2016 Reds

A proposal for better bullpen management

[This post was written by loyal Nation member, Warren Leeman, who is a regular commenter under the name Shchi Cossack. Thanks, Warren!]

The word “leverage” is a sabermetric measurement of how crucial the game situation is. High leverage situations occur when the game is close. Low leverage situations are those when the game is a blow out.

The most important way leverage enters into baseball strategy is bullpen usage. Even without statistical documentation, savvy individuals can recognize high leverage situations. Big league managers must recognize and deal with high leverage situations on almost a daily basis.

Based on comments attributed to Bryan Price and watching his bullpen management, high leverage consists of two situations: when opposing runners are on base with the tying or go-ahead runs, or the ninth inning with a 3-run lead or less. Price still chooses to manage by the same principles as always: a designated 9th inning closer and designated setup relievers for the 7th and 8th innings. Price is missing the riverboat as it departs the dock.

Bryan Price is certainly not unique in this respect among major league managers, but he’s the only major league manager the Old Cossack cares about since he manages the Reds.

Bullpen management is one facet of a manager’s in-game strategy that is almost universally accepted as having a significant impact on the team’s success (read W-L). During the first half of the season the Reds bullpen management was nothing more than struggling in quicksand with no good options other than to wait out the morass for help to arrive.

Now that the cavalry has arrived, Bryan Price has the opportunity to step up his bullpen utilization as a real weapon. He can organize a true shut down bullpen, a shut down bullpen beyond any seen in recent history.

The Reds have three legitimate high-leverage relievers in Raisel Iglesias, Michael Lorenzen and Tony Cingrani. Each is capable of pitching multiple innings. A fourth reliever, Blake Wood, is available if necessary to fill a high leverage situation if needed.

Cingrani has issues, but for this season he can fill the role in the spirit of finding out who can and who can’t perform going forward. Beyond the 2016 season, the Reds will have additional pitchers headed to the bullpen as casualties of the starting rotation competition. We don’t know who will succeed as starting pitchers, but quality arms will be headed to the bullpen.

What can’t happen is burning up the high-leverage relievers warming up in the bullpen rather than on the mound in a game. The Reds should also minimize the number of times one of these pitchers only goes one inning. The high leverage reliever should find his way to the pitchers mound and fully utilize the opportunity by pitching at least two innings. That’s basically one time through the opposing lineup.

With 3 high-leverage relievers available and a 4th reliever available to fill in for a high leverage situation as needed, at least one is available at Price’s beck-and-call for every single game, every game (!), without making a serious effort to manage the situation.

Price should set a high-leverage reliever rotation. In a full 162-game schedule, all 3 relievers would pitch ~100 innings during the season with no bullpen burnout due to excessive, unnecessary warm-ups when they are not needed and not used. If a high-leverage reliever enters the game, he pitches at least 2 innings and no more than 3 innings.

Of course there will always be exceptions, but those should be minimal if properly managed (i.e. double switches, proper preparation and anticipation, etc.). Wood would have a role to fill in for opportunities less than 2 innings in high leverage situation as needed. This is the same concept as a shutdown bullpen with a designated 7th, 8th & 9th innings reliever, except it’s easier to manage, provides more flexibility and utilizes the best arms for more innings.

There are other situations when a high-leverage reliever isn’t necessary, like when the starter fails to go six innings or if they are ahead by 4+ runs when the starting pitcher leaves the game. This system means the Reds wouldn’t have a ninth inning closer since all three high-leverage relievers will pitch the ninth inning.

The Reds bullpen has ample pitchers to fill low or medium leverage situations.  If the starter goes short, Keyvius Sampson and Josh Smith are options out of the bullpen. They need work to stay game ready too, but not at the expense of the high leverage relievers. If a high leverage reliever isn’t used in a game, the rotation gets bumped back a game, just like the starting rotation during an off day in the schedule. If the high leverage rotation gets bumped back by two games, the next high leverage reliever pitches at least two innings, no matter what the game conditions regarding leverage, since the high leverage rotation will remain intact at that point.

