Bullpen Threat Level: Orange(ish)

The Cardinals bullpen has taken a lot of flack lately, and rightly so. Over the last two series spanning six games against the Reds and Nats, they’ve been cooked like a Christmas ham to the tune of 18 runs in 25.1 innings. It’s no fluke, either. They’ve yielded a .389 weighted on-base average (wOBA) over those six games, tied with the Nats for worst in the league. Their walk rate (BB%) is 14.9%, dead last in MLB over those games. Their FIP is second to last, surpassing only the Nats. It was a dreadful week for the bullpen. This comes on the heels of so many national writers summoning the regression monster on the rejuvenated second half Cardinals. The expectations for bullpen regression showed up around the time the team removed Shildt’s interim tag, with lots of folks (including our own A.E. Schafer) accurately pointing out that the second half bullpen was outperforming their peripherals. It’s a cause for concern. Here are their bullpen ranks in several key categories:

I split it out across three easy to find designations on Fangraphs- the last month, the last two weeks, and the second half. They are struggling in a lot of ways, and it doesn’t matter which time frame we use. They aren’t striking many hitters out. They’re walking a lot of hitters, and the combination of strikeouts and walks (K-BB%) is the worst in all of baseball. They’re in the bottom ten in all three time periods in getting ahead of hitters (1st strike %), swinging strikes (SwStrk%), and chase percentage outside the strike zone (O-Swing%). When they allow contact, they’re allowing loud contact, posting a hard hit percentage in the bottom ten league-wide in all three time frames. Add it all up and you have one of the worst bullpen FIPs since mid-July.

Somehow, they’ve managed to collect a reasonable win probability added (WPA) until their meltdown over the last few weeks. That’s partially because they’ve done better in high leverage situations. Their hard hit percentage in high leverage situations rises to 16th in the 2nd half (up from 20th in all situations). Their strikeout percentage also improves slightly in high leverage, from 29th to 23rd. Additionally, Mike Shildt has generally done a good job of using appropriate relievers based on leverage, which will help improve WPA. Even with optimal bullpen use and some clutch pitching, it’s hard to ignore the red flags in that chart.

Not all Contact is the Same

As we’ve seen, their inability to miss bats and low strikeout rate mean they have a contact-heavy profile. Usually, that’s a recipe for trouble. It’s a recipe for volatility, with wild fluctuations in quality at the absolute worst possible times. But not all contact-heavy bullpens are created the same. Jordan Hicks and Dakota Hudson have thrown 24.3% of all Cardinal bullpen innings in the second half, with Hicks at 20.2 innings and Hudson at 19. Those are the two largest inning totals out of the bullpen. Those are contact-heavy pitchers, to be sure. They’re also pitchers who can bully hitters inside the strike zone with high octane sinkers.

In the second half, Statcast has recorded 1,017 sinkers league-wide at 96+ miles per hour. Hicks alone is responsible for 257 (25.27%) of them. Hudson has thrown another 65, sixth most in baseball, good for another 6.39%. In other words, 31.66% of all 96+ mph sinkers thrown in the second half have come out of the Cardinals bullpen. The Cardinals’ bullpen is responsible for more 96+ mph sinkers in the 2nd half than 27 other MLB teams combined.

Being able to put the bat on the ball is a great way to score off of a bullpen, but not when the best you can do is hammer it into the ground. For some perspective on the difference between a normal sinker and the types of power sinkers employed by Hicks and Hudson, here’s a simple line graph of wOBA on all sinkers by velocity.

Most of Hudson’s sinkers are in the 96-97 mph range, while Hicks is throwing sinkers in the 98+ range. Once you get into the 93+ range, the wOBA allowed dips below league average. The Cardinals’ bullpen has thrown 433 sinkers at 93 or higher in the 2nd half, which is 93.5% of their second half sinkers. It’s also 6.3% of all pitches thrown. Their wOBA allowed on those sinkers is .308.

In theory, this should lead to better underlying numbers for the Cardinals bullpen. Specifically, it should lead to opposing hitters producing weaker contact. Let’s take a look at their rankings for Statcast data and soft hit percentage (note: soft % in this table is via Fangraphs). This includes both the 2nd half, and the last few weeks.

