Craig Stammen Returns to the Padres’ Good Bullpen

In December, the free agent signings flowed like wine. You could hardly finish reading about one marquee free agent’s landing spot before another domino fell. Stephen Strasburg, Gerrit Cole, and Anthony Rendon signed on consecutive days, and Edwin Encarnación signed with the White Sox on literal Christmas. It was a hectic month, to say the least.

Of course, taking so many names off the board in December portends a slower January, and that was certainly the case this weekend. Take the Padres. In November, they signed Drew Pomeranz, the sabermetric darling of the offseason, whose relief work in 2019 makes him a potential bullpen ace. In December, they went wild with trades: they added Tommy Pham, Trent Grisham, Zach Davies, and Jurickson Profar to the team over three deals.

In January — well, January has barely started, but so far, they’ve moved down a rung on the urgency ladder. This Saturday, they signed Craig Stammen to a two year, $9 million dollar deal, with a club option for a third. Stammen has been a valuable part of the Padres bullpen over the past two seasons, and he’ll likely continue to appear in high leverage situations for the team, albeit now behind Pomeranz and Kirby Yates rather than just Yates.

You’ve seen relievers like Craig Stammen before. He relies heavily on a sinker that sits in the low 90s, complementing it with a slider and curveball. He survives on a steady diet of grounders — for his career, he has a 1.7 GB/FB ratio, well above league average. He also adds excellent control to the mix — he walked only 4.4% of the batters he faced last year.

Of course, there’s a reason the Padres were able to sign Stammen for $4.5 million a year. While his walk and groundball rates were excellent in 2019, other parts of his game took a step back. His strikeout rate dropped from 27.8% to 21.5%, and that was no fluke. He got swinging strikes on 14% of his pitches in 2018, then only 9.1% in 2019. Batters swung at fewer pitches out of the zone (32% vs. 37.7%) and more pitches in the zone (66.9% vs. 63.8%), and made more contact (79.7% vs 71%) when they did swing.

What drove this deterioration in bat-missing prowess? A little bit of everything. All three of his pitches saw their whiff rate decline. There was no obvious reason this should happen — he actually threw his sinker harder in 2019 than in any previous year, and the slider and curveball looked roughly the same as they did in 2018. He didn’t allow particularly harder contact, either — his average exit velocity allowed declined, and he still allowed a sterling rate of barrels per batted ball.

What does all of this mean for 2020? Steamer projects Stammen to be just okay:

And that makes sense to me. There will always be risk in a 36 year old’s arm, but if he stays healthy, another season of just enough grounders and missed bats to get by is as reasonable of a prediction as any.

With Stammen’s expected 2020 production in hand, the Padres project for the fourth-best bullpen in baseball. That’s nothing new — they were among the league’s best relief units in 2018 and 2019 as well:

Of course, trying to read too much of a pattern into bullpens is a sure path to folly. The Padres had terrible bullpens in 2016 and 2017. And much of 2019’s performance was simply Yates putting in a superhuman year; a 1.19 ERA and 1.30 FIP over 60 innings would make any relief corps look good.

But even behind Yates, the team has excelled in recent years at getting good performances out of formerly journeyman relievers. Yates toiled in anonymity with the Yankees (and briefly the Angels) before finding his groove in San Diego. Stammen himself was indifferent on the Nationals before turning in a stellar 2018 for the Friars. Matt Strahm has been better in San Diego than he was with the Royals, Luis Perdomo went from Rule 5 pick to useful swingman, Brad Hand went from bad to good in San Diego; the list goes on and on.

So if you squint, and if you’re willing to connect a few dots, you could say that the Padres have a talent for constructing bullpens, a skill that should come in handy as the rest of the team takes a step forward into contention. A good bullpen, after all, is most useful to get a team the last few steps of the way to a playoff berth. Attach a dominant bullpen to a 70 win team, and they’re still bad. Attach a dominant bullpen to a 90 win team, and now you’re cooking with gas.

But honestly, that’s not that interesting of a fact. Replace “a dominant bullpen” with “a 6 WAR player” and you’d get the same result. Good players are most useful to good teams, if you’re measuring utility in terms of how much they help you reach the playoffs. Just because they happen to come out of the bullpen doesn’t change the way you should analyze things.

