clock menu more-arrow no yes

Filed under:

The reinvention of the Rockies bullpen

New, 17 comments

According to context-based stats, the Rockies have one of the best bullpens in baseball

About a year ago, I wrote the following paragraph about the 2016 Rockies bullpen:

The 2016 Rockies’ bullpen has been terrible by every measure. Everything from the context neutral ERA to the context reliant Win Probability Added (WPA) shows that the committee has been one of the worst in baseball in 2016, as well as one of the worst in recent memory. Bryan Kilpatrick covered the Rockies’ woes in terms of WPA recently (and also defined the statistic for those unfamiliar). The bullpen’s WPA is the most damning because it shows how bad the relievers have performed in high leverage situations, which in turn highlights how many winnable games the bullpen has cost the Rockies. That, and their current WPA is the 13th worst mark since 2002; that’s also pretty damning.

In the article itself, I showed that it wouldn’t be unprecedented if the Rockies turned around and had a great bullpen. The gold standard turnaround belonged to the 2008 Tampa Bay Rays, who posted a league-best 9.26 WPA a year after having a league-worst -8.61 WPA in 2007. Those figures reflect a bullpen that excelled in high-leverage situations one year after being awful in them. The 2016 Rockies ended up with the worst WPA in baseball in 2016, -5.82.

Here’s that same paragraph, updated for 2017, even accounting for this past Sunday’s collapse:

The 2016 2017 Rockies’ bullpen has been terrible excellent by every measure. Everything from the context neutral ERA to the context reliant Win Probability Added (WPA) shows that the committee has been one of the worst best in baseball in 2016 2017, as well as one of the worst the best in team history recent memory. Bryan Kilpatrick covered the Rockies’ woes year-old troubles in terms of WPA recently (and also defined the statistic for those unfamiliar). The bullpen’s WPA is the most damning joy-brining because it shows how bad great the relievers have performed in high leverage situations, which in turn highlights how many winnable close games the bullpen has cost secured for the Rockies. That, and their current WPA is the 13th worst 21st best mark since 2002; that’s also pretty damning swell.

The Rockies have gone about their resurgence a little differently than the 2008 Rays. The Rays mostly had big seasons from returning members of the bullpen. JP Howell, Grant Balfour, and Seth McClung were the bullpen’s WPA leaders, and they were also a part of the terrible 2007 ‘pen. Troy Percival, the player the Rays signed to be their closer, was effective but not among the best members of the bullpen.

If the Rockies were going to have a truly resurgent bullpen, I expected Chris Rusin, Adam Ottavino, and Carlos Estévez to be the major reasons why. While Rusin has lived up to that hope—he’s been the team’s best reliever—Estévez has fallen into the Pervical-esque “effective but not great” category, and he’s done it while shuffling back and forth from Triple-A to the majors. Ottavino’s been flat-out bad. Besides Rusin, the best returning member of the bullpen has been Jake McGee, who looked broken in 2016.

Instead, the Rockies turned their bullpen from the worst in baseball to one of the best through acquisition. Greg Holland and Mike Dunn have the second and third most WPA on the team, after Rusin, and Pat Neshek has the seventh most despite only being on the team since late July.

Rockies WPA and Clutch leaders, 2017

Player WPA Clutch
Player WPA Clutch
Chris Rusin 2.88 1.14
Mike Dunn 1.65 1.50
Greg Holland 1.54 -0.25
Jake McGee 1.08 0.19
Carlos Estévez 0.67 1.29
Pat Neshek 0.63 -0.01
Scott Oberg 0.18 -0.18
Antonio Senzatela 0.12 -0.55
Adam Ottavino 0.05 0.62
Tyler Chatwood 0.01 0.13
Chad Qualls -0.23 -0.45
Jordan Lyles -0.40 0.55

Through Sunday’s game, the Rockies’ 8.30 bullpen WPA ranks second in baseball, behind Cleveland’s 8.56. Regarding Clutch, the Rockies’ bullpen is the best in baseball, and it’s not really close. The 5.99 mark is well ahead of the Reds’ 4.64. It’s also worth pointing out a couple of stats that deserve more attention than they get, Shutdowns (SD) and Meltdowns (MD).

Think of SDs and MDs as WPA approximations of saves and blown saves. A SD is earned if a reliever increases his team’s odds of winning by six percent (at least 0.06 WPA) or more in a given outing, and he gets an MD if he decreases his team’s chance of winning by six percent (-0.06 WPA) or more. For instance, a closer getting a save with a three run lead wouldn’t get a SD because when a team leads by three runs in the ninth the win expectancy is about 97 percent. Similarly, if a reliever gives up a grand slam—say, to a pinch-hitter named Pat Valaika—in the seventh inning of a 12-0 game, that decreases his team’s chance of winning by less than 0.01 percent. That’s not a MD.

As a team, the Rockies have the third most SDs in baseball and are tied for the second fewest MDs. Because those metrics are WPA-based, and because the Rockies have the second best bullpen WPA in baseball, it shouldn’t be surprising. They’re just a couple more ways of looking at how successful the Rockies bullpen has been.

The Rockies were bound to improve in 2017 because it couldn’t be much worse, but that doesn’t necessarily mean it was destined to be good. That only happened because of the personnel. Heading into the offseason, it would have been perfectly reasonable for the Rockies to not do anything with the bullpen. Ottavino looked like a shutdown closer, Rusin looked like a guy who would just go get as many outs in relief as needed, and it wasn’t hard to see the potential in young fireballers like Estévez and Jairo Díaz. Instead, the Rockies gambled on the return of Holland, shelled out for Dunn, and added Neshek before the deadline. Things have gone spectacularly well for the most part.

Unfortunately, the future always demands attention too. Context-based metrics like WPA and Clutch are only predictive insofar as the extremes from one year are unlikely to be extremes next year. The Rockies, from 2017-2018, have experienced extremes from both ends. So, in other words, the 2018 bullpen will be bound to be worse because it can hardly be better. The calculus of who’s in, who’s out, and how the pieces all fit together will determine how the paragraph we started with will be revised a year from now. Where next year’s strikethroughs land will tell the next story.

★ ★ ★

All stats through Sunday’s game.