The Kansas City Royals beat the odds, across the board, for 2015 — and they did it with a full month to spare.
By September 1, the Royals sat atop the American League Central with a record of 80–50. The second-place Twins trailed them by a whopping 13 games, meaning the division race was essentially over. But the same metrics now predicting a K.C. cakewalk were far more cruel to the Royals just a few short months ago. Despite capturing its first American League pennant in 29 years last fall, preseason projections had the Royals—with very few exceptions — falling short of the .500 marker. Baseball Prospectus’s PECOTA system, invented by the statistician Nate Silver, foresaw a 72-win campaign. The Royals surpassed that mark by mid August — a feat the team’s Twitter faithful were all too eager to note.
Fangraphs was slightly more optimistic, but even its 79–83 projection looks silly in retrospect, since it, too, already has been rendered moot. On Sept. 8, Fangraphs predicted a final record of 95–67 for the Royals — a 16-win improvement from March.
And Kansas City wasn’t the only team to shatter the projection machine this season. The Boston Red Sox, for example, were favored to win the American League East, with the New York Yankees and Toronto Blue Jays expected to hover around .500. Instead, it’s the latter two which find themselves racing to win the division, while the former clings fast to the cellar. The Seattle Mariners were tabbed as preseason favorites in the American League West; as of Sunday, they were 69-75, one of the worst records in the American League. The Washington Nationals? They were odds-on darlings in the NL East, but now are looking up at the rampaging New York Mets, and are a near lock to miss the playoffs.
Speaking to PECOTA’s whiffs, Baseball Prospectus editor-in-chief Sam Miller explained to the Kansas City Star that he “felt apprehensive about the [Kansas City] prediction,” postulating as to why the Royals, specifically, might have caused such a blind spot in the predictive formulas: “Bottom line: I think we know that this system has a hard time with elite, highly leveraged bullpens, which the Royals have had; that it probably has a hard time with great defenses, which the Royals have; and that it has a hard time with any team that can outperform its raw stats for whatever reason, which the Royals have done.”
Miller’s first hypothesized system failure — elite bullpens — is thought-provoking. Once upon a time, relief pitchers were considered afterthoughts, one-dimensional guys who simply weren’t good or versatile enough to start. Today, we know that an effective late-inning specialist can be far more valuable than a mediocre starter, and our attitudes towards pitching staff construction have adjusted accordingly. Given PECOTA’s recent shortcomings, however, are we still undervaluing the elite relief pitcher?
The most common, and useful, sabermetric stat used to evaluate individual players has been Wins Above Replacement. By Fangraphs’ measurements, the Indians' Cody Allen and the Yankees’ Dellin Betances are the most valuable individual relievers in all of baseball, each registering a 2.3 WAR. By that measure, there is not a single relief pitcher among the 100 most valuable players in baseball. Moreover, Paul Goldschmidt, of the sub-.500 Arizona Diamondbacks, is considered more valuable than any whole bullpen. This isn’t meant to be a knock on Goldschmidt, who fully deserves to be placed among baseball’s elite, but that’s the point — so do elite bullpens! And it is a failing of the WAR metric — across the board, as measured by every single stat house — that positional value is measured so poorly.
Relief pitchers make a briefer impact on the game than any other position, or so the traditional logic goes. Betances, whose comparatively heavy innings load helped yield his inflated WAR total, is one of only three relievers to have topped 70 innings pitched. Still, that’s less than half the workload of a starter, and so his WAR sags. In the elite Kansas City bullpen, nobody has thrown more than 61.2 innings.
But when you look at team bullpen stats across Major League Baseball, a few of important trends emerge:
Great bullpens make their entire team better.
Of the top 11 bullpens in baseball, eight belong to likely playoff teams. Is this to suggest that an elite relief pitcher is more valuable than, say, an elite shortstop? Hardly. But it’s become increasingly clear that, given the game’s financial-statistical landscape, late-inning pitchers are getting the short end of the stick. Particularly in light of the bottom of the table, where it becomes apparent that ...
