First Pitch: The Common, Yet Wrong Way to Manage a Bullpen

There is no reason why Mark Melancon should be handcuffed to the eighth inning. Photo by: David Hague
There is no reason why Mark Melancon should be handcuffed to the eighth inning. Photo by: David Hague

After today’s game, Clint Hurdle announced that he would only be using Jason Grilli and Mark Melancon when the team is ahead. Hurdle didn’t use either reliever in the ninth inning, instead turning to Tony Watson for a second inning in a game that was tied 4-4. Watson gave up the eventual winning run, and the Pirates lost 5-4 with their two best relievers on the bench.

I can understand not wanting to over-use your top relievers. Grilli and Melancon have both been worked hard this season. I don’t have as much of a problem with not using either guy in the ninth inning. My problem comes in the sixth inning.

Now I know I’m crazy for suggesting this, but the time to use Grilli or Melancon would have been with no outs in the sixth inning when there were two runners on base. That’s crazy because you don’t use your best relievers until the 8th and 9th innings. Everyone knows that. That’s how every baseball team runs their bullpen. If you use them to preserve a lead in the sixth inning, then you don’t have them to preserve that same lead in the eighth inning. You’d have to turn to one of your lesser relievers in the later innings to preserve those leads, rather than relying on those lesser relievers to get the lead to the top guys.

It may seem smart to use your best relievers earlier in the game to ensure that the game reaches the later innings, especially in a situation like today’s game where you’ve got a two run lead and two runners on with no outs. But that ignores a key fact. Pitching in the eighth and ninth inning is special. It’s magical. It’s not something that any reliever can do. Pitchers who have success in the seventh inning can’t be trusted in the eighth inning, because it takes something special to pitch in the eighth. Pitchers who have success in the eighth inning can’t pitch in the ninth, because the ninth is even more special. Don’t ask me how anyone ever got to pitch in these innings to begin with. If you can only pitch in those innings after you have that magical ability to pitch in those innings, but you can’t get that magical ability until you get established in those innings, then how does one get that magical ability in the first place?

There might be those that tell you that the eighth and ninth innings aren’t always the most important innings. Take FanGraphs, for example. They have a stat called “Leverage Index”, which profiles the leverage for each at-bat. A higher LI means a higher leverage situation. A lower LI means an easier situation. Only 10% of all game situations have an LI greater than 2.

When Justin Wilson entered the game, he had an LI of 3.28 against Adam LaRoche. That went up to 4.19 against Tyler Moore, and 4.35 against Wilson Ramos. He then had a 3.01 against Roger Bernadina, and a 3.18 against Stephen Strasburg. Wilson entered with a 3.28 LI, and all five of his at-bats were over 3.0.

By comparison, Jason Grilli has had four at-bats this season where he’s had an LI greater than 3.0. He has only entered the game once with an LI greater than 3.0. That was a game where he entered down by one run, so it wasn’t even a save situation. Mark Melancon only has three at-bats this year with an LI greater than 3.0. He has yet to enter an inning with an LI greater than 3.0.

The argument that you should use your best reliever in the highest leverage situations might make sense. You might look at the above, realize that Justin Wilson just had more high-leverage at-bats in one inning than Jason Grilli and Mark Melancon each have had the entire season. You might think that because of this, one of Grilli or Melancon should have been pitching.

I’m going to cut the sarcasm here and be straight on this subject. On average, the eighth and ninth inning will produce higher leverage than any other inning. If you look at the results this year, and look at gmLI (the Leverage Index when a player enters the game), you’ll see that Jason Grilli and Mark Melancon rank first and third on average this year. That means that, on average, those two are entering the game with some of the highest pressure situations. If you look at pLI (measuring the LI of all game events), they rank first and second. I don’t think you can argue that, on average, pitching in the eighth and ninth innings will produce the highest leverage.

