You won’t find the next Fernando Rodney simply with statistics.
Perhaps Rodney’s first 10 innings or his first 100 pitches — or some other relatively early sample — contained something to suggest he’d be able to maintain his incredible quality throughout an entire season. But by that point it was too late: By making him closer, the Rays guaranteed Rodney would fly off the waiver wire.
Many — like me — tried to get a jump on the Rays’ closer situation following Kyle Farnsworth’s injury and picked up Joel Peralta. There were plenty of reasons to think Peralta would get the nod. He saved six games for the Rays in 2011. He struck out a batter per inning in 2010, and he was near that mark in the two surrounding seasons. His ERA- and FIP- were each less than 90 in both 2010 and 2011. He was a quality reliever and — at least looking at recent performance — he was Joe Maddon’s seemingly best choice after Farnsworth.
Instead, the Rays liked what they saw from their little tweaks with Rodney’s delivery. Obviously they were right. Rodney set the major league ERA record at 0.60, posted a sharp 2.13 FIP, struck out more than a batter per inning for the first time since 2008 and saved 48 games.
But Rodney is the exception, not the rule. He’s not a guideline. As Nate Silver would put it, he is the Noise, not the Signal.
If it were easy to find a signal with respect to these relievers, we wouldn’t have articles like these every year. (I wrote one looking at shutdown percentage and big save gainers for last year’s FG+.) We wouldn’t have things like “Closer News,” which is a real website.
But let’s try and cut through the noise. Which statistics and characteristics are shared among relievers who save 10 games or fewer one season but add 20 or more saves the next year?
I’ve selected seven reliever stats to analyze: ERA-, FIP-, K%, BB%, HR%, WAR and fastball velocity. Let’s see which ones hold at least some key to finding precious waiver-wire or late-draft saves.
What follows are frequency graphs for each statistic. Bar height shows how many relievers with 0-10 saves achieved a certain mark in the staistic. Parts of the bar colored in FanGraphs Green™ denote a reliever who added at least 20 saves the next season. If we see most of the green on one side of average, we’re probably onto something.
It’s not that these are poor statistics to look at. Many of them go far in determining the actual quality of relief pitchers — something that’s important when a player gets a chance to close games out. But we’ve seen plenty of mediocre relievers rack up saves in a closer’s role, and most fantasy players will take a hit on rates to add some saves, especially in deeper leagues. These stats simply don’t add much value in determining who gets an opportunity:
No surprise here. Given seasons only spanning between 50 and 80 innings, one season of reliever ERA typically isn’t helpful in even determining the next season’s ERA. It helps to have an above-average ERA-, to be sure, but not in a statistically meaningful way for determining saves for the following year.
Many relievers not only manage but thrive with high walk rates. Without other information, BB% doesn’t tell us much about reliever quality, much less a pitcher’s likelihood to close.
Although pitchers with legitimate home run problems are unlikely to get a shot at closing, there’s enough volatility in home run rates as to give us essentially zero information here. Our closers-to-be aren’t any different from the average reliever.
This one might be a bit of a surprise, but the counting nature of WAR likely drags it’s usefulness down. In many cases, we see players like John Axford or Tom Wilhelmsen come up in the second half of the season and dominate hitters, but because they don’t pitch many major-league innings, their WAR won’t match up. That’s why we see such a wide spread here.
When we reduce WAR to a rate stat, we effectively get FIP- for FanGraphs’s implementation. Unsurprisingly, this moves us closer to our holy grail, but not all the way there. Still, 11 of the 37 pitchers who gained 20-plus saves were below the reliever average in FIP-. It’s better than WAR, but not by a significant margin.
So far, the search has been mostly fruitless. But our two remaining statistics offer the light we’re looking for.
K% illustrates what this search is all about. The point isn’t simply to find above-average relievers — anybody with a FanGraphs leaderboard or even your fantasy service’s basic player add screen can do that with a simple sort. Instead, it’s about the characteristics unique to the pitchers whom managers select as closers. Debate whether or not closer mentality is a thing all you want, but managers think there’s a special “closer class” of relievers. As fantasy players, we need to be more interested in this perception than the capital-t Truth.
