Baseball, and more specifically the way fans watch baseball, has changed a good deal over the last 10-15 years thanks to the flood of analytics into the game. This insurgence of numbers and metrics hasn’t just seeped its way into the front offices, but into the general public as well. We can now see definitive data that leads us to view games, player, and situations in a new light. Not all of this data has been well received however, and there are still some very polarizing theories that create fantastic discussions on social media and radio across the country. One of the most divisive of these that has cropped up recently has been the redefining of bullpen roles.
Analytics leads us to believe that the best way to use relief pitchers is to call on your best bullpen arms in the highest leverage situations. Leverage is a metric that was created about 10 years ago based off of work by Keith Woolner in win expectancy. In a nutshell, leverage tells us how important a specific situation is on the overall outcome of a game. Every inning, base, out, score scenario is given a leverage based on win expectancy with a neutral leverage being 1. The higher the leverage, the more critical the situation is to the outcome of the game.
For example, a tie game with two runners on base and no outs in the bottom of the 7th inning would have a higher leverage than a three run lead for the home team with two outs and no baserunners in the top of the 9th inning. In this scenario, your best bullpen pitcher should come in to close out the 7th in an attempt to preserve the tie.
Going back to the advent of the save, and the closer utopia of the 80’s and 90’s, this rocks what has been an established tradition in baseball. While analytics has put the value of a closer under a microscope and shows its flaws, we still have this mainstream idea that relievers need defined roles and any deviation from those roles causes such an upheaval in the pitchers psyche as to render him virtually useless. I see this argument all over social media and I hear it chubbed up weekly on podcasts and radio shows.
To dig into this idea that relievers are so fragile that they can not alter their expected routine in any way, I pulled up every pitcher that threw at least 40 hitters in more than one inning in 2017. Being a Cleveland native I started my search with Andrew Miller, who has become sort of the poster child for the leverage movement ever since his amazing playoff run in 2016.
Obviously Miller has no qualms about mowing hitters down in any situation. He did most of his work in the 8th inning of ball games where opposing hitters slashed a dismal .134/.183/.232 over 121 plate appearances. He 81 batters in the 7th inning, allowing a slash of .155/.250/.155; this is slightly worse than the 8th but still an All-Star performance for Miller. While hitters were able to reach base at a bit of a better rate in the 7th, he allowed zero extra base hits in that inning. Miller is clearly a pitcher that is indiscriminate when it comes to sending hitters back to the bench.
Miller is just a catalyst for this research. Taking the numbers from all of MLB for relievers used in multiple innings, I’ve compiled a chart to show the outcomes by inning. I used tOPS+ to judge the pitcher's performance in each inning. tOPS+ takes a hitters OPS, adjust it for the league, and adjusts it again for the hitters own average. In the final number, 100 is still a league average.
I refer to the inning that a given pitcher saw the most usage as the Main Inning, and his secondary inning as Off Inning. There are some anomalies and outliers like there are in every study, but as a whole relievers in the league don't seem to waiver one bit when it comes to pitching in different innings. In actuality as a group they performed better in their Off Innings, suggesting that possibly leverage has more of an effect on a pitcher than the inning in which they perform. This idea that relievers need to know that they pitch the 7th or the 9th and that those roles need to remain undeviating in order for that reliever to be effective, is just not consistent with reality. Relievers that pitch well, pitch well no matter the inning.