It’s a mixed bag. The good news is that the overall 2nd half performance shows a bullpen keeping launch angles down, suppressing exit velocity, and giving up a decent amount of soft contact. The bad news is that, over the last few weeks, the soft contact percentage has cratered while exit velocity has skyrocketed. They’re getting hitters to drive the ball into the ground more, but it’s cold comfort.

Who are the culprits for the recent slide?

Let’s figure out which pitchers are causing this decline by parsing out the difference in performance when we compare the recent two weeks with the rest of the second half (All-Star break through 8/22). Of second half relievers, we know we can eliminate the following pitchers who began the 2nd half in the bullpen but have not pitched in relief over the last two weeks: Austin Gomber (1.2 IP in the 2nd half), Daniel Poncedeleon (10.2 IP before moving to the rotation), Greg Holland (1.1 IP before he was released), John Gant (1.0 IP), Luke Gregerson (3.1 IP), Matt Bowman (2.2 IP), Sam Tuivailala (3.2 IP), and Tyler Lyons (2.0 IP). That group collectively gave them -0.4 fWAR before moving out of the bullpen, so it’s not as if their absence is causing the slide.

Here’s a table showing the difference in FIP and soft contact percentage over the last few weeks for the remaining individuals:

Obviously, we’re dealing with small samples, which will yield extreme gaps. Still, there are some clear culprits. Hudson, Hicks, and Webb have all regressed to varying degrees in FIP. Some- but not all- of that has been balanced out by the presence(s) of Carlos Martinez, Luke Weaver, and a few acceptable performances from John Brebbia. The regression group has thrown 11.2 innings, while Weaver, Brebbia, and Martínez have thrown 10 over the last few weeks.

Complicating matters is the fact that Hicks and Hudson had settled in as trustworthy higher-leverage options. Their slide has had greater impact than it would for a generic low-leverage reliever. Speaking of trustworthy high-leverage options suddenly declining, the truly terrifying collapse in that table is from Bud Norris. The 16.35 gain in his FIP over the last few weeks is not a typo. Brett Cecil also qualifies to a lesser degree.

They’ve lost Mike Mayers to the disabled list, but replaced him with Dominic Leone. Leone had a hiccup Tuesday night when he gave up two singles and a double. However, he has yet to walk anyone or give up any homeruns in his 5 innings since coming off the disabled list. His presence has been helpful, particularly in the type of metrics that enhance a team’s FIP. Ross and Shreve (along with Webb) aren’t getting as much soft contact, but it hasn’t hurt their productivity compared to where they were earlier in the 2nd half. They are generic relievers with defined roles, ideally in low leverage.

It doesn’t show up in the table, but it’s worth noting that this rocky two week bullpen stretch overlaps with the injuries to Kolten Wong and Jedd Gyorko. Both injuries, especially the injury to Wong, open up the possibility for more groundballs in the infield to sneak through. Moreover, the injury to Wong limits the possibility of using Yairo Muñoz as a defensive replacement for José Martínez. Those injuries increase the chance for an already fragile bullpen to get BABIP’ed to death.

Conclusions

It would be easy to look at the alarming peripherals earlier in the second half, then the recent two week stretch, and assume the regression monster attacked. That said, the nasty peripherals were likely overstated to a small degree because of Hudson, Hicks, and their ability to force weaker contact. As Craig Edwards pointed out a few weeks ago, Shildt’s choice of defense (full-time for Wong and Bader, limiting Jose Martínez, more time for Gyorko at third base) helped them beat their peripherals a bit. That’s not to say that the Cardinals are the one magical team in MLB history that can sustainably beat their FIP by a run. I certainly won’t dispute that their peripherals foretold regression. I also think it’s fair to suggest that the potential for regression has been overstated a little bit. We’ve seen regression sting them, but I’m not convinced it’s the only factor.

Other factors:

• the sudden slide from Bud Norris
• the return of Brett Cecil and his rocky early season performance
• the loss of Mike Mayers to injury
• the Wong and Gyorko injuries

It all combines into a rich mélange of frustrating late inning performances. It’s worthy of your concern. It’s also worthy of reasonable hope that they can find a middle ground between their early 2nd half performance and the current goat rodeo. The Cardinals aren’t in a code red yet, but it’s a hue in that general vicinity.