As mentioned there will always be exceptions (a reliever struggles, extra inning games, etc.), but those situation can be managed as exceptions rather than blowing up the bullpen rotation. If the high leverage rotation is completely rested, all three high-leverage relievers could pitch 6-9th innings in a single game if needed and the situation dictates such a commitment. The definition of high-leverage for bullpen utilization just needs to be expanded beyond a ninth inning closer definition.

41 thoughts on “A proposal for better bullpen management

  1. I like the idea. Would also take pressure off starters as it would diminish need for them to go more than 6 innings. Go out, give us a quality start, and take a shower.

  2. So, when you pitch your best guy in the 8th (due to a higher leverage situation) then aren’t you stuck with the guy you didn’t trust in the 8th to pitch the 9th? Or..if you pitch your best guy(s) multiple innings, then doesn’t that make them available for fewer games….therefor increasing how often you need to use the guys you don’t trust?

    Ultimately, when you lack depth, you lack depth. Its still a zero sum game. If Iggy pitches in the 8th, then someone has to pitch the 9th.

    • If Iggy pitches the 8th he can also pitch the 9th. The point is to not ask him to pitch 1 inning every day, but 2-3 every 3rd. If you have 3 guys you trust you are fine and the article is written on the premise that we do have 3 guys we trust.

      • Suppose Iggy gets rocked in the 8th, though. It can happen, and will sometimes, as it will to Lorenzon and–certainly–to Cingrani. The rotation is now in disarray, possibly for multiple games. I agree with the premise that a guy who can go 2 innings should go two innings when the game situation calls for it, but bullpen effectiveness is ultimately predicated upon abundant talent.

    • I don’t see the problem with his plan. If the guy who pitches the 9th will be facing the 6-7-8-9 batters, then maybe you do trust someone there that you didn’t trust in the 8th (against the heart of the lineup).

      The point of his post is that you best the best pitchers in the toughest situations, and use the lesser pitchers in situations where throwing up a zero is less important. Seems pretty simple to me.

      • It’s simple, but its not that simple.

        Aren’t most good relievers in that role because they don’t have the stamina, physique and/or the “pitch variety” to effectively last more than an inning? Also, wouldn’t it just take 1 bad outing in a close game or a long extra inning game to completely derail this? This sort of plan may work for a week, but there is no way it has any staying power. Even “progressive thinkers” like Theo Epstein have reverted back to the defined role bullpen. Of course, the Cubs have bullpen depth so it works.

        If Iggy were to theoretically get shelled in the 7th, but it’s still a winnable game….do you bring in Lorenzen on his “off” day? Or if the game goes 12 innings, do you bring in Iggy on his “off” day….therefor making him unavailable for his “on” day…..forcing you to pitch one of the guys you don’t trust the next day in a close game?

        Over a 162 game seasons, you can’t scheme your way out of lacking depth. Yes, I would love for the best pitcher to be involved in the most high leverage situation all the time….but there is always a domino affect. The Reds lack bullpen depth and you can’t fix that on an excel spreadsheet.

        • A lot of relievers are there because they don’t have the repertoire to go through the lineup more than once. This would not be an issues if you only pitch 2 innings (unless you get shelled, in which case he’s probably getting taken out anyway.

          And yes, you still bring a guy in on his “off” day if it’s a winnable game, so long as he is rested and available. You don’t save a pitcher for the next day, because who knows if he will even be needed. Maybe our starter goes the distance. Maybe we get shelled by 8 runs and the “high-leverage” guys are not needed. Regardless, you take the chance to win the game today and let tomorrow worry about itself.

          And with regards to depth, there is not a team in baseball without at least one pitcher in their bullpen who is not cut out for high-leverage outings. Simply doesn’t exist. Enlightened managers know to use their best guys when it matters most, and their lesser guys when it doesn’t. That’s it.

    • Also Chuck, if you pitch someone for one inning, he will still need to have a day off here or there. Conversely, if a guy can pitch 2 innings every 2 days, then he can pitch more innings. Your best pitchers should pitch more innings, and higher-leverage innings, than the lesser guys.

