World Series time! Enjoy Premium-level access to most features through the end of the Series!
December 28, 2012
In A Pickle
When you do an article search on this site for the phrase "Silver Slugger," you get 57 results, the first of which, by Gregg Pearlman, is apparently the 18th article ever written for Baseball Prospectus and the most recent of which is Geoff Young's piece about the Padres throwing their heft around the N.L. West this offseason. (That's #19056.) Young's Silver Slugger mention came because Jason Marquis won one. Pearlman was writing about Barry Bonds. (Or really about sportswriters' relationship with Bonds. This was October 1997. We were innocent once, and young.)
By contrast, when you search "Gold Glove," you see just a smidge over nine times the results. (The first of which, hilariously, is another Gregg Pearlman article -- this one includes a lamentation of the J.T. Snow trade—which is numbered "1" in our content system.)
Now that I've mentioned both phrases in this piece, I'm not actually righting the imbalance, but since I'm going to actually discuss the Silver Slugger but not the Gold Glove, at least the scales will be tipped a tad bit more in favor of the batting award in terms of mindshare and Q Score.
So. The Silver Slugger awards are decided the same way that Gold Gloves are: coaches and managers vote and they're not allowed to write in someone who plays on their own team. The website describing the award actually mentions batting average, on-base, and slugging, which is interesting, and says nothing about RBI, which is heartening, but then also includes the "general impressions of a player's overall offensive value" as a criterion, which is pretty vague.
Two things I can't tell, though I'll take your input, Dear Reader, if you have thoughts on the matter: (a) whether the awards are meant to value level of production as long as the player achieved a reasonable amount of playing time (as opposed to a cumulative-offensive-value approach) and (b) whether baserunning is considered part of "offensive value." Glancing at the history of the American League winners and noting that Rickey Henderson was out-Sluggered in his career by such players as Gary Sheffield, Albert Belle, and Juan Gonzalez, I think it's fair to suppose that baserunning is not at the fore of voters' minds when they mark their ballots. (Then again, maybe it shouldn't be—our Baserunning Runs shows Henderson being 170 runs above average for his career, compared to 598 Batting Runs Above Average. For context, those 170 runs are the most in our database and it's not particularly close (Henderson doubles up the ninth-best runner, Paul Molitor, which is astounding), yet Henderson's batting dwarfs his historic running in terms of contributions to his overall value as an offensive player.)
The intriguing thing to me about Silver Slugger voting is that we as outsiders have a much better handle on offensive value than we do on defense and pitching. Without downplaying the amount of research there still is to be done on hitting and hitters, the question of offensive value in the sabermetric realm has essentially been settled for decades—as Jim Albert and Jay Bennett describe in their book Curve Ball, George Lindsey published what we would now describe as linear weights for offensive events all the way back in 1963. While it took a while longer for researchers like Pete Palmer to fully develop the methods (and add elements like park factors), when we talk about True Average or Weighted On-Base Average, we're talking about offensive value metrics that, despite the "new-fangled" and "new-age" and "WAR WHAT IS IT GOOD FOR" descriptors with which they're saddled, were discovered/invented before Ruben Amaro was born and solidified around the time Billy Beane was struggling to hit in Double-A.
Given all of this, even if, say, Bob Melvin doesn't mark his Silver Slugger choices by consulting our humble website, the ideas of what makes good offense have probably seeped into the baseball establishment to a fair degree. (Contrast this with, say, evaluating pitchers by strikeouts, homers, and walks alone, which, I mean, who can blame teams for not buying up pitchers by fWAR when a million nerds on Twitter can't even agree on whether that's the right approach.) It also doesn't hurt that, while the Moneyball A's as a team had stellar pitching, Moneyball as a book was in no small part about offense and the characteristics of good offensive players.
