Revitalizing the Long Relief Pitcher

In the last few decades, the league has clearly put higher emphasis on role specialization.

The most salient example of specialization is the closer, a role that evolved until it lost some utility. As a reliever improved, he became increasingly likely to be moved into a later-inning role, until he became the go-to guy on his team. It just so happens that the number-one guy would often enter in the last inning with a lead and a clean slate; it eventually became apparent that these elite arms were too often being saved for the 9th when the outcome of the game had been in danger during previous innings.

The problem of the game being “on the line” is often only clear in hindsight, or in the last couple innings, (when it is also obvious to everyone in the stadium). It’s clear that managers are reluctant to pull their starting pitcher too early, because they want keep their relievers fresh, and maybe want their starter to record the win.

In 2015, relief pitchers entered games in which their team was either winning or losing nearly 84% of the time. This suggests that managers are not likely to go to their bullpen when the game is still tied, meaning that relievers are often brought in after the momentum has already shifted.

I feel there exists an optimal starter/reliever dynamic, in which managers intuitively sense the balance of the game changing earlier on. Essentially, instead of trying to mitigate disasters created by leaving a starter in too long, you may attempt to avoid them entirely by preemptively yanking your starter out of the game. This happened extensively with the Tampa Bay Rays in 2015; manager Kevin Cash “pulled [his starters] quicker than anybody,” Pulling starters early means leaning more heavily on your bullpen, though. Those innings have to go somewhere after all.

Bullpen or Rotation?

A 2011 baseballprospectus article outlined the benefits of a “quick hook” on starting pitchers when they come up to bat, (both NL and interleague matchups), during high-leverage situations after pitching at least 5 innings. Your pitcher in this case doesn’t hit in a high-leverage situation, but the drawback, of course, is that you must go to the bullpen immediately afterwards. The projected advantage was about 5 extra wins if a team’s best starters are sometimes allowed to stay in the game, reducing their being pulled for a pinch-hitter by “more than 15 percent.”

Bullpens are an integral and delicate part of every baseball team though. Frequent usage wears players out over the course of a season, and having them pitch in longer stints than they are prepared for is often an ill-fated idea. Pitchers’ numbers dip as their limits are pushed, and introducing less experienced relievers to spread out the workload seems just as likely to fail as it is to succeed.

In comes the swingman. He is almost always a developing or middling starter that is utilized according to his versatility, but in the modern era, there really hasn’t been a dominant one. Obviously, if a young guy pitches well as a starter, he often becomes a starter. If he pitches well as a reliever, he often becomes a reliever. Older pitchers differ in that typically, as they become swingmen when their skills diminish.

So it seems that to survive as a swingman, you must be more-or-less consistently mediocre. But I don’t want my team to sport mediocrity just to fill up innings or save better arms. What I want is a true long reliever.

A True Long Reliever?

By “true” long reliever, I mean a pitcher who was developed specifically for the role. This is a pitcher who doesn’t start games, but will work strictly as a relief pitcher. This long-reliever also won’t be subject to mop-up duty any more often than his teammates.

Much like starters and closers, his role should not fluctuate throughout the season. The long-reliever should be groomed and poised to fill 2 to 4 innings every 3 or so days at the major league level by deliberate development. Of course, no plan is truly disaster-proof, but let’s leave the drawbacks for the end, and look at some potential positives.

In 2015, starters averaged 5.8 IP (17.4 outs) and 93 pitches per game started (GS). Relievers averaged 1 IP (3.0 outs) and 17 pitches per appearance. This leaves a typical 9-inning game up to 1 starter and 3 different relievers, or 4 total pitchers. Of 15,108 appearances by relief pitchers, 23% covered more than one inning, while about 19.5% went for more than 3 outs per baseball-reference.

This manifests itself, in that 6-inning starts have seemingly become a benchmark for success. However, in this article, it appears that the number of pitches a given batter has seen from a particular pitcher influences his wOBA more than the number of innings the pitcher has thrown. High and low pitch counts to each batter in this case were 4 or more and 3 or less, respectively. Maybe lasting 6 innings isn’t an important threshold for managers to consider.