  3. I really like your suggestion here. The problem is, I don’t think there would be very many managers that would actually do this.

  4. Seen where Max Wotell did not last a inning. Did we trade for two players that were damage goods from the Mets. Will the Reds Management Keep Quiet about then getting screwed.

  5. Good stuff, I like your thinking here. It could really reinvent bullpen usage. I think it would also allow the bullpen to carry only 7 relievers instead of the typical 8 we’ve seen the last two years, allowing for a deeper, more specialized bench.

  6. Love it
    Great concept.
    I’ve always felt that a 9th inning closer was foolish.Bring your best guy into the most critical spot whether its the 7th,8th or 9th inning.Why use him only in the 9th?
    I cringe when they waste a guy after 8/9 pitches and dont allow him to come out for a second inning.I truly believe that is part of the reason for all the TJ surgeries.
    Kinda like a car battery.Turn it on and let it go.The short little trips kills the battery a lot quicker.

  7. Please notice all the teams leading divisions have a designated 9th inning closer. Having quality relievers who can pitch the 7th and 8th is essential for success. A lefty specialist is also needed. Attempting to reinvent reliever strategy is something the fans of losing teams do for lack of a having better players to win games. The need to pinch hit for pitchers in close games works against multiple innings on many occasions in the NL. Hopefully the Reds will come up with a solid closer and the good set up men. Then we will enjoy watching the Reds relievers the same way the fans of the Cubs, Cards, Dodgers, Giants, Nationals, Pirates, Orioles, Blue Jays, Rangers, Red Sox, and Yankee do.

    • The lack of set up men or closers is not the issue. The Reds had the most DOMINANT “closer” in the game, to make him effective the talent has to be in the rest of the players to make a “closer” relevant. The teams you mentioned are not there because of set up men and closers some of them traded to get a closer and a couple still don’t have one. The teams you mentioned for the most part have a real oood team to make the back end of the bullpen relevant.

      • OK. Changing to a commitee of multiple relievers is not the answer to anything. All the mentioned who are doing well teams have closers and if they thought they needed a better closer or set up man they went out and got one….Cubs, Nats, Indians, Sawx to name several. Good teams have good overall balance and they all have a closer.

    • Well said. The Reds aren’t going to be the first to do anything. In order for the Reds to find the road to success they will have to build their models off of what other teams do that lead to success. A traditional model of having a setup man, a closer, and a left handed specialist along with a few innings eaters appear to be the traditional model and I do believe that the Reds will also find success doing the same thing. We just need to get pitchers developed up that can go 7 or 8 innings at least 50% of the time instead of lasting just 5 or 6. It would be nice to have at least one pitcher that we could count on to give us a complete game from time to time. I think it is quite possible the Reds could go the entire season without a complete game.

  8. Mets-Familia, Nats-Melancon, Marlins-Ramos/Rodney, Cubs-Chapman/Rondon, Cards-Oh, Dodgers-Jansen, Giants-Castilla, Pirates-Watson, Sawx-Kimbel, Yankees-Balances, Indians-Allen/Miller, Rangers-Dyson, Orioles-Britton, Blue Jays-Osuna, Tigers-Rodriquez. White Sox-Robertson, etc…..

    • The designated Closer Role started with LaRussa using Dennis Eckersley exclusively in the 9th for the A’s. Before that, teams pretty much used bullpens the way the Cossack suggests here; check how Goose Gossage and Rollie Fingers were used. Statistics show that teams with a lead going into the 9th win pretty much the same percentage of games now, as they did in the 50s, 60s, 70s, etc., so the Closer Role hasn’t really been as fruitful as you suggest. And why on earth use a guy like Chapman or Rivera to get a save with a 3-run lead?

      The simple point is to use the best pitchers in the most high leverage situations, and then don’t take them out if they are effective. I’ve pulled my hair out for 25 years, watching managers take out a reliever who is pitching well, and then watch the new guy blow the game up. Removing a pitcher who you can see is throwing well, and replacing him with somebody who may or may not throw well, seems to be silly. It is what I call WWLD–What Would LaRussa Do.

