---

Monday, October 17, 2011

Should the Yankees Swap Brett Gardner and Curtis Granderson?

By just about every objective measure, Brett Gardner is probably a better defensive player than Curtis Granderson.  So logically you’d think it probably makes sense to move Gardner to CF and Granderson to LF.  What I don’t know is how much of a difference it’d really make.

I’ll look at this using zone rating, since it’s more granular and transparent.  I don’t think using other metrics would change things much, but if anyone wants to show that I’m wrong go for it.  Standard caveats about defensive metrics apply, etc.,  Here’s the methodology I use for converting zone rating to plays/runs compared to average.

In zone rating, plays are either made or not made.  There are no adjustments for difficulty of chances, which can be problematic over a single season but less so as we move into larger sample sizes.  I’m going to ignore the fact that lumping OF ZR together isn’t necessarily the best approach either, just because I want to ignore it.

Zone rating is a decimal and it’s calculated as plays made divided by chances.  Chances are defined as plays that are converted into outs at least 50% of the time in the assigned zone for each defender.

Since 2008, Gardner has seen 901 chances in the outfield and converted 821 of them into outs.  That’s a ZR of .904.  Since 2008, Granderson has seen 1575 chances and converted 1404 into outs, good for a ZR of .891.

The average AL LF had a ZR of .848 in 2011.  The average AL CF had a ZR of .879.

The primary reason to make the switch is that CF will typically see more chances.  In 2011 Yankee CF saw 444 chances and Yankee LF saw 368.

So using 2008-2011 ZR as an estimate of current ZR, here are the two scenarios.

Player	Pos	Ch	Diff
Granderson	CF	444	5
Gardner	LF	368	23
Total		812	29

Player	Pos	Ch	Diff
Gardner	CF	444	14
Granderson	LF	368	16
Total		812	30

Diff is difference between Granderson/Gardner’s ZR and an average LF/CF over the chances shown.  The difference is in terms of plays, not runs.  A play made/not made in LF is generally worth about 0.83 runs on average, in CF it’s about 0.84 runs.

These tables show little difference.  Part of the reason Gardner has had such good defensive numbers is because he’s being compared to worse defenders.  He’d still likely be quite good in CF, but he wouldn’t be as valuable defensively relative to position, even with more chances.  We can see that moving Granderson to LF makes him more valuable defensively relative to position, again because he’s being compared to worse defenders now.

This assumes that Granderson would be able to make the move to LF seamlessly, which I’m not sure about.  It also ignores the fact that Granderson’s going to be 31 next year and probably in defensive decline, even if last year probably overstates it.  If you were to just use 2011 ZR #s for both (which we shouldn’t do)

Player	Pos	Ch	Diff
Granderson	CF	444	-12
Gardner	LF	368	21
Total		812	9
Player	Pos	Ch	Diff
Gardner	CF	444	12
Granderson	LF	368	1
Total		812	13

If you think using the 2008-2011 data is too generous to Granderson, I could see splitting the difference.  In that case you’re talking about a couple of runs.  I suppose you could probably move Gardner closer to RF with Granderson in LF and gain some additional benefit, but I doubt it’d be much of one

So technically speaking, the Yankees are probably better with Gardner in CF and Granderson in LF, but it likely wouldn’t be noticeable.

And anyone that mentions any player whose name begins with Yu and ends with Darvish in this thread will be put in timeout.