All of this is a roundabout way of posing a hypothesis that I won't fully test in this piece but into which I will make a brief examination. That hypothesis is that Silver Slugger voting will conform better to sabermetric understandings than other awards do. (And in particular than the Gold Gloves, though as to the knotty question of what "sabermetric understandings" of defense are in a world where some metrics have significant uncertainty and others have significant data problems (and uncertainty), let's just wave our hands and mumble something about how voters value looking smooth in the field over the nerd ideal, which involves actually making more plays than one's competitors.) Like I said, no big formal or even informal tests are coming in this article, but here's a review of the 2012 awards with some BP metrics (True Average (the hitting-value rate stat), Batting Runs Above Average (the hitting-value cumulative stat), VORP (the total-value cumulative stat that includes everything—including position—but defense)), league ranks therein (minimum 502 PA), and commentary on the general state of things as far as we can tell from the winners:
Posey was a slam-dunk. It doesn't take much balls or sabermetric cred to give the catcher Silver Slugger to the guy with a .408 OBP and .549 SLG. Pierzynski, on the other hand, is not a slam-dunk on the pure numbers, but likely benefited significantly from the fact that Joe Mauer got half his starts at first or designated hitter, and Carlos Santana played 48 games at those spots himself. Pierzynski was easily the best hitter among "pure catchers," with the next-best being Matt Wieters.
Fielder was nearly as dead easy a choice as Posey, with the second-best American League first baseman being Albert Pujols, who helped matters along by not only having a year that was clearly secondary to the large Tiger's, but doing it in a way that aroused the Narrative Machine to whisper about him being kaput as a superstar player. One suspects that had the two tied in TAv by splitting the difference and finishing at .317, Fielder would have won anyway, but one's suspicions aren't fair to the voters—the whole point here is to figure things out, not to cast the same old aspersions.
LaRoche is thornier. He was 3rd in TAv, but Allen Craig was one of the players ahead of him, and Craig started only 86 games at first (and played 119 overall), so this is where cumulative vs. rate value comes in: Craig outhit LaRoche in his time on the field, but not by enough to make up for the at-bats he lost to injury, as LaRoche beat him by about three runs of BRAA.
The leader in BRAA was Paul Goldschmidt, but he topped LaRoche by less than half a run, which is hardly a difference about which we can have supreme confidence. Goldschmidt, as it happens, has been an above-average baserunner for his career, while LaRoche has not. In 2012, that difference contributed to Goldschmidt finishing almost seven runs above LaRoche in VORP. That LaRoche won the Silver Slugger anyway could be evidence that baserunning doesn't matter, evidence that Goldschmidt's baserunning is underappreciated, or evidence of nothing at all. (Speaking of the Narrative Dragon, LaRoche having a comeback season, his best since 2009, on a suddenly nationally prominent team may have played a bigger role than any of the individual differences between the top candidates.)
Cano was a legit MVP candidate and his closest challenger, Dustin Pedroia, finished 37 points of TAv behind him, with that difference reflected equally as well in their raw slash lines. Cano also had more plate appearances than any second baseman but Ian Kinsler (who had a down year with the bat).
While Hill didn't pull away from the pack the way Cano did, the gap between his basic stats and Neil Walker's are blindingly obvious. These picks should not have been close. I'd even bet that Cano was unanimous (excepting the Yankee manager/coach votes for reasons of ineligibility. I hope they voted for Jemile Weeks in protest).
Jeter vs. Erick Aybar is a tough call for a sabermetrician, but not tough at all for someone who's just going to look at the basic stats. The Yankee not only won by 186 PA, he also beat Aybar's OPS by 51 points. Where Aybar pulls himself into contention is the massive difference in offensive environment: our personal park factors for Aybar and Jeter were 93 and 104, respectively. That does not represent the biggest gap among positional peers (that would be Dexter Fowler's 115 to Cameron Maybin's 92), but that's still a sizeable gap, one that makes Aybar a defensible choice to you or me, especially if we're rewarding rate over accumulation, but might not to someone who doesn't realize that the A.L. West is simply a brutal place for a hitter to play ball.
Desmond, as you might guess from the rankings listed above, had playing-time issues. He destroyed Jimmy Rollins (the VORP leader) in TAv and edged Jose Reyes (the BRAA leader), but was topped by more than 150 plate appearances by each. Desmond's .511 slugging (compared to Reyes's .433 figure or Rollins at .427) may have won the day. Overvaluing slugging as compared to on-base ability or other aspects might be considered a hallmark of non-sabermetric thinking. Just for what it's worth.
