As snubs show, WAR won't carry you to Cooperstown

Trevor Hoffman missed the Hall of Fame by just five votes this year. Fellow closer Billy Wagner (above) didn't even come close. Your definition of "snub" might vary, but was Hoffman really THAT much better? Henny Ray Abrams/AP Photo

It has never been less clear what a Hall of Famer is.

Oddly, it has never been more clear what a Hall of Famer did. Wednesday, for instance, we nearly elected Trevor Hoffman, who missed by just five votes. Sixty years ago, if his role as a relief ace had existed at all, his signature statistic -- 601 saves -- certainly didn't, as the save itself was invented only later. That method of carving a baseball game into more specific achievements is what Hoffman's fame depends on.

Today, we know about the saves. We also know about his wins above replacement, a single-number accounting of his career. We know his WAR is higher than Bruce Sutter's or Rollie Fingers'. They're both in the Hall of Fame, so we know Hoffman measures up to, at least, precedent. We know his win probability added -- the measure by which his team improved its chances of winning while he was on the mound -- is the second-highest in baseball history for a reliever.

We also know this:

  • Hoffman: 28.0 Baseball-Reference WAR; 26.1 FanGraphs WAR; second year on the ballot, 74 percent of the vote

  • Billy Wagner: 27.7 bWAR; 24.1 fWAR; second year on the ballot, 10.2 percent

I'm not going to try to convince you Wagner should be in the Hall, or that Hoffman shouldn't. Or even that they're mutually exclusive; if the majority of voters felt Hoffman Yes/Wagner No was a perfectly consistent position, the majority of readers probably will, too, and majorities aren't to be ignored casually.

But over the past five years, something strange has been happening in baseball's left-out club. There have always been "snubs," players whose total-value metrics argued they were good enough but whom voters ignored. These snubs usually fell into the same few categories: Their value came from "quiet" skills like defense or drawing walks; they were overshadowed by a teammate or a similar, superior contemporary; they played positions traditionally underrepresented in the Hall; or they started their careers late, or ended them early, missing out on compiler milestones. Their "snubs" were generally prophesied years in advance, by lackluster MVP finishes or sporadic All-Star selections. These categories described most of baseball's uninducted WAR leaders: Dwight Evans and Graig Nettles, Buddy Bell and Bobby Grich, Alan Trammell and Gene Tenace, Reggie Smith and Willie Randolph.

But this Venn diagram of invisibility seems less relevant to the most recent crop of worthy players looking in, who have given us five paired sets -- including the bullet-pointed one above -- that illustrate how subjectivity has retained its place in Hall of Fame voting even as objectivity has gained strength in baseball writing, analysis, history and team-building.

We're not talking about the omissions of Roger Clemens (the greatest pitcher of the modern era, by WAR) and Barry Bonds (the greatest player of the modern era, by WAR); those guys are being punished for PEDs, simple enough. But how to explain the untainted starting pitcher who outpitched (by WAR) Bob Gibson, or the one who outpitched (by WAR) Bob Feller, or the one who outpitched (by WAR) Jim Palmer? Or the right fielder who outplayed (by WAR) Tony Gwynn and Dave Winfield, or the center fielder who outplayed (by WAR) Duke Snider, or the slugger who outhit (by WAR) Harmon Killebrew?*

As Hall of Fame voting moved into the WAR era, what could easily have turned into an exercise of leaderboard sorting became anything but. Voters remain as willing to ignore WAR as they ever have been.

Consider these four pairs:

Pair 1

Mike Mussina: 82.7 bWAR; 82.2 fWAR; fourth year on the ballot, 51.8 percent
Tom Glavine: 74.0 bWAR; 66.9 fWAR; elected on first ballot (2014)

Pair 2

Edgar Martinez: 68.3 bWAR; 65.5 fWAR; eighth year on the ballot, 58.6 percent
Frank Thomas: 73.7 bWAR; 72.0 fWAR; elected on first ballot (2014)

Pair 3

Larry Walker: 72.6 bWAR; 68.7 fWAR; seventh year on the ballot, 21.9 percent
Vladimir Guerrero: 59.3 bWAR; 54.3 fWAR; 71.7 percent on first ballot

Pair 4

Curt Schilling: 80.7 bWAR; 79.8 fWAR; fifth year on the ballot, 45 percent
John Smoltz: 66.5 bWAR; 79.6 fWAR; elected on first ballot (2015)

Schilling's vote total likely suffered this year from his inflammatory social media posts -- especially those directed at the media -- but he was dangling even before that. Smoltz, meanwhile, was the 50th player in history elected on the first ballot. It's baffling how differently these careers are remembered. There have rarely been two more similar careers up for induction at the same time.

