Pitching Labels Are Increasingly Irrelevant

Current Cy Young-candidate Chris Sale and former Cy Young-winner Justin Verlander faced off last Monday in Game 4 of the ALDS on a dreary afternoon in Boston. It’s the kind of matchup that grabs our collective attention. It’s how they were matched up, however — each pitcher entering the game out of his respective team’s bullpen — that merits further consideration. For Sale, it marked his first relief appearance since 2012. For Verlander, it was the first time he’d pitched out of the bullpen as a major leaguer.
During LDS play, David Price, Jose Quintana, and Max Scherzer were among the other starting pitchers employed as relievers.
A year after Buck Showaler failed to use Zach Britton in an elimination game and the Indians creatively employed Andrew Miller in the late summer and October to nearly advance to a World Series title, it seems managers (and, by extension, the clubs they represent) are attempting to replicate the latter approach, thinking unconventionally, moving away from tradition to best leverage pitching talent.
We did, of course, witness relievers like Miller, Aroldis Chapman, and Kenley Jansen getting employed in unconventional, sabermetrically sound ways last October. We’ve also seen some aggressive usage patterns from Joe Girardi — who essentially bullpenned the Wild Card game — early this postseason.
Girardi didn’t hesitate to go to his bullpen at the first sign of trouble during New York’s Game 5 in Cleveland, summoning David Robertson and Chapman for multiple innings of work.
The result of it all is that roles are blurring.
At Yahoo, Jeff Passan wrote last week about the trend in pitcher usage towards less defined roles. He quoted the man who convinced Rich Hill to lean on his curveball as his primary pitch, a voice willing to test tradition.
“I almost wonder if we’re getting to a point,” Red Sox VP of pitching development Brian Bannister said, “where roles aren’t defined. I’m not a starter. I’m not a reliever. In the postseason, I get outs.”
There were some interesting opinions circulating the Twittersphere last week regarding whether this pitcher usage is applicable to regular-season play.
The postseason game is very different, of course. There are more off days, there are greater stakes, there are fewer games with which to be concerned, and sometimes there is the prospect of no tomorrow. We should expect the game to be warped here. We should anticipate October to be a time for experiments. There’s often a greater need to try new strategies and depart from accepted practices.
While we should expect the postseason game to be different, the interesting question is whether teams and managers willing to test limits in the postseason will be willing to employ ideas like these in the regular season. Or whether they will soon be forced to move further away from traditional labels in the regular season.
This author wrote last week that bullpenning of some sort represents the future of the game, whether managers and players like it or not. Starters accounted for the lowest percentage of innings in the history of the sport this season. There were only 15 starters to reach the 200-inning threshold in each of the last two seasons, an all-time low. Moreover, only 35 major-league starters reached 180 innings. Just 45 threw 170 innings.
In 1990, 42 pitchers reached 200 innings, 69 threw at least 180 innings, and 83 reached 170 — and that was with four fewer franchises.
Because of the types of athletes being developed, because of the stress on arms in the modern game and the specialization of the sport, the labels and roles assigned to pitchers are becoming less defined.
Said Pirates general manager Neal Huntington to this author and then newspaperman back in spring of 2016:
“Could there be something five, seven, 10 years from now, where instead of five starters trying to go seven innings, you have two trying to get to the seventh inning and (three) trying to get through five?
“That’s where that guy who has two pitches, that’s always been a gun guy (fits). … He can’t give you 100 good pitches, but he can give you 50.”
There’s still room for an ace to pitch deep into games. There’s still a place for an outing like Verlander’s on Saturday night. It’s a classic look that won’t go out of style easily. But there are fewer and fewer workhorse aces like Verlander. There are more and more pitchers being developed to fit less defined roles — pitchers who, when not compelled to face a lineup two or three times, can be very effective. Even Clayton Kershaw was not permitted to see the Cubs for a third time on Saturday night.
The 162-game season presents a different challenge than the postseason — namely, that workloads must be managed. But I do wonder if most pitchers are eventually going to be interchangeable in roles and labels.
With the advent of the 10-day DL and increasing tendency of clubs to use the last bullpen spot as a revolving door to bring up fresh arms, it’s possible that the 200-inning arm will soon become extinct and that the majority of staffs will be occupied by 100- to 150-inning pitchers. The game is already moving there.
Perhaps we’re moving toward a future where there are a few aces but generally no relievers, no starters. Just pitchers.
A Cleveland native, FanGraphs writer Travis Sawchik is the author of the New York Times bestselling book, Big Data Baseball. He also contributes to The Athletic Cleveland, and has written for the Pittsburgh Tribune-Review, among other outlets. Follow him on Twitter @Travis_Sawchik.
