August 2, 2013
Reliever Value at the Deadline
Baseball lore preaches that a team “can never have enough pitching,” but we rarely hear the same thing said about hitting, perhaps because of a sister proverb, “Baseball is 75 percent pitching.” Pitcher fragility plays a big part, of course. But sometimes one poor start or relief outing will cause a team to press for more pitchers: a marginal arm blows up, and suddenly the team needs assistance. When a hitter goes 0-for-4 or 1-for-5, on the other hand, the line is common enough that we don’t bat an eye.
The “never have enough” axiom seems doubly true for bullpens. Scoreless relief innings are so common that when a reliever allows several runs, it becomes a referendum on the state of the bullpen: we need help, because this shouldn’t be happening. We should be seeing that unblemished 1 0 0 0 0 1 box score line every night. But variance can’t be beaten.
Despite the abolishment of draft-pick compensation, trades for relievers haven’t become less common (I’ll show this later). Pedro Strop, Matt Thornton, Francisco Rodriguez, Jesse Crain, Scott Downs, Jose Veras, Marc Rzepzynski, and Joe Thatcher comprise 2013’s reliever trade deadline class. Seven of those players went to contending teams. Whether the reliever filled a crucial role or offered peace of mind as depth, teams were willing to part with a prospect to acquire him.
Nonetheless, most teams that featured attractive relievers in their display windows found them new homes. Consider the Houston Astros: for two straight seasons, they’ve declared a closer at the start of the season and shipped him off by July 31st. Brett Myers procured three prospects for Jeff Luhnow, and Jose Veras netted another.
I wondered how much value these relievers ended up producing, and whether their worth to their new teams could be forecasted. I gathered every July trade since 1995 and filtered for relievers, whom I defined as having over 10 innings pitched with both their old and new teams. I also eliminated pitchers who had more starter innings than relief innings with their new team. Of that reliever set, I took only trades in which the acquiring team was over .500 on the day of the deal to select for contending teams.
This left me with 114 relievers. Pre-trade PWARP per player had a modest correlation (R=0.16) to post-trade PWARP (p=0.083, statistical significance at your discretion). So even for contending teams, which presumably would be trading for good relievers, PWARP with the old team isn’t a meaningful indicator. Relievers tend to be very volatile in small samples.
Under similar parameters (minimum 25 PA with both teams), hitters’ pre-trade BWARP was a significant predictor of post-trade BWARP (R=0.39). This obviously follows from conventional wisdom about batter and pitcher variability.
Here are the number of relievers who were dealt to contending teams each year:
These relievers accrued 28.5 PWARP in their new homes. A quick-and-dirty division by number of relievers—which doesn’t control for talent or usage—says that these relievers averaged a quarter win, plus any playoff value. Not game-changing, but not insignificant. Still, there’s always the risk of hurting the team with a below-replacement value performance. Twenty relievers with positive pre-trade PWARPs accumulated a negative PWARP with their new team.
As I mentioned earlier, seven relievers went to contending teams in 2013, no different from in the past two seasons. Teams won’t receive any draft-pick compensation for these relievers, so if they’re in the last year of their contract (which goes for Crain, Rodriguez, and Downs), they’re merely two-month rentals.
An interesting question is whether contending teams pay a premium for a hitter’s increased reliability. Given a batter and reliever of equal value, is the caliber of prospect traded away different? These continued transactions suggest that teams believe that relievers are worth the prospects it takes to get them. And they might be—prospects are unpredictable too.
The Braves had a clear role for Downs to fill, Veras made the back of Detroit’s bullpen more stable, and in Crain, the Rays stockpiled some late-inning talent. One wonders, though, whether the output of an acquired reliever could be replicated by an in-house promotion or by stretching someone else’s one-inning relief outing to two. If so, contending teams might be overpaying to stock their bullpens with name brands.