Smoltz was never really the best pitcher in baseball -- he won a Cy Young but never finished higher than fourth in the National League in ERA and never led the league (or, often, his own rotation) in WAR. But his case for the Hall is sturdy, his fine career bolstered both by association (with the Hall of Famers Greg Maddux and Glavine) and his exceptional postseason pitching. Indeed, his postseason win probability added is third in history, and second among starting pitchers.

But the one starter ahead of him is Schilling, who had his own reputation-bolstering associations: With Randy Johnson he formed the most dominant 1-2 punch in postseason history, and with Pedro Martinez and the 2004 Red Sox he broke a curse. He starred in the most famous postseason start since -- what, Don Larsen's perfect game, at least? And at age 40 in 2007 he went 3-0 with a 3.00 ERA in October, pitching Boston to a second championship.

There's little room to differentiate between these two. Schilling had a slightly better adjusted ERA than Smoltz, won 20 games three times and finished second in Cy Young voting three times. Like Smoltz, he was a workhorse who led the league in strikeouts twice and innings twice too, and his black ink overall is more impressive than Smoltz's. Like Smoltz, he won the Roberto Clemente Award, given annually to baseball's kindest and most sportsmanlike player. They retired with almost identical win-loss records. Schilling's most comparable peer, according to Baseball-Reference's player similarity scores, is Smoltz.

There is one difference in these careers: Smoltz spent three years during his prime as a very good closer. In any career retrospective or here's-my-ballot column about Smoltz, the fact that he's the only pitcher with 200 career wins and 150 career saves appears. As that fact is the most substantial difference in these careers, we're left to assume that made the difference in votes.

But what is that fact, and how significant is it? It isn't as though young pitchers all start out aspiring to join the famed 200/150 Club, and only Smoltz got there. It isn't as though those 150 saves means Smoltz was a better pitcher than Maddux, whose career ended with a pathetic zero saves. Surely, Schilling (or Maddux, or most any other Hall of Fame starter) would have had great success pitching one inning in relief. And why is 200/150 more compelling than, say, 215 wins, 20 saves and 3,000 strikeouts -- a club that includes only Schilling and two Hall of Famers? The degree to which 200/150 is meaningful is subjective. A choice is made that it means something. That it means a lot.

What's the case against Walker? That he didn't win an MVP award, like Guerrero? He did. That he didn't have the awe-inspiring arm that Guerrero did? He, and maybe he alone, did. That he couldn't match Guerrero's awesome blend of speed and power? He once hit 49 homers and stole 33 bases in a season, and he finished with more career stolen bases than Guerrero did. That his postseason record wasn't as good as Guerrero's? It was better. That his numbers were inflated by Coors Field? Well ... OK. Accounting for ballpark is a huge part of what WAR does. Measuring each player's offense, by the batting runs used in WAR:

  • Guerrero: 429

  • Walker: 420

To conclude against this that Guerrero was better is fine; it's just subjective. It's a statement that there's something illegitimate about offensive greatness at Coors Field, and that the objective attempts to control for this are insufficient.

There is no case consistent with previous snubs to explain Mussina, who has the 13th-best winning percentage in history and whose 270 wins easily clear the 250 that voters have unofficially set as the magic line. (Only three other pitchers with at least 250 have been rejected, and one is Clemens.)