I think Scott Downs would make a fine closer. He’s posted an ERA below 3.15 for six straight seasons. He has an 80 ERA- in 12 seasons despite being awful-to-mediocre for the first five. He was setting up Jordan Walden before his implosion in 2012, and he earned a few save chances (he finished with nine saves).
But the job went to Ernesto Frieri, a waiver claim who had a couple of solid seasons in San Diego but looked like the prototypical Petco Park mirage. He posted ground ball rates below 25% in each of his first two full seasons, but HR/FB rates under 5% allowed him post sharp ERA and FIP marks. But his xFIPs were mediocre-to-bad — 3.76 in 2010, 4.34 in 2011 — leaving legitimate questions about his ability to thrive in a more neutral park such as Anaheim’s.
Frieri gets strikeouts, though: about 30% of batters he faced from 2010 to 2011, and well more than a batter per inning. Of the 37 new closers in our sample, 29 had a K% greater than the reliever average.
More importantly, Frieri throws gas. He averaged 94.2 mph on his fastball with the Angels and has never averaged lower than 92 mph on it for his career. It took nearly all of 2012 for Sergio Romo to ascend to the closer role in San Francisco because he’s a junkballer, not because he wasn’t talented. Part of being in the closer class means having a blazing fastball, and it shows in the data.
The average reliever fastball comes in at 91.8 mph. Of our 37 non-closers to pick up 20 saves the next season, all but five featured at least equal velocity. Six had fastballs faster than 95 mph, on average. Perhaps here’s where we could have found Fernando Rodney, but his poor strikeout rates (under 17.5% in 2010 and 2011) muddled the picture.
The new closers averaged 93.5 mph on their fastballs, as a group. This is the basis of the profile to search for when adding relief talent at the end of your drafts or on the waiver wire as closer jobs get shaky: gas and strikeouts.
Who fits the bill? Some won’t become clear until after spring training, but this group stands out:
— David Robertsons (92.2 FBv, 32.7 K%): Mariano Rivera is back for the Yankees, but with Rafael Soriano out of town, Robertson should have the clear path to the closer’s role should Rivera — coming off his first real major injury — not be quite his old self.
— Jake McGee (95.7 FBv, 34.4 K%): Could Joel Peralta get skipped in Tampa again? McGee has always been considered a future closer since his time in the minors, and he has the stuff.
— Carter Capps (98.3 FBv, 25.7 K%): Between Capps, Tom Wilhelmsen and Stephen Pryor, the Seattle bullpen was sneakily one of the most fun parts of 2012. Capps gave up a .357 BABIP in his debut season but was otherwise excellent.
— Jim Henderson (95.0 FBv, 34.4 K%): Henderson briefly earned closer opportunities in Milwaukee last season. He’s had a similar career arc to John Axford.
— Trevor Rosenthal (97.6 FBv, 28.1 K%): The Cardinals practically bleed power arms. Rosenthal put together a solid 22.2 inning performance to finish 2012 in his MLB debut.
— Rex Brothers (95.3 FBv, 28.1 K%): Colorado is a difficult environment for any pitcher, and Brothers’s struggles last year should be forgiven. After all, he still finished with an 86 ERA- and 74 FIP-. Also: most hard-boiled name in MLB history?
Many — most — of these players won’t close in 2013; a good number probably won’t even record a save. But the beauty of selecting relievers in this fashion — particularly in the standard 5×5 game — is you can’t lose. Worst case scenario: You get a solid relief pitcher capable of picking up some vulture wins, loads of strikeouts and perhaps solid rates.
Again, stats alone won’t find you the next Fernando Rodney — only a sharp eye on the news and some luck can do that. But if you want to be ahead of the curve, target velocity and strikeouts. Only good things can happen.
Graphs made with Tableau Public
Jack Moore's work can be seen at VICE Sports and anywhere else you're willing to pay him to write. Buy his e-book.