Edwin Diaz, Blake Treinen, and the Greatest Reliever Seasons Ever

Reliever performance is volatile, fluky even from year to year. One season, a closer is dominant; the next, he’s just average. Over the past 40 years, there have been 59 relief seasons of at least 3.0 WAR. Only Rob Dibble, Eric Gagne, Rich Gossage, Tom Henke, Kenley Jansen, and Craig Kimbrel have produced seasons of that standard consecutively. By comparison, 10 starting pitchers have exceeded 7.0 WAR in consecutive seasons (67 seasons total), and 10 position players have exceeded 8.0 WAR in consecutive seasons (83 seasons total). Those 59 relief seasons were compiled by 41 different relievers, and three of those seasons are happening right now.

Josh Hader’s second half hasn’t been as good as his first after a forgettable All-Star Game, but with a 1.83 FIP and a 2.08 ERA, Hader is right at 3.0 WAR. In a lot of seasons, a solid finish to the year would make Hader the highest-rated reliever by WAR. This year, however, Hader is solidly in third place behind Edwin Diaz and Blake Treinen.

A year ago, Diaz posted a 4.02 FIP and a 3.27 ERA. That’s not bad, but it’s also not great. Diaz struck out 32% of batters faced, which is quite strong, but he also walked 12% of batters and gave up 10 homers. This season, Diaz is using his slider a bit more to get swings outside of the zone. The results have been staggering: he’s increased his strikeouts by about 50% while decreasing his walks and homers by 50% as well. With a few weeks to go, Diaz has piled up 3.7 WAR thanks to a 1.38 FIP — or 34 FIP- when factoring in league and park, which allows us to compare across eras. Only four relievers have ever put up a FIP- that low: Wade Davis, Gagne, Jansen, and Kimbrel (twice). The increased specialization of the closer role means that those four players all come from the past 20 years. Although Diaz’s 1.95 ERA and 48 ERA- are very good, they are not the best marks in the game. That honor goes to Treinen.

The A’s reliever is part of an improbably successful bullpen that appears to be working miracles and has helped Oakland into playoff position. A year ago, Treinen split his time with the Nationals and A’s, recording a 3.42 FIP and 3.93 ERA. Like Diaz above, this made Treinen a decent, above-average reliever, but nothing close to a relief ace. This season, Treinen has added a 95-mph cutter and another mph to his fastball. Like Diaz, he increased his strikeout rate by about 50% — Treinen’s jump from 23% to 33% isn’t nearly as impressive — while keeping the same walk rate and seeing his home-run rate and BABIP fall precipitously. Treinen is pitching better than last season, but he’s also benefiting from a very good Oakland infield defense and a park that suppresses homers. Both of Treinen’s long balls this season have come on the road.

The combination of Treinen’s good pitching and Oakland’s defense — and, perhaps, a little bit of luck — means that only four relievers have ever posted an ERA- lower than Treinen’s current figure of 21: Zach Britton, Dennis Eckersley, Jonathan Papelbon, and Fernando Rodney. When you combine that run-prevention with a ton of high-leverage situations, you get one of the top Win Probability Added (WPA) seasons of all-time.

I think it’s fair to say that, when we think of great relievers, we consider some combination of performance, results, and situation. To that end, I compiled a list of relievers. I began with a sample of 3,617 relievers total who pitched at least 60 innings since 1871. I looked at the 84 player-seasons that featured a FIP- below 50, the 281 seasons that featured an ERA- below 50, and the 124 seasons where WPA was at least 4.0. There were 52 seasons meeting the ERA and FIP requirements. The lowest WPA was Tom Henke’s 1989 seasons, in which he somehow managed a 0.40 WPA with 24 shutdowns (WPA of at least .06) balanced out by 13 meltdowns (WPA of -.06 or worse). Adding in the WPA requirement left 23 seasons. Only Craig Kimbrel and Mariano Rivera appear on the list more than once, as the table below indicates.

Both Diaz and Treinen have time to add to their totals before the season. If we just wanted a reliever list sorted by WAR, it would look like this:

If we want to incorporate runs and WPA, we can take the list of 23 players above and average WPA, RA9/WAR, and WAR. This might end up double-weighting the run-prevention aspect, but the requirement of 50 FIP- or lower filters out a lot of pitchers to potentially provide some balance. Here’s what those averages look like:

That Bruce Sutter season was something else. The split-finger pioneer had a higher WAR than CY Young-winner Steve Carlton in 1977, though both pitchers finished behind (in order) Steve Rogers, Rick Reuschel, Tom Seaver, Phil Niekro, and Burt Hooton. There might not be a best way to determine top reliever seasons, but the final method used here puts a Hall of Famer, a Cy Young winner, and a future Hall of Famer in the top-three. Diaz and Treinen both have some time to move their way up that particular list, but they are pretty clearly having two of the best seasons we’ve seen in the era of the modern reliever.