It would not surprise me if some team started using tandem starters with the 4th/5th starters very soon. Teams need about 10 starters. The 6th through 10 reserve starters would need to have options so they could be bouncing back and forth from AAA to majors such that there are still enough roster spots for relievers that stay up in the majors.
In various ways, this is the question that keeps popping up all the time. How do you roster enough pitchers to use them in increased roles while limiting the numbers of innings that the traditional rotation pitchers have thrown in the past. I doubt if more than a few teams could find even a couple of decent starters who have options left. If starters are now only going to give a team 700 innings a year, then relievers will have to throw those innings. There are not many who can pitch 70 games and 100 innings out of the pen and then you are down to the 20th guy on the depth chart. Expanding the rosters would only exacerbate the time and pace of game problem and I already find 13 man staffs appalling.
We’re going to have to find a new definition for “qualified starters” if this happens. As it is, most leaderboards tend to be filtered by pitchers who qualified for the ERA title, but if hardly anybody is throwing enough innings anymore, those leaderboards won’t be nearly as meaningful.
Guessing when number of qualified starters drops below 40, they will make a change.
Averages of qualified starters in 10 year blocks:
2008-2017: 82
1998-2007: 88
1988-1997: 83
1978-1987: 87
1968-1977: 83
1958-1967: 65
1948-1957: 58
1938-1947: 62
1928-1937: 67
1918-1927: 70
For just the last 10 years, here are the season-by-season counts:
2008: 88
2009: 78
2010: 92
2011: 93
2012: 88
2013: 81
2014: 88
2015: 78
2016: 74
2017: 58
the more recent blocks are a bit propped up by expansion [more innings to go around]. Going from 4-man to 5-man rotations presumably also helps bump up the counts.
Maybe teams will go towards 6-man rotations [or further MLB expansion], which may drive the numbers back up again?
And the 5 IP win requirement. How many times do you see a manager with a big lead leave a gassed pitcher in there in the 5th inning just to get him the W ?
I don’t like the death of the starting pitcher, but it looks like that is whats happening and at an increasingly rapid pace. Based on Fangraphs data of starting pitchers (started with 1974 as that appears to be as far back as they clearly separate SP data) we are at an all time low for IP/GS for starting pitchers. I used only pitchers that started 10 or more games in a season. In 1974 the average IP/GS was just over 6 2/3 and in 2017 it was just under 5 2/3. The big drop-off to 6 IP/GS started around 1999 and was steady until 2015. Then it dropped 5.9 IP, then 5.75 in 2016 and now 5.61 in 2017.
That’s not even the worst part. Look at that data on how quickly the pitcher gets pulled. 1974 93% of the SP’s averaged over 5 1/3 IP/GS. From 1975-2015 that percentage ranged around 80% – 90% with 2014 being just below 91%. In 2016 only 75% of pitcher averaged more than 5 1/3 IP/GS and in 2017 it took another huge dip to 68%.
Just to validate the data I ran the numbers including all starters regardless of number of starts and it was even more dramatic. In 1974 the number was 69% of starters averaged 5 1/3 or more per start and in 2014 it was 67%, 2015 59%, 2016 49% and 2017 47%.
It really looks as if managers do not want to let the other team face a SP a third time around and are pulling them after the 5th.
Regarding Huntington’s quote, are we really that far away his idea right now?
If I’m using the Fangraphs split tool correctly, regular season starts for 2017 averaged approximately 5 1/2 innings. That number has been trending downward fairly steadily (which Travis notes) and it was 5 9/10 innings as recently as 2012 (so a slightly over 1 out per start decline in just the past 5 years).
Granted, this average is showing how long starters actually pitched, not how long the manager might have hoped that they pitched at the start of the game. The average is also picking up starters who left games in early innings due to ineffectiveness or injury. The average length of a start, however, seems to be inching ever closer to 5 innings.
For that quote, some teams are there already or close. Astros typically had 1-2 pitchers going about 7 and 3-4 guys going for less than 6 for most of the season. However, Astros had a deep pen until 2 of their relievers fell apart which may or may not have been due to workload.
I think the first sentence of the last paragraph is the key. The Chad Kuhls of the world get removed after 5 innings because they aren’t very good, but are still be left in when they’re cruising, even when evidence shows they should be taken out.
The postseason is a sprint while the regular season is a grind. I am not a proponent of keeping things as they are simply for the sake of tradition, but you do have to keep this perspsective.
