Chris Young and Minute Maid Park

Chris Young, the former Diamondback and current A's outfielder, is a Houston prep star who seems to inflict a lot of damage on the Astros. Assuming the A's don't pick up Young's $11 million club option--a fairly likely outcome, given his poor season last year--some TCB commenters have suggested that Young might be a good free agent target for the Astros.

Young has amassed good batting stats against the Astros--despite a decline in his overall performance over the last couple of years. Some Astros fans wonder whether he might continue the level of offense that led to his moniker as an "Astros Killer" if he changed into the Astros' uniform. This article isn't aimed as a "pro" or "con" article on the merits of signing Young. Instead, I want to explore whether Young's domination of the Astros should carry any special meaning or predictiveness if he were to join the Astros.

CHRIS YOUNG VS. THE ASTROS

In 220 career plate appearances against the Astros, Young has an excellent batting line:

.325, .389, .590, .979

His career OPS against the Astros is 60% higher than his overall OPS against everybody. But I just don't see how a player's record against an opposing team, in itself, should have much predictive power for his performance if he joined that team. If he gets more motivated against his hometown team, that won't be a factor if the Astros are no longer the opposition. If he bats well against the Astros because the Astros' pitching and defense has been poor, that also won't be a factor, since he would no longer benefit from hitting against he Astros.

If a player has performed particularly well against a specific opposing team, it's human nature for the opposing teams and its fans to over value the player. Perhaps that entered into the Oakland A's decision to trade for Young last year. Surprisingly, Young's batting line against the Astros is not his best performance against an opposing team. Young's career slash line against the Oakland A's is .400, .419, .725, 1.144. Then the A's signed him last year, and he promptly responded with a .200, .280, .379, .659 batting mark for the A's.

Regardless whether it entered into the A's decision making, the disparity between Young's performance for and against the A's is a cautionary tale about the predictiveness of small sample splits versus a particular team.

However, this isn't the end of the story. Perhaps Young's elevated performance against the Astros is really due to Minute Maid Park. If Young matches up well with the ballpark, this could be a causal link.

CHRIS YOUNG VS. MINUTE MAID PARK

Chris Young's career batting line at Minute Maid Park is Ruthian. In 107 plate appearances at MMP, Young has put up the following slash line:

.392, .430, .732, 1.162

If we knew he could repeat that line in the future at MMP, there would be no question as to whether the Astros should sign Young. But there's is a phrase that sabermetricians use frequently that prevents us from making that assumption...

SAMPLE SIZE

107 plate appearance isn't a large enough sample size to predict much of anything. Do you remember a first baseman named Chris Shelton? In his first 105 plate appearances of 2006, he hit 10 HRs, posting a line of .326, .404, .783, 1.186. He hit 6 HRs in his remaining 300 or so plate appearances that year, producing an OPS around.660 for the remaining season. He became a career minor leaguer.

We could just give up at this point, and chalk up Young's dominance at MMP to a small sample size aberration. But let's move forward.

DOES YOUNG MATCH UP WITH MINUTE MAID DIMENSIONS?

MMP definitely has features which can give an advantage to certain hitters. The most prominent feature is the short LF porch--the Crawford Boxes.

Chris Young is a dead pull hitter to LF. This ESPN HR tracker scatterplot shows Young's 2013 HRs overlaid on MMP. Young hit a 401 foot HR into the Crawford Boxes, which you can watch...if you can stand to re-live that game.

Young has a 206 OPS+ at MMP, and the ballpark is accompanied by other Young favorites--- Fenway Park with a 170 OPS+, Kauffman Park with 130, Rogers Center at 224, Coors Field at 140, and Wrigley Field at 130. With the exception of the Royal home park, all of these ballparks have above average HR park factors for RH hitters. Fenway and MMP are similar ballparks featuring a short LF distance and a high wall.

Young's MMP stats may be a small sample size, but there is reason to believe that Young is well suited to exploit the LF advantage at MMP. Young may have an advantage in MMP, but that doesn't mean his 206 OPS+ is repeatable.

PROJECTING YOUNG'S MMP PERFORMANCE

Young's LF power may play well in MMP, but the small sample size of his performance there limits our ability to project his future performance if he played in MMP regularly. Faced with this problem, I will turn to an unusual projection tool, Fielding Independent Offense (FIO), which is akin to a hitting version of the FIP and x-FIP pitching stats. The FIO formula is explained in this Fangraphs article.