Bad bullpens and tragic collapse go hand in hand.
Three of the five teams bringing up the rear for bullpen performance are among 2015's cautionary tales. The Seattle Mariners and Boston Red Sox were both preseason favorites to win their divisions. The Tigers were slight underdogs to Cleveland, but still touted better than one-in-three odds. These three teams are now among the worst in the American League.
The Red Sox, in particular, stand out as a textbook example of what happens when a roster construction project neglects the importance of a bullpen. And while Japanese imports Koji Uehara and Junichi Tazawa have both been quite good, Boston doesn’t tout a single other reliever definitively better than replacement level. The "best" of the bunch, Robbie Ross, has a 0.2 WAR, and eight different Red Sox pitchers have registered 10 innings or more in relief while falling short of replacement level. Craig Breslow and Alexi Ogando have somehow managed to log 111 innings combined, even though it is apparent that neither of them belongs on a Major League roster at this point. Problem is, the alternatives have been no better.
Thanks in large part to their shorthanded relief corps, the Red Sox pulled a reverse Astros and went from first to worst in 2015 — and they did so after spending approximately $50 million combined on Hanley Ramirez, Pablo Sandoval, and Rick Porcello. Both Ramirez and Sandoval have served largely as vestigial appendages in 2015, while Porcello has failed to actualize the potential he showed during his tenure with the Tigers. Indeed, it’s mind-blowing to ponder how far that money could have gone had it been put, at least in part, towards assembling a competent bullpen, because ...
High-end relief arms are a rare bargain (for now).
There are 100 Major League players making a salary of $10 million or more this season, and only two of them — David Robertson of the White Sox and the injured Joe Nathan of the Tigers — are relief pitchers. Both barely qualify, too, making exactly $10 million for the season. Uehara, the Red Sox’ closer and the fourth-highest paid reliever in all of baseball, was paid a lower salary than seven of his teammates this past season.
In recent years, pursuing a fully stacked bullpen has been an affordable enough approach even for the budget-conscious. For all its recent success, Kansas City remains very much a small-market team. But even now, in the midst of yet another competitive run, the team’s payroll still falls outside the top 10. Greg Holland was lights-out in 2014, yet the Royals — by avoiding arbitration with him in the off-season — were able to sign him for a comparatively pithy $8.5 million. Meanwhile, the extension for Wade Davis, inked back when he was still a starter, only bumped his pay up to $7 million, including a club option Kansas City quickly picked up. Combined, the Royals’ eighth and ninth-inning guys — their most expensive relievers by a long shot —will make about as much this season as Melvin Upton. The rest of their bullpen, combined, earned just under $12 million this year.
Since 2007, baseball’s strategic pendulum has swung from an offense-dominant dynamic all the way back to a lower-scoring pitcher’s game. For starting pitchers, the advanced metrics — and the salaries — have more or less adjusted. (See: Clayton Kershaw, the third-most valuable player in baseball according to WAR, and the first $32 million man in U.S. team sports.)
But for these and many other bullpen bulwarks, neither of these adjustments have come to fruition—not yet, anyway. For all the talk of sabermetrics becoming more nimble and nuanced, they have yet to capture or properly measure the contributions of relievers. If anything, pinning down their precise value is as much a crapshoot as it’s ever been. Still, sabermetrics remains an evolving science. With any luck, the failed prognostications of 2015 will only serve to refine our approach in the years to come.
The financial discrepancy might take longer to fully sort itself out, as teams will always be hesitant to roll their free-agency dice — and the many dollars behind them — on players who make up, individually at least, a comparably slim wedge of the workload.
But this off-season, there’s an enticing slate of 30-something relievers set to hit the market. History tells us that they’ll all be available for short-term, high-value deals. Perhaps the Major League franchises currently in to the doldrums will realize what almost all of this year’s postseason hopefuls have in common: a deep, dynamic bullpen helping lead the way.