The Astros defense isn't that bad, with two brief interludes

Nothing confounds me more than the Astros' defense.

Nothing.

Are they going to be an average defensive team? Are they going to be terrible? Have they been terrible for years and my eyes have been deceiving me? Are the numbers fair or unfair to them?

I think the Astros will be a decent defensive team next year, especially in the outfield.

Is that just because of a fan's optimism? Certainly, I've never lacked that. Heck, I'm optimistic about teams that I don't actively root for. Probably why I'm not the best sportswriter in the world, but let's get back on track.

Everyone expects the Astros to be terrible defensively, but I don't expect that to be the case.

I just can't prove it.

What we know

We know that the eyes lie.

We know that defensive metrics have come a long way in the last two decades.

We know that they're still contentious.

We know that any defensive stats take much longer to stabilize than other stats.

We know that the Astros have plenty of players projected to be bad defensively in 2015.

Yet, we know they've been very good at doing things that the metrics don't judge well.

What they say

The projections, though, are pretty dire. On FanGraphs' depth chart, only one everyday regular in Houston's lineup is projected to have positive defensive value. That guy? Jake Marisnick, who very well might lose that time in left field to Gattis. Or to Bob Grossman, or Preston Tucker or whoever.

That's why everyone assumes the Astros will be a bad defensive team. They've been one, on paper, for the past three seasons.

In 2012, Houston had negative-70 Defensive Runs Saved.

In 2013, the Astros had negative-43 DRS.

In 2014, they had negative-16.

Not once have they ranked in the top half of team defense under Luhnow.

That's all despite getting a ton of support from the shift, which many popular defensive metrics have problems with.

Photo by Dale Zanine-USA TODAY Sports

Evan Gattis

A great test case for the Astros dissonance on defense is Evan Gattis. A poor performing catcher, the Braves had already set in motion a plan to transition him to left field before Gattis was traded.

The baseball community generally took a dim view of this, thinking Gattis would perform horribly in the outfield.

How much of an impact will he have?

Well, that all depends on how much he plays in the outfield. Let's assume Gattis will play in 125 games next year. That's about 20 more than he's ever played in the majors before, so we're going out on a limb a little.

If he gets there, he'd play about five games a week. Give him one start at catcher and one at DH. That means he's only spending about 60 percent of his time in left field. Throw in a start at first base and that number drops to two games a week, but that's not a given. Let's stick with three starts in the outfield.

Now let's say the Astros strategically deploy those three starts with pitchers who have high ground ball rates or against lineups more prone to hit the ball to left field. Let's say they focus on getting him those starts in the outfield at Minute Maid Park, where his limitations may be, er, limited.

What will his defensive impact be then? I don't know the answer and I don't think the advanced metrics do, either. I suspect it would be bad, but not terrible.

The shift

What we do know is that the Astros have shifted and shifted and shifted some more under Luhnow. The team shifted more than any other team in the league last season.

Sometimes, the shift has phenomenal results.The Pirates dramatically improved their defense because of the shift. The Astros have shifted a bunch, but have faced more pushback than they may want.

Some defensive metrics account fort shift plays. Ultimate Zone Rating (UZR) does not, which makes the Astros look worse in comparison to reality. Their defensive WAR totals on FanGraphs are depressed thanks to that. But, not all defensive metrics ignore the shift.

The ones that do show the Astros added more value from the shift than almost any other team in baseball. Add that to their other defensive numbers and those Astros from 2014 look like a decidedly average defensive squad.

The 2015 squad will feature a possible defensive upgrade in the outfield, but the infield may have gotten worse. Luis Valbuena is at least Matt Dominguez' equal defensively, but had a worse defensive season than Matty D did last year. Jed Lowrie may bring some pop to shortstop, but is materially worse in the field than the players Houston trotted out at short in 2014.

Photo by Steven Bisig-USA TODAY Sports

Jose Altuve

Another player who will be on that infield and who is routinely diminished by defensive metrics is Jose Altuve.

We know why.

It's the eyes.

A player that small shouldn't be able to play adequate defense. Heck, we even debated this on the podcast. "What about those line drive that are hit over his head, out of his reach?" we said.

I can't tell you how many conversations I've had with old-school baseball fans to this end. Altuve doesn't look like a great fielder, so he must not be. The defensive metrics say he's not a good fielder, so he must not be.

Except, he kind of is.

Two points I'll bring up here. The first is this FanGraphs piece on plays made. The Inside Edge data lets us see which players make the plays they should make. Call this the "Miguel Tejada Corrollary," because when Tejada was in Houston, he excelled at this very thing. He wasn't the flashiest shortstop, but he gobbled up anything that he got to, turning it into an out.

In this article by Jeff Zimmerman at FanGraphs, we find that Altuve is actually four percent better than league average defensively on plays made. His defensive metrics have been average every other year, but have been abysmal in 2014 and 2012.

But, there's an even simpler measure to see how good a fielder Altuve is. Bill James has been talking about defense all month, really, and Rob Neyer has been summarizing that work.

In this one, he talks about middle infielders turning double plays. Jose Altuve led the majors in double plays turned in each of the last two seasons.

He may not have the same kind of range as an "average" shortstop, but he makes plays and he turns double plays.

What to do?

Altuve sort of exemplifies the Astros problems with quantifying their defense. Our eyes don't tell us the whole story. The numbers don't tell us the whole story. How do we know what's going on?

We don't. We have to guess. And that should make you feel a little queasy. We make these overarching assumptions on player value, but we can't accurately say how to rate defense, 15 years or so after the problem was first broached.

Maybe the Field F/X or whatever it's going to be called will solve this. By tracking player movement on the field, we can judge things in components, rather than trying to come up with one number to rule them all for defense.

What I keep coming back to is a fundamental question about these Astros. They hired a defensive whiz in Colin Wyers. They have the Trackman data. They shift more than anyone else and have success doing so.

Yet, they continue to nab below-average defensive players and attempt to play them every day.

Is this a case where the numbers don't tell us what's actually going on? Do the Astros have metrics that show their defense is actually quite good and we cannot see them? Can they hide defensive deficiencies with shifts and positioning?

Or, are they simply buying what they can, imperfect players who have defensive flaws?

The answer is probably all of the above.

It makes thinking about the Astros defense tough and makes projecting their team performance even tougher. Should we believe the BaseRuns model, that this is a 78-win team right now? Should we believe ZiPS projections that only one projected Astros starter will be a good defensive player?

Unlocking those questions could be the key to figuring out these Astros next season. Your guess is as good as mine.