Here’s where you run into a problem. On average the eighth and ninth innings produce the highest leverage. But that doesn’t mean relievers should be chained to the eighth and ninth innings. Tonight is a perfect example. The situation Wilson inherited was tougher than 90% or more of the situations Melancon and Grilli inherit. Those two are starting innings with no one on base. Wilson came in with two on and no outs. On average the eighth and ninth innings are going to have a higher leverage than any situation Wilson pitches in. But that doesn’t mean you won’t have higher leverage situations before the eighth and ninth innings.

The smart thing to do would be to use Jason Grilli in that situation. He’s your best reliever. The argument against using Grilli in that situation is that you’d have to have someone like Wilson throwing in the ninth inning. But as shown above, Grilli has entered 14 games this year, and only once has he entered a game with a situation as difficult as the one Wilson inherited. The argument against using Grilli in the sixth is that you’d put Justin Wilson in a situation where he might enter a high leverage situation. That doesn’t make much sense. For one, you’re guaranteeing Wilson will see a high leverage situation in the sixth. If you can’t trust him in the ninth inning, why would you trust him in the sixth to get the lead to the ninth inning. Two, as shown with Grilli’s numbers, there’s a very small chance that Wilson is going to enter the game in the ninth with that high of a leverage index.

Even if you can’t possibly use a closer outside of the ninth inning, you’ve got Mark Melancon. There’s no reason why Melancon has to be handcuffed to the eighth inning. The choice seems like a no brainer. You’ve got either:

1. Melancon throws in the sixth with two on and no outs. Wilson throws in the eighth with no on and no outs.

2. Wilson throws in the sixth with two on and no outs. Melancon only throws in the eighth with no on and no outs if the team is leading.

On average, Melancon and Grilli are going to have higher leverage situations pitching in the eighth and ninth innings. But they are your two best relievers, which means you want them pitching in as many high leverage situations as you can get. Not every high leverage situation will come in the eight and ninth inning. By handcuffing these two to those innings, you force your lesser relievers into those high leverage situations. That’s ironic, because the main reason you don’t go with Grilli or Melancon earlier in the game is due to a lack of trust in those lesser relievers to pitch in the traditionally higher leverage 8th and 9th innings.

This is a trend all across baseball. It’s not just the Pirates. One day a manager is going to manage games based on leverage. He’s going to bring in his best reliever when the first high leverage situation comes up, rather than saving him for later and hoping for the chance to use him. That manager will be criticized and mocked, because it’s crazy to use your closer or eighth inning man in the sixth or seventh. But that manager will also see fewer blown leads in the sixth or seventh inning. Eventually people will catch on, and every team will adopt this method. People will think about how crazy it was that teams once only used their best reliever in the ninth inning when the team was up by no more than three runs. Relievers will be paid based on the numbers that matter, and how they handle high leverage situations, rather than their save totals, which will remove the pressure for teams to only throw their closer in the ninth inning.

Or maybe that’s just my dream. Chances are, things won’t change. Teams will continue throwing their lesser relievers in high leverage situations, all because they have to save their best relievers for the final innings, in the event that those lesser relievers manage to get the lead to those late innings. It doesn’t make any sense typing it out, which is why it makes zero sense that every major league team operates this way.

Links and Notes

**The 2013 Prospect Guide and the 2013 Annual are both available on the products page of the site. If you order them together, you’ll save $5.

**Be sure to check out the new podcast: P3 Episode 2: The Returning Pitchers, The Hot Start, and a Robby Rowland Interview.


**Prospect Watch: Cole and Pimentel Have Good Starts, Lambo and Garcia Homer.

**Minor League Schedule: 5/5/13.

**Draft Prospect Watch: Bryant Homers, Two Catchers To Follow.

**Minor Moves: Barrett Barnes to the DL.

**Pirates Sign Jeff Gibbs.


**Pirates Pregame: Mercer’s Role and Hurdle on Strasburg.

**Video: Hurdle Talks Locke, Bullpen; Strasburg More Efficient.

**Nats’ Ninth Inning Rally Beats Pirates.

Oldest Most Voted
Inline Feedbacks
View all comments
Kevin Anstrom

Over the next 10-20 years I expect to see major changes in the way teams use their pitching staffs.

