As far as I know:
1) The Diamondbacks are still shopping Justin Upton.
2) The Diamondbacks lost a major pitching prospect in Trevor Bauer.
3) The Diamondbacks have a lot of outfielders, and they don’t like Justin Upton very much.
4) The Mariners need a good outfielder.
A major problem, as cited here, in working a trade with Upton is that the Mariners are on his no-trade list. I was once confused about no-trade lists, but I believe I have been straightened out. A no-trade clause, in this case, simply means that the Dbacks must have Upton’s approval before trading him to the Mariners. The clause may have been there because Upton doesn’t like rain, doesn’t like pitchers’ parks, doesn’t like coffee…who knows? But if he’s sick enough of Arizona—and he has heard of the fence transplantation in SafeCo as well and something closer to a temperate summer—then perhaps he’d okay a trade to Seattle. Probably not likely, but worth exploring in the Sodomojo pages, nonetheless.
The deal would likely include a major pitching prospect (to replace Bauer), and something else fun, but that’s not what I’m here to talk about. I’m here to talk about the powerful force known as regression to the mean. Dave Cameron alluded to it in an article over at Fangraphs, but maybe some specific examples can advance the discussion.
The biggest knock on Justin Upton comes from his home/road splits. The guy has been a monster at home in the friendly Phoenix atmosphere (138 wRC+), but he’s essentially been Ty Wigginton on the road (96 wRC+). Over the last ten seasons, that 42 wRC+ difference ranks him third most polarized among all batters with at least 1000 plate appearances both home and away. In other words, Justin Upton is an extreme case when it comes to splits.
Let’s take a break to play the Regression to the Mean Game. Heard of it? I’m going to take the best and worst 10 batters from 2011 in terms of a few different stats, and then see how they fared in 2012. Let’s start with batting average.
|Miguel Cabrera|| |
|Adrian Gonzalez|| |
|Michael Young|| |
|Jose Reyes|| |
|Ryan Braun|| |
|Victor Martinez|| |
|Matt Kemp|| |
|Jacoby Ellsbury|| |
|Hunter Pence|| |
|Joey Votto|| |
Besides Joey Votto, every single player declined, or in the case of Victor Martinez, did not play. The average decline was 9.5%.
Now let’s use wRC+ to check for regression in the opposite direction. In other words, did the worst hitters in 2011 rebound? Qualified players only, of course.
|Name||2011 wRC+||2012 wRC+|
|Alex Rios|| |
|Mark Ellis|| |
|Casey McGehee|| |
|Gordon Beckham|| |
|Alcides Escobar|| |
|Yuniesky Betancourt|| |
|Alex Gonzalez|| |
|Miguel Olivo|| |
|Aaron Hill|| |
|Ichiro Suzuki|| |
Ignoring the fact that two Mariners are on the list (and a third would have been if Figgins had been qualified), again we see a lot of regression—this time in an upward direction. The overall regression back up was 35.1%! This isn’t exactly a biased sample either. There’s not a majority of young players or injured players in the sample, where we would have projected obvious improvement.
There are probably many theories as to why extreme players regress. I’m sure psychology plays some role, for instance. But I believe the best explanation for regression is simply the unsustainable nature of extreme play. Whether you want to call it chance, randomness, or luck (or bad luck), it takes more than one’s skill level to perform extremely, in my opinion.
We notice Upton’s splits because they represent an outlier among baseball players. We can go back and try to explain that it’s caused by his home ball park, or maybe his swing, or that he loathes airplanes, etc. But as with any metric in baseball, it would be foolish to assume that Upton’s stats are representative of such an extreme inability to hit outside Chase Field. Just having a hitter-friendly home ballpark doesn’t explain it all.
We saw that Miguel Cabrera won the batting title in 2011. We know Miguel Cabrera is a good hitter. We should still have expected regression in the following the season.
We see that Upton is among the three most extreme players when it comes to home/away splits. We know Upton should have distinguished splits, playing in Chase. We should still expect regression in the 2013 season.
In Upton’s case specifically, we should probably assume he’s not as bad on the road as his 96 wRC+ suggests, and perhaps he’s not quite as good at home as his 138 wRC+ suggests. If regression is up to something, as usual, then Upton is not likely to move to Seattle and continue hitting at a .250/.325/.406 clip on the road. Maybe in SafeCo, but not on the road.
The trade is unlikely, but I would welcome Upton with open arms. He’s good, he fills a major need in the organization, he’s 25, and he’s signed through 2015 on a team-friendly deal.