Can we come up with some unique role of bullpens in helping teams overachieve? We can certainly try, though “we can certainly try” is true of a lot of things. As an example, let’s try to find how bullpens show up in terms of teams outperforming their Pythagorean expectations.

First, we calculate every team’s winning percentage and Pythag-expected winning percentage over the past four years. Next, we try to come up with some bullpen factor that is correlated with under- or over-performance. First up is ERA, and nope:

What if we used ERA-, to control for those pesky park factors? Still no — the data still looks like soup, and the r-squared is still de minimis. Regressing against FIP doesn’t promise to be any better, since it’s a step further removed from actual results than ERA, and indeed, FIP and FIP- explain less of the variation in outperformance than even ERA.

What if we’re just missing something obvious? Use WAR as a predictor, Ben! That should account for leverage, due to the leverage multiplier to individual reliever WAR. Picture two teams. One has a 3.00 FIP reliever and a 5.00 FIP reliever; each pitch 60 innings, with the 3.00 FIP reliever pitching in high-leverage situations. Another team has two 4.00 FIP pitchers who each go 60. The two teams will have identical FIP’s (and what the heck, let’s specify that all of these pitchers have ERA’s to match FIP), but the team with two disparate relievers will have a higher WAR. That team will also have better pitching when the game is on the line, so they should theoretically beat their Pythagorean record.

Does WAR help? Kind of. It gets the r-squared all the way up to a whopping 0.01. In other words, you can explain 1% of the variation in how much a team outperforms Pythagorean approximation using bullpen WAR. That’s nothing, a statistically insignificant correlation. And even if you do think that it’s real, the slope is small; adding 1 WAR to your bullpen adds 0.2 wins of outperformance. In other words, if you believe this effect, 1 WAR of relief pitching is “actually” worth roughly 1.2 wins.

So okay, maybe there’s nothing particularly special about the Padres combining a good bullpen with an already good team to achieve some kind of synergy. That doesn’t mean that having a good bullpen isn’t good just on the merits. After all, we’re merely speculating on whether relievers are worth more than a WAR formula would suggest, and relievers already have value in that framework.

To me, the Stammen signing is less about some belief in the intrinsic value of a good bullpen and more about the construction of the Padres roster. With their earlier trade bonanza, the team simply doesn’t have many easy spots to add talent. Catcher? They’re unlikely to move on from a tandem of Austin Hedges and Francisco Mejía, and even if they did, it wouldn’t be cheap to improve on that tandem’s overall production.

Center field is a bit of a hole, but not one they can easily patch; no marginal player is going to be more useful to the team than Franchy Cordero, Manuel Margot, and Wil Myers, the current third through fifth outfielders on the squad. First base is a place where you could improve — Eric Hosmer has totaled -0.5 WAR over the past two years — but the team is unlikely to move on from him and his eight year contract, whether or not that’s logical.

That leaves only pitching, and the team has enough depth in the starting rotation that adding a Homer Bailey type (a rough match for Stammen’s contract) is unappealing. A pitcher like Bailey would probably have slotted in as somewhere between the sixth- and eighth-best pitcher on the team, a swingman to go along with Perdomo and Strahm. There’s some chance that type could transition to relief, but if you ignore that, the team doesn’t really have much choice. If they wanted to make marginal improvements, the only place to do it was the bullpen, where the team has plenty of innings to fill. There doesn’t need to be any magical value to a good bullpen on a good team; the Padres simply got the most bang for their buck at the price point they shopped at by adding to the bullpen.

Of course, just because Stammen was an optimal choice if the team was spending $5 million or so to improve the roster doesn’t mean that they shouldn’t have spent more. Adding Gerrit Cole or Stephen Strasburg would have made the team better. But if you accept that the team’s big moves this offseason have already happened in Pham, Grisham, Profar, Davies, and Pomeranz, then this move is a perfectly logical complement.

It’s not a satisfying conclusion, necessarily. There’s no bigger picture at play in this small potatoes signing. The Padres simply wanted to spend a little money to improve, and they did it in the most logical way. Maybe Stammen will surprise in 2020, and maybe he’ll disappoint. But it’s hard to fault the tactics that led the team to sign him.