During the regular season, you play 162 games in 182 days. You can only have 12-13 pitchers in your roster and need to have at least 5 available for any particular game, and so decisions which may seem suboptimal strictly from the perspective of an individual game start to make more sense in the larger view.
Managing a pitching staff is about balancing health with quality and quantity; you need to strike the perfect balance between all three to have a winning team. Realistically, any sort of radical departure from the current status quo would require to teams to come up with possibly twice as many pitchers as they currently have who are capable of throwing 80-100 effective innings in a given season. Something tells me that if teams were able to do this, they would have already.
That’s why it’s an evolution and not an overnight change.
With the current starting pitcher paradigm, a starter goes 5 games/days between appearances. It may very well be that a starter conditioned to throw only 3-4 innings might need only rest 3-4 days. I don’t think anyone knows that yet. Instead of planning for 30+ 7 inning starts, maybe managers will start to plan for 40 shorter starts from the same pool of players.
Also I wouldn’t be terribly surprised if we see a roster adjustment in the next decade, something akin to the healthy scratch in hockey. It would allow for teams use to effectively roster 26-27 players and designate a couple of them to simply not play that night. Rather than playing games with the 10-day DL or running a shuttle service to the AAA squad.
It isn’t that some teams don’t have the guys, it is about getting around the 25 roster limitation. Dodgers used DL this year, but that was more about giving guys more rest instead of limiting innings during starts. Other teams have cycled guys back and forth to AAA to give guys rest. I think at some point these tricks will be used instead of just to give guys rest, but to limit pitches thrown per outing for the bottom of the rotation guys as well.
The Dodgers still followed a fairly conventional pattern of usage; most starters averaged >5 IP per start, and most relievers averaged an inning per outing.
I think the problem is that whether you call them relievers or starters, there are only so many good pitchers you can allocate innings to, period. Take a look at the Rays pitcher page; excluding their top 6 pitchers, the remainder of their staff threw 723 innings and combined for 3.3 WAR; under this model, they’d either need even more pitchers at MLB level, or some existing mediocrities to throw more innings. Clearly, those extra pitchers don’t exist or else they’d have seen action, and it’s hard to see how the existing bottom of the staff could have done better with a different inning allocation.
Building on kuri’s point, I think part of the way to do this (with current rosters) would have to include something like using a lot of relievers who pitch less often but typically pitch for 2 innings at a time rather than one. It’s very difficult to get relievers up to 80 innings per season, never mind 100 innings, with current usage patterns.
I think there’s still some question how to balance that out – barring an idea like HappyFunBall’s healthy scratches – because the 10-day DL and options to AAA (which have to last at least 10 days before recall) are fairly crude tools to manage workload for relievers. I don’t know that riding relievers really hard for a couple weeks and then giving them 10 days to recover is a recipe for effectiveness or health. There’s also the question of how to manage it in light of game situations and leverage if usage patterns dictate “2 innings in a game and then time off” and come close to prescribing in advance which days pitchers have to pitch, using pitchers regardless of platoon disadvantages, etc.
I agree that what the Dodgers did, at least with starters, was really more a way to spread innings across the season because they had more than 5 legitimate MLB starters, at least when all or most of them were healthy. Using the 10 day DL strategically can allow a starter to skip only 1 turn in the rotation when off-days line up correctly, so it works well with that particular plan.
The top 7 pitchers by starts for the Dodgers all averaged over 5 innings per start, though only Kershaw averaged 6 or more. Those 7 combined for 151 of the Dodgers’ starts this year. All but 1 of those 7 starters were on a rate basis well over 2 WAR / 180 innings this year, and Ryu was below that but not a train wreck (about 1.2 WAR / 180 innings).
http://www.fangraphs.com/leaders.aspx?pos=all&stats=sta&lg=all&qual=0&type=8&season=2017&month=0&season1=2017&ind=0&team=22&rost=0&age=0&filter=&players=0&sort=7,d
So part of what makes the Dodgers’ 2017 usage difficult to replicate is that “have 6 good quality MLB starting pitchers and 1 below-average but OK starter” isn’t an easy roster to replicate. It stands out even more considering that 2 starters expected to be in the mix aren’t even on that list of seven due to injury (Kazmir and Urias), though Darvish is one of the top seven and was added during the year.
Creativity on this topic is interesting, but there are real limitations from factors including the 25-man roster size, how many effective innings relievers can pitch, the number of talented enough pitchers available, and not wanting to have many pitchers with reliever profiles face batters twice or face a lot of opposite-handed hitters.