While both Cabrera and Headley were the "right" choices, it should be noted that the gap from them to their nearest competitors was not like Cano-to-Pedroia: Cabrera beat Adrian Beltre by 16 points of TAv, not nearly 40, and Headley topped David Wright by 12 points. These are still large enough differences to qualify, in the context of top major leaguers, as running away with it, but voting for Wright would not have looked insane, especially since he actually out-OPS'd the Padre by nine points. The difference is made up and then some by, as you'd expect, Petco Park. In contrast to Anaheim's stadium, however, Petco's run-suppression ways appear to have penetrated mainstream thought, so Headley likely benefited from an upward mental adjustment even if the voters were not applying mathematical changes to his numbers. (To me, the key distinction these days between sabermetric thought and non- on park factors is whether they're quantified and, relatedly, whether non-obvious parks are recognized as being outside of the norm. Anaheim, as has been noted many places, does not look particularly cavernous or feature very obviously ball-suppressing weather, for instance.)
Procedural issues! First, based on Hamilton-Trout-Willingham (i.e. no right fielder) as well as some past results, I believe the voting for Silver Slugger is as the Gold Glove used to be: three outfielders from any position, not a left, center, and right, despite the fact that a left, center, and right is exactly what we ended up with in the National League. Second, I've not counted Ben Zobrist as an outfielder here. I don't know what to count him as, but I suspect the voters don't know, either. Third, I've omitted VORP from the discussion because among the outfielders, a substantial difference is going to be on the positional adjustment—comparing Jay Bruce (we'll get to this in a minute) to Michael Bourn leaves Bourn dominating Bruce in VORP. Most of this is baserunning, actually, but a not-insignificant part is the difference between center field and right field, which difference should not be relevant if the question is hitting (or offensive value) by an outfielder. This question obviously does not arise between two full-time shortstops, which is why I felt fine using it in the previous sections.
Substantive issues! Braun, Trout, and McCutchen were obvious choices, MVP candidates all. As with Cano or Cabrera or Fielder above, we're not going to be able to distinguish between an enlightened electorate and a Dark Ages one based on these choices. Josh Willingham was not an MVP candidate, but his raw stats look just as obviously deserving as his advanced ones, so he fits into the same category for these purposes.
On the other hand, Josh Hamilton prevented Yoenis Cespedes from taking home the shiny bat that should rightfully have been his by virtue of finishing ninth in baseball in TAv and second among A.L. outfielders. Park effects probably come into play again (Cespedes's personal factor was 95, Hamilton's 109), though there's always a lingering possibility of rookie bias, a "he'll get his if he keeps earning it" mentality. There are a variety of possible explanations, as always, but none are terribly flattering.
But I've saved the most egregious for last: What on earth is with Jay Bruce's award? The Reds right fielder had essentially an identical line to Jason Kubel, of all people. I wonder if Kubel got any votes. What really jumps out from Bruce's line is the magic "5" that begins his slugging percentage. Outside of McCutchen and Braun, the players ahead of Bruce by TAv topped out at a .497 slugging (Matt Holliday). The tougher parks (Angel Pagan, Andre Ethier), significantly higher on-base percentage (Holliday), hell, the better batting average (everybody) —none of this apparently mattered. Jay Bruce gets hardware and Jay Bruce had a pretty slugging percentage. I'd like to think that something else went on, that there's some secret value the voters are attempting to suss out, but you just can't justify an outfielder with a .327 on-base percentage winning an offense-only award with an explanation other than a fetish for power.
(It has occurred to me that some voters might take "slugger" in the name of the award overly literally, but given that the description, which presumably reflects the voting instructions, explicitly mentions on-base percentage, I'd say that "overly literal definition of 'slugger'" and "overly sexual feelings toward power" are just symptoms of the same disease here.)
Tell you what, the results above do not suggest what I expected they would suggest: there appears to be less of a focus on the things stat-savvy baseball fans appreciate (on-base ability, park adjustments) and more of an emphasis on dongers than I anticipated/hoped. It isn't clear, from this review, that the voting for Silver Slugger is any closer to our understanding of the offensive game than the Gold Gloves are to our ideas of defense. It's just one year, sure, but the Cespedes and Holliday omissions (and to a lesser extent those of Aybar at shortstop and maybe Goldschmidt at first) have shaken my confidence that the concepts of offense we by now take for granted are being appreciated by the men on the field.