Nor Edgar Martinez, for that matter. There is almost nothing easier than measuring the value of a bat. Defense is hard to measure, especially in retrospect. Baserunning can be controversial. Positional adjustments aren't household words. Pitching stats get into philosophical questions. But batting? Batting is so easy to measure. Batting is almost all that Edgar Martinez and Frank Thomas did. Thomas was better, to be sure. But the difference between them is far too small to explain their vote totals. Thomas is an inner-circle, first-ballot inductee. Martinez is stuck hoping the internet will adopt him, like it did Bert Blyleven and Tim Raines. Alas, there's competition for that position.

Finally, back to Wagner and Hoffman. I can admit: When I was growing up, watching these players, I had no doubt that Thomas was better than Martinez, that Glavine was better than Mussina, that Guerrero was better than Walker. I might have been wrong, I might have been right, but I knew it in my gut. But it would never have occurred to me that Wagner wasn't at least as good as Hoffman. He made as many All-Star Games. He made more money. And I wouldn't have traded a Billy Wagner card for a Trevor Hoffman card.

Hoffman pitched longer. He threw 186 more innings in his career, but in those innings he allowed 115 more earned runs. That's three extra seasons of a 5.56 ERA on top of Wagner's career, hardly a bullet point for the brochure. But Hoffman once held the all-time saves record, the most significant achievement he can boast above Wagner. That puts voters in the subjective position of weighing such an achievement. One voter might look at such a record and brush it away: It's a record that could only have been held by a closer in the post-1990s reliever era, so it isn't as though Hoffman beat out a century of baseball greats to set it. It's a record he didn't even hold long enough to make it to the Hall of Fame ballot, having ceded it to Mariano Rivera.

Another might look at such a record and weigh it heavily. If not the record holders, who is the Hall of Fame for?

Who is the Hall of Fame for? We can conclude, without getting too far out there, that the Hall of Fame isn't for the best players; and further, that this is the way we all like it.

Eventually, most of the best ballplayers do get there; lose the correlation entirely and it becomes a farce. But over the past decade there has been a rupture between what the Hall of Fame is and what the list of the best ballplayers is. Bonds, Clemens, Rafael Palmeiro and Mark McGwire are part of that. Once we start to allow that two of the five greatest players in the game aren't in the Hall -- for now, at least -- we lose the binding definition of what a vote is even for. Hall voting becomes less an interrogation of a player's stats and more a way of honoring the stories we want to honor.

That this has come, paradoxically, at a time when it's easier than ever to objectively identify the best players might not be a coincidence: With WAR leaderboards just a Google search away, you have less need than ever for a panel of experts to tell you who the best players are. Future fans will have even more objective data. The need voters fill might be shifting, toward something more like a memory bank, for those future fans who didn't see a particular player in person.

This can be very frustrating. (This can be especially frustrating if you're, say, Schilling.) The voters might not want to honor the same stories I want to, and without a basic logical process we all agree to -- e.g. "who is the best player, and how can we decide" -- there's no real room to sway a person to your side, any more than you can convince them to like tomatoes. But it's pretty close to the voting that was laid out at the beginning of this process: no advanced stats, few leaderboards, just a bunch of expert writers chosen to vote their memories and experiences.

Tell players that you're going to throw a ceremony for the top name on a WAR leaderboard, and nobody will show up. Tell them the ceremony is in honor of a person chosen by a collection of writers given only vague instructions on how to make their decisions, and it becomes a signature event on the baseball calendar. This is, kind of, science: Japanese researchers in 2012 asked subjects to learn and repeat a finger-tapping test. One group got personalized compliments; another got only their objective performance results. The group that was complimented responded more, and showed more improvement. Objective measures simply weren't as meaningful to the brain.

The WAR leaderboard isn't going anywhere. It (or its statistical descendant) will play an important role in documenting what happened in hundreds of thousands of baseball games spread out over dozens of cities and multiple eras. But subjective Hall of Fame voting isn't going anywhere, either. It might be as interesting to see what happens to it in the next few decades as it will be to see what happens to the leaderboards -- and to see whether the two merge or diverge.

* In order: Mussina, Schilling, Kevin Brown, Walker, Kenny Lofton, Martinez. All WARs, unless otherwise noted, are the Baseball-Reference model.