Going into the season, the Reds bullpen looked like the second coming of the Nasty Boys. It was not hard to count backwards from the ninth inning: Aroldis Chapman to save, Sam LeCure in the 8th, Jonathan Broxton or Alfredo Simon in the seventh, and Sean Marshall ready for the occasional lefty. The plan was perfect. If only that plan would have made it out of spring training.

The injuries piled up quickly: Chapman came down with a bad case of ball-to-the-forehead, Sean Marshall’s shoulder injury flared up, and J.J. Hoover has been downright dreadful. At the same time, Mat Latos’ injury forced Alfredo Simon to the starting rotation, further depleting the bullpen.

In light of these injures, the Reds had to rely on AAA talent against MLB-quality hitters. With five games against St. Louis, three at Atlanta, three against Tampa Bay, and four against Pittsburgh in the first month, the Reds would quickly find out if their young bullpen arms were up to the challenge. They weren’t.

Now that we are almost thirty games into the season, it is worth asking: How has our new skipper managed the bullpen? Who does he trust in the most important situation, and who is there just to clean up when the fans are already heading for the parking lot?

To get an idea of whom Bryan Price has turned to this season, here is a breakdown of games and innings pitched:

Sam LeCure 10 11.2
Nick Christiani 9 11
Manny Parra 12 10.1
J.J. Hoover 10 8.1
Logan Ondrusek 7 6.1
Jonathan Broxton 6 6
Curtis Partch 3 4.1
Sean Marshall 4 3
Trevor Bell 2 0.2

This is roughly what you would expect given the state of the Reds bullpen this year. Sam LeCure appears to be Brian Price’s favorite arm, using him almost as much as Broxton and Ondrusek combined. The number of games and the number of innings pitched give us an idea of who is getting the call, but not how important each situation is. 

To measure the importance of situations, sabermetricians have developed “leverage” statistics. The “leverage index” measures the importance of each situation during a baseball game. A normal situation is LI 1, while a LI of 2 is considered very important.

“Leverage” is (roughly) calculated by the change in a team’s winning percentage from one play to the next. If a team is ten runs ahead, then giving up a home run and narrowing the lead to 9 will not change the “leverage” index much because the difference between being ahead by 10 runs vs 9 is negligible. If a team is ahead by 1 run in the bottom of the ninth, then giving up a home run has a large impact on a team’s winning percentage. Tom Tango at the Hardball times has a detailed explanation of the index (which, coincidentally, discusses a collapse by Dusty Baker’s Cubs. Stupid Cubs).

Two subdivisions (there are multiple others out there) of this index are relevant to this article:

“gmLI” is the index for how dangerous the situation is when a pitcher enters the game. If Price were to bring in Chapman in the bottom of the 9th with the bases loaded, no outs, and the Reds up 1, gmLi would be around 3.0. If you enter a clean inning, it will be lower. This stat is useful to see whom the manager trusts when the fire is burning, and what reliever Price throws out there when the game is long gone.

“exLI” is how dangerous the situation is when the pitcher leaves the game. By comparing the exLI to the average gmLI, you can tentatively see if pitchers are doing more harm than good (note: if a pitcher ends a game, they will not have an exLI – hence the missing values below).

gmLI exLI
Sam LeCure 1.6 1.35
Jonathan Broxton 1.51
Manny Parra 1.38 1.14
Logan Ondrusek 1.34 1.31
J.J. Hoover 1.32 1.4
Nick Christiani 0.66 0.52
Trevor Bell 0.62 0.97
Sean Marshall 0.59 0.42
Curtis Partch 0.33

Normal caveats apply: Small sample size, biased sampling frames, not enough innings for pitchers to return to the mean, etc.  Keep in mind that Sean Marshall is still coming back from an injury and might be selected for low-leverage work as he gets back into his normal pitching routine. 

From this you can see that Brian Price turns to Sam LeCure when the Reds need to buckle down a win, followed by Broxton and Manny Parra (who, with a 3.84 xFIP, appears to have pitched better than his line this year). Logan Ondrusek and J.J. Hoover round out the pitchers who enter a game when the Reds are in tight situations.

Two pitchers on the list, J.J. Hoover and Trevor Bell, have the inglorious designation of leaving the game in worse shape than they entered it, on average, a rather amazing feat. [I will reiterate small sample size bias here: J.J.’s numbers are not helped by giving up a grand slam walk-off home run. His work this week will help his numbers.].