In today’s game we rarely see complete games and it’s not just an issue with the Pirates. Take the Colorado Rockies as an example. Thru 30 games their starters have pitched 6+ innings 16 times (53%), 5+ innings 23 times (77%), and 4+ innings 29 times (97%). No starters have pitches 8+ innings thus far.

I would propose they consider the following structure …

4 short starters
5 majority pitchers
3-4 short pitchers

The short starters would begin the game and pitch to a maximum of 9 hitters. They would start a game every 4th day.

The majority pitchers would be used in a similar fashion to today’s starters except they would enter the game in the 4th inning. They would pitch every 5th day.

The short pitchers would fill in the gaps in the 3rd inning and in the 8th/9th/extras as needed.

This structure would essentially have 9 pitchers going on a set schedule. It would be easy to keep those 9 pitchers under pitch / innings counts.


Just more baseball nincompoopery. My girlfriend (who doesn’t really watch baseball) asked the same question recently. I answered without thinking “those are the guys the Manager trusts the most.” which then led to her restating the question like she was talking to a child. I see a movie being made about the first Manager/GM who embraces this concept, Joe Morgans voice can be heard throughout the movie saying negative things like “this just won’t work” like in Moneyball.


Clint Hurdle will never, ever get this. It’s pansy thinking whereas he’d rather use macho gut management. All this sissy-stats-boy stuff doesn’t belong in a dugout. Leave it for the prissy college boy front office types.

Lee Young

Personally, I wouldn’t have minded seeing what the grizzled vet, Contreras had left in his splitter.


We are agreeing more often lately.

I’ve argued for years, that managers use their best relievers the wrong way. There are four reasons for this:
1. They truly believe the myth that some pitchers can pitch in the last inning. It doesn’t matter that 95% of all closer were pitchers who couldn’t pitch innings 1-8 and had to be relegated to the pen.

2. Managers manage by stat. Managers are managing according to the save opportunity. Its why Hurdle (and every other manager) won’t use his closer unless his team is leading by less than 4 runs. They only change this pattern in the post season.

3. Managers are managing based on “hope”. By keeping their best relievers out of the game, they are hoping they still have the lead later in the game. Hurdle’s fiat has nothing to do with rest, its all about getting Grilli saves. That thought process will has impact on and off the field. See how much Grilli costs when his deal is up.

4. Finally, managers don’t want to think or be held accountable for their thoughts. By placing even more restrictions on his 2 best closers, he’s taking in game thought out of the process. He’s managing by a code, not by analysis. He now has a book that states that he will not use his closer, when his team is behind, up by 4+ runs, and tied at home.


In the past closers were referred to as “firemen” and while you occasionally still see that term it’s rarely appropriate anymore (the exception being when a manager brings in a closer to get out of a jam in the 8th to get the last 4-5 outs instead of just the last 3).

So I propose a new role in the bullpen–the fireman. This will be the guy you bring in the first time you face a jam after the starter is done. Of course then we need a new “common stat” so that these players are fully appreciated when it comes to arbitration and free agency. Publicizing LI would seem to be a good place to start.


While I tend to agree that using your best reliever to close is vastly over-rated and therefore closers over-rated, I wonder how realistic it is to use your best reliever or two in the highest leverage situations the majority of the time. The reality is, the guy needs to be warmed up to enter the game and managers will rarely have the luxury of having the best guy ready to go when the situation calls for it. Is a manager gonna have either Grilli or Melancon throwing in the pen every inning from the sixth on, just so they’re ready in the event the first batter or two reaches base and a high leverage situation presents itself? I don’t think this is appropriate either. Just goes to show the importance of having a deep and talented bullpen.


Thank you. I feel like I got to vent vicariously through your article.

Comments are closed.

Most Voted Comments