      The whole idea of designated roles for 7th, 8th and 9th inning guys assumes that the team will be ahead after 6 innings. That didn’t work too well for the Yankees this year, because they weren’t ahead very often after 6 innings. Maybe if they had used those guys earlier in the game, when the game was actually on the line, then they would not been behind later in the game.

  9. Many times bullpens built “from the closer forward” put the cart before the horse. When it’s all said and done, Every major league staff works to figure out a way to get 1,458 innings while giving up as few runs as possible.

    But to split the difference with those who oppose this idea, let’s swap say Josh Smith in for the Cingrani. Then make sure you pitch Lorenzen, Iglesias, and Josh Smith every third game.

    Over a whole season they’d each pitch 108 innings, and given their stats from this year, they’d give up roughly 100 runs total. That’s 324 of the total innings needed at an ERA of only 2.77.

    That relieves an awful lot of pressure from the remaining 1,134 innings in the season.

    • But I think the point is that you can’t “make sure” to pitch these guys every third day. Things happen–they throw lots of pitches in their previous outing, they get hurt, they are ineffective in their first inning, etc. I certainly agree that rigid closer rules are questionable, and that aiming to have your best relievers pitch the highest leverage innings makes great sense, but as–I think–Chuck is saying, substituting one rigid set of roles for another won’t work over a full season, and will not disguise a lack of depth.

  10. Love the idea. Any means of putting your best pitchers in the most important situations is advantageous compared to what we have. Another option is to send out your highest strikeout rate guy in situations that demand strikeouts (close game in the 7th, 2 on 0 out).

    Unfortunately, this is not attempted because of clubhouse/ego/star management surrounding the “Save” stat. Great relievers want saves cuz saves==$$. Come up with a way for relievers to get sexy power stats outside the 9th inning and then it can be done.

    • So, you’re suggesting that a manager should be able to anticipate several batters beforehand that the ” strikeout guy” will be needed and he should have him warm up? Do you think having the ” strikeout guy” potentially warm up 2-3 times per game is a good idea? Do you think that may cause some problems over 162 games?

  11. Thanks to the Old Cossack for this interesting article. I have been thinking about this lately and I believe that bullpen management is an area where mathematical modeling can be applied. Obviously, as pointed out above by several writers, you are constrained by the bullpen talent. However, even with talent constraints I am convinced that modeling is applicable and can make a difference.

    In operations research there is a problem known as stochastic bin packing. An application of this problem is managing server farms where each server has a capacity and corresponds to a bin. As jobs arrive online there is an algorithm that assigns jobs to servers. At each point in time the future is stochastic — you do not know how many jobs will arrive nor their capacity requirements.

    Bullpen management is a complicated version of stochastic bin packing. The relief pitchers correspond to the bins/servers. They have a multi-dimensional capacity (innings per year, per game, how recently they pitched, etc.). The jobs that must be packed correspond to high leverage situations. There is also a different potential return associated with assigning a job to a bin. An Ohlendorf bin is different than an Iglesias bin. Like the server farm problem, this problem is highly stochastic in that over the 162 game season we do not know how many jobs must be packed, their arrival rate, or their capacity requirements.

    I would not be surprised if an appropriate modeling of bullpen management could lead to a four or five game improvement in a season. This would not matter for the Reds this season, but for teams that are in the 90+ win category this would be huge. Consider the poor Pirates in 2015 who finished two games behind the Cardinals. Their reward was a single playoff game against Jake Arrieta. I wonder if some of the teams with progressive management are attempting to model this problem. At the minimum it is worth a masters or Ph.D. thesis in operations research.

    • KMartin….I believe the key hiccup to a scientific bullpen is that pitchers have to warm up. This also impacts creating the optimal “high leverage” match up.

      • Could you explain the “hiccup” a bit more so I understand the issue better. I envision an algorithm producing prior to each game a ranking of pitchers for each scenario in a rather large set of scenarios. As soon as a scenario in the list of scenarios occurs the first pitcher at the top of list warms up in the bullpen. I can also envision the the algorithm being real time and making adjustments to the scenario list as the game progresses. This may seem like fantasy, but I think it is possible.

