Splitting Kendry Morales
First base is Kendry Morales’ job to lose this year, so sabermetricians are spinning their statistical heads trying to prove he’s the next big bust.
Sam Miller at the Orange County Register, who’s doing quite a nice job breathing fresh air into the paper’s Angels coverage, posted a blog entry on February 25 reporting that a formula concocted by Baseball Prospectus writer Nate Silver called PECOTA predicts that in 2009 Morales will post an AVG/OBP/SLG of .253/.295/.389.
Part of the denigration, Sam writes, is due to PECOTA trying to adjust for Triple-A Salt Lake’s hitting friendly environment.
What PECOTA knows is that Morales has spent most of his minor league career in extreme hitting environments. It adjusts his minor league numbers to Major League Equivalent numbers. Going from Salt Lake City (elevation: 4226), in the Pacific Coast League, to the majors strips Morales’ numbers of most of their shine.
For instance, his .341/.376/.543 Salt Lake line of 2008 translates to .262/.297/.422 in the majors. The first line looks like a star; the second line doesn’t deserve a starting job in the majors. Giving credibility to PECOTA’s translations, Morales has hit a very similar .249/.302/.408 in 307 real major league at bats, which gives us more reason to be suspicious.
To quote Shakespeare, “Ay, there’s the rub.”
If you’re a regular reader of my blogs, you know I’m hard on statheads, and this is one reason why. They look for macro-trends and then apply them in individual examples, which is the lazy way out. What they should be doing is analysis of the individual player.
I’ve written many articles in the last couple years about how to properly analyze Salt Lake players. You don’t just look at home/road split, or the nonsensical Major League Equivalent formula.
In the PCL, there are five hitter-friendly parks — Salt Lake, Albuquerque, Colorado Springs, Las Vegas, and Tucson. So what you do is look at a player’s performance in those five parks versus the rest of the league.
It requires a little bit of effort, namely collecting the game-by-game data in a year for a player, then sorting them into “hitter-friendly” versus the rest.
I spent a half-hour last night looking at Kendry’s splits for 2007 and 2008. Here’s what I found (AVG/OBP/SLG):
2007 (Age 24):
OVERALL: .341/.385/.486 (255 AB)
HITTER-FRIENDLY: .359/.406/.503 (181 AB)
THE REST: .297/.342/.446 (74 AB)
2008 (Age 25):
OVERALL: .341/.376/.543 (317 AB)
HITTER-FRIENDLY: .323/.362/.494 (164 AB)
THE REST: .359/.404/.595 (153 AB)
2007 is what you’d expect, better numbers in the hitter-friendly parks. But in 2008, Morales had much better numbers in the neutral/pitcher-friendly parks!
I checked those 2008 numbers looking for any mistake I may have made, but it looks legitimate. And the sample size is about the same for both groups (164 vs. 153 AB).
So does this mean that Morales is a future Hall of Famer?
Of course not.
But what it does show is that off-the-shelf statistical formulas designed for fantasy leagues are no substitute for context.
If Morales had posted that .359/.404/.595 line in a neutral park instead of Salt Lake, he might not have been dismissed so easily by the calculator crowd.
The Register‘s Angels blog is becoming quite the place to be. Sam and the paper’s other sports writers participate all day long, interacting with readers. Click Here to read the blog and join in the fun.