The primary variables in FIO are BB rate, K rate, HR rate, and BABIP., which together produce a .95 R-squared for predicting wRC+.(weighted Runs Created). That means FIO explains 95% of the variation in wRC+. FIO is a simpler form of wRC+, but it does a pretty good job of approximating wRC+. FIO can be used as a tool for eliciting information from small samples. It also allows us to examine the relationships between variables which stabilize quickly (BB, K), a variable which stabilizes over larger samples (HR), and a variable which is volatile (BABIP).

The FIO formula is shown below:

FIO = -51.57 + 275.21(BB%) – 180.52(K%) +
1184.34(HR%) + 151.75(SB/PA) + 422.14(BABIP)

To begin this exercise, let's plug Young's MMP stats into the FIP formula.

		Young Career At Minute Maid Park
BB%		K%		HR%		SB%		BABIP
7.5%		17.8%		7.5%		2.8%		0.417
FIO Estimated wRC+		205.8

As we can see from the FIO result, the formula accurately estimates his wRC+. Young has a 206 OPS+ at MMP, and FIO produces an almost identical wRC+.

By testing variables in the formula, we can deduce that Young's elevated performance is driven by a sky high BABIP and a high HR rate. Since the .417 BABIP at MMP is unsustainable, we can re-run the formula with Chris Young's career BABIP, .275.

		Adjust BABIP
BB%		K%		HR%		SB%		BABIP
7.5%		17.8%		7.5%		2.8%		0.275
FIO Estimate wRC+		145.8

Decreasing the BABIP to Young's career average reduces the FIO estimate of wRC+ to 146. 146 is still a very high wRC+. (As a comparison, Jeff Bagwell has a career 149 wRC+.) At this point, it is clear that the HR rate--roughly double Young's career rate-- is driving the elevated performance. And this also puts the small sample size effect into perspective. The MMP HR rate is based on 8 homers in 107 plate appearances. Randomly subtracting a couple of HRs would have a significant effect on the result.

Perhaps we can estimate Young's expected wRC+ at MMP by adjusting the HR%. I will adjust Young's overall career HR rate upward to reflect the impact of MMP's LF. I averaged six years of ESPN's HR park factors for MMP to arrive at an overall HR park factor of 1.104. I then used this Hardball Times article by Greg Rybarczyk to arrive at a ratio between MMP's overall park factor and MMP's LF park factor. Applying the ratio to the ESPN overall HR park factor resulted in a LF park factor of 1.25, which can be multiplied by Young's career average HR rate.

The new FIO estimate for Young's MMP offense, adjusted for MMP's LF effect is shown below.

		Adjust HR Rate
BB%		K%		HR%		SB%		BABIP
7.5%		17.8%		4.5%		2.8%		0.275
FIO Estimate wRC+		111.1

The estimated wRC+ for Young in MMP is an above average 111---and also more believable. I should note that the results, above, are sensitive to the strike out rate. Young's actual BB rate and K rate in MMP are both less than his career averages. If his career rate for BBs and Ks is inserted in the FIO,, the wRC+ declines, as shown below.

		Career BB and K Rates
BB%		K%		HR%		SB%		BABIP
10.0%		22.9%		4.5%		2.8%		0.275
FIO Estimate wRC+		108.7

So, I'll put Young's estimated wRC+ for MMP somewhere between 108 and 111. And that's pretty good. Justin Maxwell had a similar wRC+ (112) at MMP in 2013.

Of course, Young's overall performance will include road games too. Plug Young's career averages across the board into FIO and the result is a wRC+ of 98, which I will utilize as the estimated offense for road games. Arguably this estimate is on the optimistic side, given Young's decline last year. But it is also the same as Steamer's estimate for Young in 2014, which seemingly confirms the FIO result.

Combining the MMP and Road wRC+, above, results in a 103 - 105 wRC+ for Young's hypothetical 2013 season with the Astros. That would certainly be a substantial improvement over Young's 2012 season. And that's no surprise, since Young's offense is likely to regress upward next season. That's probably good for a .725 - .740 OPS, and is in the neighborhood of outfielders like John Jay, Michael Brantley, and Josh Hamilton.

I can't guarantee this projection for Young as a member of the Houston Astros. I think there is a good deal of risk in any projections for Young. But I think it's a reasonable estimate.

As I said, I'm not staking out a position in either direction advocating that the Astros sign Young. But I have addressed the "Minute Maid Effect" for Young. Does it surprise you or change your mind about Young?