The Rockies’ offseason has been a weird mixture of doing a lot and nothing at the same time. On the one hand, they committed over $50 million dollars for at least six years of service to two relievers. But on the other hand, one of those relievers is a returnee and the other one is an approximation of their 2017 trade deadline acquisition. Because the (in)activity has been on the bullpen, it provides an opportunity to explore one of my favorite baseball truisms: You can’t rely on repeat performances from a collection of relievers whose seasonal activity range from small to smaller sample.
That’s not to say that the Rockies shouldn’t invest in the bullpen. It’s to say that the Rockies shouldn’t only invest in the bullpen.
The best measure of a bullpen’s success is its Win Probability Added (WPA). WPA is a context-based stat that rewards good performance in high-leverage situations, penalizes bad performances in the same situations, and considers any sort of performance in low-leverage situations as essentially neutral. It’s firmly a “what happened” metric.
Because it’s based on context and “what happenedness”, WPA from one season is not a very good predictor of performance in the next. The chart below shows the bullpen WPA from all 30 teams over the past 10 seasons. Find the pattern:
The substantial movement from year to year sticks out most. Even teams on the plus side of WPA fluctuate quite a bit up and down. The pattern is that there is no pattern. You’ll also notice the Rockies moving up and down, but having spent the years prior to 2017 below zero. The dramatic bounce from 2016 to 2017 is a graphical way of showing the team going from the worst bullpen in baseball in one year to the best in the National League the next.
Every reliever who threw at least 20 innings for the Rockies in 2017, with the exceptions of Scott Oberg and Jordan Lyles, ended up on the right side of the WPA ledger. The collection of these players wasn’t accidental. The personnel were the product of shrewd acquisitions, but the Rockies were fortunate to have so many things work out so well from so many people.
Not only were the Rockies the best bullpen in the NL in 2017 by WPA, but they were also one of the 15 best bullpen seasons over the last 10 years—that’s the top 0.05% of 300 team seasons. It makes sense for the Rockies to want to re-create that by getting the band back together. But we can’t ignore how hard it is to replicate such seasons. The following table highlights the Rockies’ 2017 along with the other 14 seasons that compose the top 15 since 2008:
:format(webp):no_upscale()/cdn.vox-cdn.com/uploads/chorus_asset/file/9886395/15_best_BP_WPA_table.png)
Pay attention to the top seasons along with the seasons that come afterward. It makes sense that the 12 bullpen seasons previous to 2017 declined—that comes with the territory of being in the top 0.05%—but it’s surprising just how much some of them fell off. While some bullpens followed up with good seasons, the only team to have led their league in bullpen WPA the following year were the 2012 Braves, and they even trailed four American League teams. The team that fell off the most was the 2013 Orioles, who declined by a whopping 13.37 WPA. Let’s look a little closer at these two teams.
The 2012 Braves had an advantage that, it’s safe to say, the Rockies won’t—one of the best relief seasons of all time. Craig Kimbrel in 2012 struck out half the batters he faced and put up a 1.01 ERA. He also, by himself, posted 4.25 WPA. I’m not one for exclamation points, but: !. They also benefited from Eric O’Flaherty’s 2.69 WPA. The smallest marginal decline among the top 15 bullpen WPA seasons was on account of a historic season coupled with an excellent one.
The 2013 Orioles weren’t quite as mediocre as their WPA suggests. They had solid seasons from Tommie Hunter and Darren O’Day, but everyone else was closer to average. Pedro Strop’s terrible -2.72 WPA in 221⁄3 innings dragged the entire bullpen down. But even without Strop, the 2013 Orioles’ bullpen wouldn’t have come close to the 2012 performance. In that year, Jim Johnson wasn’t spectacular with his stuff but managed to log 5.27 WPA (!!) by himself, mostly on the strength of saving a bunch of close games. O’Day contributed 3.45 WPA of his own, too. A handful of pretty good seasons and the lack of any terrible ones complemented the two big contributions to produce one of the best bullpen seasons baseball has ever seen. It wasn’t even close to being replicated the following year.
The obvious point is that when a team produces one of the best bullpen seasons in a decade, it’s unlikely to reproduce it. The less obvious point is that the Rockies shouldn’t put all their eggs in the bullpen basket, regardless of whether the relievers are new or old. A lot broke right for the team in 2017, and while bringing in Bryan Shaw and re-signing Jake McGee should certainly help, it will be more surprising if the 2018 Rockies’ bullpen matches their 2017 than if they don’t.
That’s fine though. The bullpen doesn’t need to be the best in the National League to compete. The Giants won three championships over the past decade, and they didn’t even come close to registering among the top 15 bullpen seasons in any of those years. The Rockies just need to get a couple really good seasons from the bullpen, a handful of middling ones, and avoid catastrophe elsewhere. They need something else too: support for the bullpen. If the past decade of bullpen seasons teaches us anything, it’s that it’s a gamble to take excellence from one season and assume it can be the supporting mechanism for the next.