The Rays—surprisingly, considering all their injuries—are in first place in their division, and people on the national scale have started to notice. Jeff Sullivan wrote an article at FanGraphs the other day investigating why the Rays are winning. Among other factors, Sullivan discussed the most interesting component of their success and one we've been debating since the first week of the season when Kevin Cash showed a willingness to pull his pitchers before most fans are used to seeing. The Rays starters, despite having lost a ton of quality arms, are being very effective when they pitch, but they have pitched the fewest innings in baseball
The idea behind the strategy is not new. The best work on it was done by MGL, which you can read here. As batters see a pitcher more times, they gain a better understanding of his stuff and approach and hit him harder. So a team can gain an advantage by limiting the exposure of their starters and moving innings to their bullpen, rather than let the starter face a team for the third and fourth time through the order.
But what is actually gained by this strategy and what is lost is a much more complex question than that. When you force the bullpen to pitch more innings, they too become exposed. And as they tire, they should theoretically pitch worse as the season wears on.
How should a manager weigh improving his chances of winning a game today with hurting his chances to win tomorrow? A week from now? In several months? Ted Berg wrote about how the Rays are working their relievers over at For the Win, and that's worth a read, but friend-of-the-site Jason Hanselman just completed the best and most detailed breakdown of Rays bullpen usage. His article now becomes jumping off point for any discussions of the merits and demerits of the Rays early-hook strategy.
His first look is at who's carried the load within the Rays 'pen:
Anything in the positive tells us that that guy has pitched more than the average and the more positive the further he is from the norm. Same in the other direction for the negative values. In this example Kevin Jepsen leads the Rays relievers in pitches thrown and is around 1.7 standard deviations from the norm. He checks in well below Justin De Fratus who leads all MLB pitchers in relief pitches thrown with 654, which equates to 2.4 standard deviations.
Note that the biggest load is being carried by the top guys (although Jake McGee obscured things a little bit by arriving late to the party after rehabbing from offseason surgery). So Jason next looks at how the use of the top pitchers compares to other teams.
What becomes clear is that, while the Rays are working their bullpen hard, they're not actually working any particular relief pitchers to an outrageous extent. Instead, their strategy of shuttling young pitchers with options back and forth from triple-A Durham is keeping those innings spread around more guys than the 25-man roster allows them to carry.
There are more visualizations in Jason's peice breaking down the Rays' implementation of the bullpen strategy, and some of them are very clever, so click over and read his full article. The basis for understanding a practice and arguing over it has to come from the facts of what that practice actually means, and Jason has taken us significantly closer to being able to intelligently discuss the early hook and the Rays success.