Meh. It’s not really that Kershaw wasn’t allowed to see the order for the third time. He was PH for in a high-leverage spot. There’s no indication that Roberts was going to remove Kershaw otherwise. Don’t shade small things to support your narrative, or else you’ll lose some credibility.
Moreover, the use of starters in bullpen roles is explained by context, especially their use in elimination games (Verlander, Sale, Quintana, Scherzer, Lester, etc.). In contrast, basically, there are no similar “elimination” games in the regular season (except, perhaps, for one or two teams in the last week of the season who are about to be mathematically eliminated from the playoffs).
While I do like to see the pitcher hit now and then, I think it’s a travesty when they have to pull a pitcher in a high-leverage spot at the plate. There shouldn’t be a situation where you have to consider pulling the best pitcher over the last 7-10 years for a PH because he is a career hitter with a .157/.198/.180 slash and a 25% K-rate. The fact that of the 119 pitchers that had 10 or more plate appearances, only 6 had a BA .250 or above, 4 with OBP .287 or above and 9 with SLG .333 or above should make everyone want to allow the DH in the NL. As a whole in 5277 PA’s their slash was .125/.157/.163.
Players and managers all say that one of the most important things about managing a pitching staff is that the pitchers know their roles. And that will be difficult to change.
In some regards I think we’ve gone too far the other way. Bringing in a regular season starting pitcher with little to no experience pitching back to back days or pitching after abbreviated warm up in the bullpen when you have better qualified relief arms is not the right answer.
It wouldn’t surprise me if in a couple of decades it’d be rare for pitchers other than stars to go more than two times through the order. Using most current starting pitchers for three innings or so every three days would allow you to leverage your starters as long relievers and still get 6 innings each game from your “starters” (using a six man rotation). This also allows for avoiding the TTOP.
That might be an idea that works, but there are some challenges to figure out with it.
This idea has tandem starters going on 2 days’ rest with the result that each would throw about 162 innings per year (barring injury). As HappyFunBall says above, it might be the case that pitchers conditioned to pitch this way can do so, but I don’t think that we yet know what impact (if any) there would be on health and effectiveness from this usage pattern of shorter but more frequent outings. When we take it to the extreme with relievers, we know that that they max out around 70-80 innings per season if their standard usage pattern is an approximately 1-inning appearance.
We’re also still requiring these pitchers to face hitters more than once per game, which means that we’re still going to need pitchers who profile more like at least mediocre starting pitchers than like most current relief pitchers. An issue for most of the latter is that they don’t have the pitch repertoire to face hitters more than once per game.
I think the problem with that approach is that the math doesn’t allow much room to adjust for leverage or to give the staff a rest after a tough stretch.
In other words, if you’re building a staff of 12 guys who all go three innings, you basically need to plan it out weeks in advance (try it on paper, just for fun) and stick to the formula regardless of the game situation.
There are trade offs everywhere, but I do think the current arrangement affords teams to allocate their innings in the most optimal fashion across a long season, at least from a very high level overview.
I think the takeaway from the last 2 years is not that labels have disappeared as much as the regular season / post season divide has widened.
It used to be that the only dramatic change a manager would make for the postseason would be to start a guy on short rest.
Perhaps a starter would come in for a game 7 or Mariano would pitch more than 2 innings, but even those were rare events.
Now some managers are experimenting, and as soon as one team succeeds with it, others are bound to give it a shot.
Now things have changed in the playoffs, but don’t expect regular 3 inning starts out of #1 guys next year. It’s not happening just yet.
More info on Starting Pitching. I pulled all the gamelogs for Starting Pitchers since 1957 (via http://www.baseballmusings.com/cgi-bin/DayByDayDatabase.py).
I think the most concrete evidence on the pending demise of the Starting Pitcher role is in number of starts that don’t go beyond 5 IP. Basically the early 1970’s had the lowest % of starts that were 5 IP or less (24.23% in 1972). Most other seasons ran between 28% and 32%, but in 2016 it jumped to the highest at that time (35.75%) only to be followed by this season at 39.04%.
Just in case you thought maybe this increase was caused by bad pitching, the actual ERA of the starting pitchers that only lasted 5 or less IP was it’s 2nd lowest ever (8.25) and had the 2nd highest % of starts where they allowed 2 or fewer ER in those starts (30.23%). If you can believe this, there were a total of 558 starts in 2017 where the starting pitcher gave up 2 or less runs and was pulled before the 6th inning. That’s 11.8% of all starts in 2017. No other season was higher than 9% until 2016. That means that managers are really buying into not to letting the lineup see the pitcher for the 3rd time and it’s only going to get worse from here on out.