With batters averaging 3.82 pitches per PA in 2015, I calculated that pitchers average about 24 PA/GS. This is enough to get the top 6 batters a third PA against a starter. For a combination of reasons, the third time through the order is the killer more often than not. If we scale the average PA/GS back to 18, we get a rough average of 69 pitches/GS. This obviously looks quite low, but could it become the new norm?

The Return of the Long Reliever

For sake of simplicity, we’ll assume each team gets 38 PA per game (4.2 per 9 innings). If it takes roughly 150 pitches (93 + 17 x 3 + extra) to get through a game and we assume starters only throw 18 PA (~70 pitches or twice through the order), then relievers must compensate with about 21 PA (~80 pitches) on average (I realize this adds to 39 because of rounding error, but again this is just an exercise).

With three 1-inning relievers, that shakes out to 7 PA (27 pitches) per RP. This is seems too extreme to be viable as a daily strategy. If we factor in a long reliever that throws 2 innings (8.4 PA or 32 pitches), then you have about 13 PA left for short-order relievers. A 3 IP outing leaves about 8 PA while a 4 IP outing leaves about 4 PA. This could work well when the bullpen is well rested, but what this illustrates is that it is probably not viable to have your starters average two times through the order (~4 IP). This means a third time through the order is largely unavoidable.

So from this point, I’ll operate under the condition that that starters could effectively average 5 IP/GS or rather 21 PA/GS, leaving roughly 18 PA for relievers to complete. I will also assume that it is necessary for fewer than 4 total pitchers to throw in a game on average, in order to “save the bullpen.” Since we’re working with averages, totals between 3 and 4 pitchers are considered to be giving other relievers more rest.

A 2-inning (8.4 PA) relief outing would leave about 10 PA for 1-inning relievers. A 3-inning (12.6 PA) outing would leave about 6 PA for 1-inning relievers. Finally, a 4-inning (16.8 PA) outing would pretty much finish the game. Following this process, I would wager that a 5 IP start and 3 IP relief outing would pay dividends in both saving the bullpen, and limiting the exposure of your starters. 2 IP outings could compliment 5 IP starts well enough to offset using more pitchers, but I don’t think it would necessarily be anymore effective than the current system.

piece following the previously mentioned article examined the effect of the Times Through the Order Penalty (TTOP) on pitchers who throw 1, 2, or 3+ pitches at least 15% of the time. One-pitch pitchers had the overall advantage the first time through the order. This is because if you throw just one pitch, it’s probably pretty good, but they also suffered the steepest penalty the 2nd time through the order.

Pitchers with 3 or 4 pitches had the softest penalty the 2nd time through but ended up with the worst wOBA through the 3rd and 4th times through. Given that our hypothetical long reliever would rarely, if ever, pitch a 3rd time through a lineup, we can assume that for the role to be effective, a pitcher would need 1 really good pitch and at least 1 good secondary offering.

It’s clear that many managers are recognizing these trends, and are wary of leaving their starter out too long, but the fact that relievers in general are not developed to pitch more than 3 outs at a time kind of forces their hand sometimes. I believe a solid long reliever, one developed specifically for the role, can make the short leash a viable option more often than not.

Making it Easier on the Manager

Let’s cherry pick an example of poor management. Take a look at Chris Sale, of the Chicago White Sox, in a start against the Boston Red Sox in Fenway Park on July 30th, 2015.

Jose Abreu got the White Sox off to a 2-0 start with a home run in the top of the 1st, David Ortiz answered with an RBI double in the bottom half for the Red Sox, and then there was a lull in scoring until the 5th inning. To begin the 5th, with the White Sox up 2-1, Sale sat at 4IP, 5H, 1ER, 7K, 1BB, with a wild pitch.

In the bottom of the 5th:

Blake Swihart flew out, Sale hit Jackie Bradley Jr., and Brock Holt grounded into a fielder’s choice. Then, things got iffy, as Sale’s performance began to unravel. Xander Bogaerts singled, Hanley Ramirez singled, (scoring Holt), and Ortiz singled, (scoring Bogaerts). Mike Napoli then popped up to end the inning, but the damage was done. The Red Sox took the lead, but Sale still could have ended his day with a decent 5IP, 8H, 3ER, 7K, and 1BB.