        Is the hiccup that while the pitcher is warming up, the events on the field change the scenario?

        • Yes, that, or that the situation requiring the strikeout guy arises too suddenly to get the pitcher warmed up.

        • KMartin….my concern is how quickly you can get Pitcher A into the game when the model suggest that he’s the optimal person.

          Many ” high leverage” situations arise rather quickly and my concern is that by the time Pitcher A is ready then it’s too late….or if you anticipate that Pitcher A is needed…and he’s not…then you have him warm up..sit down..warm up again later…which over the course of a season adds an enormous amount of stress.

          I trust your mathematical abilities and use of logic far more than I trust my own ability. The “warm up” issue is the key issue for me.

  12. I agree for the most part… but in a close or tie game… might be difficult to let even your best ” high leverage guy” have to bat, so you’d have to be careful to have the right bench management too and be willing to double switch out one of your regular position players to make sure that spot in the order doesn’t come up or be willing to send a reliever who has not had an at bat in weeks (if that) to the plate late in a tie or close game.

  13. One of the practical issues is money. Relievers and their agents have monetized the Save stat. You get paid more if you are The Closer. Agents aren’t going to like negotiating the “he-did-great-in-high-leverage-situations” stats.

    • All pitchers are being paid to pitch, period. They have some psuedo roles of course, but a team can and should use them as needed. I know a few will openly complain, but you need to know this about a pitcher so you can trade him or not extend him…due to lack of team attitude.

    • Very true. Doing something outside the box, rather it be this plan (which I’m not sure about) or just a general plan to try to get your best relievers in the game when it’s actually on this line, is going to come down to getting buy in from the pitchers. The team that sets this precedent is going to have to find a way to pay pitchers appropriately based on leverage. Not sure how that is going to come about but that is the problem that front-offices are going to face.

    • That’s very true.

      It’s also very baffling — if any other player on the roster complained this way (“the manager’s usage of me is hurting my individual stats, and therefore my free agent value”) he’d be tarred and feathered. Yet it’s assumed to be valid for relief pitchers – the most replaceable men on the roster.

  14. I like this plan. I wonder what this plan’s WAR is compared to what Price is doing now that the Calvary has arrived.

    One point that might help this plan succeed is that none of the pitchers currently is getting paid “closer” money. Hoover is gone and seems to be done. The rest are young players. If there was ever a time to break entitlement based on which inning a pitcher pitches, it’s right now now.

  15. In response to some of the discussion above, this is a proposal for high leverage games, not high leverage situations.

    One key to this proposal is NOT to match the BEST pitcher to the MOST CRITICAL high leverage situations. Those occur too infrequently and too often without warning to be effective and result in excessive warmup time in the bullpen without getting to the mound. That’s wasted utilization of the best bullpen resources.

    Another key to this proposal is the availability of MULTIPLE (3) high leverage relievers who CAN pitch muliple innings every time out. The Reds have that capability in the bullpen, now and in the future, while other teams do not.

    If a starting pitcher goes at least 6 innings of a close game, the FIRST option out of the bullpen is one of the high-leverage relievers and that reliever will pitch 2-3 innings to finish the game. If a high-leverage reliever warms up to enter a game and the game no longer requires a high-leverage reliever, he enters the game anyway and pitches at least 2 innings or he sits down to await the next game if the high-leverage relievers have been used a lot recently.

  16. I entirely agree with the premise. And, given a young staff, you essentially have to go with something like this.

  17. The only problem with this premise is that it essentially punts platoon advantages in the late innings. That’s a very real weapon for modern managers. And, I suspect, one that helps mask weaknesses with some of these pitchers.

    Do we really want Cingrani facing Harrison-Marte-McCutchen?

  18. When Cingrani comes in I get that same queasy feeling I always got when Todd Coffey used to sprint across the field to pitch….(shudders) horrible memories.

Comments are closed.