The good news is that the Reds will be able to bump Ondrusek and Hoover down into low leverage situations once Chapman and Simon return to the bullpen. This index might also give us some insight into who Price will designate for Louisville once the Reds bullpen is back to full strength. As the season progresses we will keep an eye on the leverage index to see who our new skipper turns to when the Reds need help.

Join the conversation! 14 Comments

  1. Hadn’t heard of Leverage Index. Neat concept. Thanks for bringing it to my attention. Based on this and the, “They’ve never made it in the big leagues before” (TNMIITBLB) Stat, safe to assume Christiani and Bell disappear as quickly as an arm is ready to recall?

    • The tough bullpen decision will be when Latos gets back and Simon is moved back to the pen. Chapman and Cingrani will return before Latos. The two pitchers demoted to make room for them will likely be Partch (who was just called up to take Cingrani’s place) and Christiani.

      But when Latos comes back, assuming everyone else stays healthy, you’re looking at a tough decision. At that point, the bullpen will be: Chapman, Broxton, Marshall, Parra, LeCure, Simon, Hoover and Ondrusek.

      Presumably it will be one of those last two guys. That’s where the leverage index comes in. You’d expect the pitcher with the lowest use-leverage to be sent down, all things equal.

      But things (like contract status) may not be equal. It’ll be interesting to watch. And welcome just to see that choice be forced by a fully operational relief staff.

      • Marshall does not appear to be healthy at all.. Lets hope he just needs “to build up arm strength” . I have my doubts . He may ultimately could back on the DL again , despite the contract status

  2. One of the uses of the leverage index – and I’m halfway suspicious that the stat might have been created for just this purpose – is that fantasy baseball players can use it to predict who the next closer of a team will be. High leverage numbers show the manager trusts the pitcher and there’s a correlation between that trust and being used in the ninth inning.

    • Leverage Index was created simply to establish how crucial a situation was.

      But I agree with you that you could use it for the purpose you describe.

  3. Question for Mike: Did you get a chance to compare the relative leverage numbers for the way Dusty Baker used the pitchers in common (LeCure, Hoover, Parra, Ondrusek) last year compared to Bryan Price’s early numbers from this year?

    Sorry if that’s stealing a topic for one of your upcoming columns. 🙂

    • That’s a future post — I think there is a lot to unpack there because Price doesn’t use the lefty/righty split as consistently as Baker did. He also has a strong commitment to using a pitcher after they finish warming up (like last night).

      I don’t think there is any available data on this, but I bet that results in Price warming up fewer pitchers during the season because he won’t feel the need to have both a lefty and righty warming up at the same time.

      • Hadn’t thought of that. Managers who tend to play match-ups more (Baker) would tend to confound the Leverage Index more than managers who tend throw pitchers for an inning (Price).

        I suppose you could turn the entire thing around and use the leverage stats to see if a manager is using his best pitchers (based on pitching stats) in the most important times (based on leverage).

      • Both warming up fewer pitchers and warming up pitcher less frequently (both sides of the same coin). So much of the wear and tear on a reliever’s arm/shoulder/elbow comes in the bullpen, not on the mound. With the approach Bryan uses, a relief pitcher with 70-80 innings will have significantly less miles on their arm during the season than with the approach Dusty sued, which was to completely disregard time(s) warming up in the bullpen as a factor on a reliever’s usage. Hopefully this efficiency for relief usage will result in future increased health in the bullpen.

  4. Has there been any talk of Chapman starting or is it a foregone conclusion that he is in the pen for here until the end of his career? If Parra and Hoover revert back to midseason last year then it still is a solid pen throwing Simon back in and possibly a starter that Chapman would replace.

    • I allowed myself the luxury of that dream for a moment, when it looked like Latos might have suffered a season-ending injury. But short of a serious injury to one of the five, I think the Reds will get Chapman back in his familiar role, with maybe an extra-long save here or there.


    • I’ve read that in a few places. Diaz is not on the 40-man roster, which apparently is a major impediment to having the best 25 players in the system on the major league team.

  6. Speaking of Simon, whatever happened to those accusations against him? They blew up all over the news one day, then it seems like the next day everyone forgot about them.

    Very odd….

Comments are closed.


Sixty Feet from Home