Manager Robin Ventura ran Sale out there for the 6th inning, though:

Rusney Castillo singled, Josh Rutledge was hit with a pitch, then Swihart singled to load the bases for Boston. Bradley Jr. then singled in one run, Holt singled in two more, and finally Sale was pulled. Reliever Matt Albers allowed one inherited runner to score, (how could he not with none out and a runner on 3rd?), leaving Sale with an ugly line: 5IP, 12H, 7ER, 7K, 1BB, 1 WP, and two hit batters.

There is much to be said about a guy you can expect to give you quality innings every time out, like Chris Sale, but no pitcher is perfect. Chris Sale didn’t lose that game; Robin Ventura did. At no point was his finger on the trigger because Sale wasn’t dominating, and he mustered a call to the bullpen just as soon as the game was already effectively over.

With a reliable long relief option, maybe Sale doesn’t go out for the 6th. A less talented starter may almost certainly be pulled in that situation.

In my mind, in close games, an intuitive manager and coaching staff should sense when opposing hitters are coming around on his starter, and pull him before any “too late” scenario is in play. A specialized long reliever or two should help make these decisions easy.

Hypothetical Qualifications:

One 55-80 pitch

At least one 50+ secondary offering

Stamina to last 2 to 4 innings or 8 to 17 plate appearances

Must be a good sport about taking on a new role

No starting experience needed

Fun for all ages!

Candidates, Past and Present

It’s possible we could see a strategy like this implemented with some frequency in 2016. Notably, the Red Sox and the Chicago Cubs have a few arms that could slide into this role with a relatively short adjustment period. There is no guarantee that these guys would be any more effective than previous typical journeymen long relievers.

The Baltimore Orioles, however, could potentially have an ideal long relief candidate in Dylan Bundy, who is currently out of options. Bundy has a good fastball/cutter combo that pairs well with a good curveball. Injuries have kept him from logging high innings totals in the minors, and he isn’t high on the SP depth chart. Bundy is definitely capable of handling a multi-inning role however.

The organization wants Bundy in the rotation in the long run, but that doesn’t seem like a possibility if they want to make a serious playoff run in a very competitive AL East. 2-4 inning outings, then, could be the norm for Bundy, as the O’s try to stretch him out and limit the exposure of their admittedly weak rotation.

Injury-risk should raise some red flags here though. No reliever has thrown more than 100 innings since Scott Proctor in 2006. No relief pitcher has come very close to 90 IP since Craig Stammen and Josh Roenicke threw about 88 IP in roughly 60 appearances each in 2012, and for good reason. Proctor pitched 102.1 innings in 2006 over 83 appearances! He matched this total the following year before becoming a disaster in 2008 at just 31 years old.

Plenty of starters exceed 100 innings, so it’s not the sheer total that raises injury risk; it’s the frequency of usage that poses health concerns. Under this system, I would see long relievers making under 50 2.5 IP-appearances (an average), netting them around 125 innings a year. This usage should make development critical, as not just any journeyman should be thrown into this role.

Issues and Concerns

The biggest obstacle to popularity among pitchers here may be financial. By limiting starters’ innings, we also limit their earning power on the open market and during arbitration. This is certainly a likely outcome because pitchers that frequently get pulled early use a track record of being durable/going deep into games as leverage in negotiation.

How much this might effect contracts is difficult to even roughly estimate, but it would have to be balanced with the benefit of getting a little more rest each season, which would hopefully lessen fatigue late in season and ultimately help extend careers.

Also, given the relative unfamiliarity of an established long reliever, it may take awhile for these pitchers to get properly compensated, assuming they do bring good value to a club. In addition to the role itself not exactly being prestigious, this could make long relief duty quite unappealing to many players.

With these things in mind, it’s hard to see any team trying to jump ship and adopt a new philosophy immediately, but it’s fun to entertain new strategies, even if they’re just old ones polished up.

Statistics provided by:,,

Image thanks to:

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s