math-based.  It doesn't take into account the possibility that Banister may feel that Moreland and Napoli will hit better if they are both playing every day, versus being in a platoon role.  It may be that Banister has information that the other options are likely to do worse than their projections, because of current conditions or what have you.  And it may be that this is a confidence thing, that Banister believes that his players have a greater level of trust in the veterans than the others, and that they are going to play better with them in the lineup.

But if we look at just the math, and use just the projections, the actual calculations don't support the move.