Back in 2016, Terry Francona’s usage of Indians left-hander Andrew Miller was revolutionary, and the way he deployed his best relief pitcher, particularly during the postseason, has since had a profound impact on the way MLB teams have used their bullpens. The Andrew Miller Effect changed the game of baseball, and was a fascinating story to watch.
Except that’s wrong. Or at least, it’s the wrong way to look at the story. What we sometimes call the Andrew Miller Effect isn’t actually a story in and of itself, but rather a single chapter in a longer novel that has yet to reach its conclusion. That novel doesn’t begin with Miller, either, and it’s not even really about relievers. At least, not as much as you might think.
In the simplest terms, a team wins a baseball game by scoring more runs than the opposing team. So obviously, there are two ways for a team to improve its chances of winning: get better at scoring runs, or get better at preventing opponents from scoring. The latter objective placed within the confines of baseball’s nine-inning, three-outs-per-inning format outlines a modified objective: the pitching staff must somehow get 27 hitters out while allowing the fewest runs possible. The only real limitation on the pitching staff beyond that is that a pitcher who is removed from the game may not re-enter.
Baseball is a game largely centered around one-on-one matchups between a pitcher and a hitter. And since the hitters must continue to bat in a pre-determined order unless replaced by another hitter, the team that’s trying to get outs in a given half-inning has far more flexibility in gaining matchup advantage. In addition, with the way rosters are usually constructed, a team has the facility to change pitchers 11 or 12 times in a game, while a batter can only be swapped out three or four times total.
The conclusion here is that teams have always had enormous incentive to get creative in the way they deploy their pitchers. It’s not an entirely new concept; teams have been using LOOGYs (Lefty One-Out Guys) against left-handed hitters for years because the pitcher has a distinct, proven advantage in such a matchup; it’s just one way of increasing the chances they’ll get an important out. Similarly, Francona using his best reliever in situations with runners on or where the opposition’s best hitters are due up is all about finding ways to get the difficult outs with the highest probability and bridging the gap from zero to 27.
The Indians’ strategy with Miller was ground-breaking because it blurred the hierarchy of “middle relievers”, “setup men” and “closers”; in some ways, the roles of Josh Hader, Chad Green and more are products of the Andrew Miller Effect. The Rays are now breaking ground by similarly blurring the lines between “starters” and “relievers”. If you’re reading this, you probably already know that Rays manager Kevin Cash has been using relievers such as Sergio Romo and Ryne Stanek to get the first few outs of a baseball game, then turning to his “starters” to come in after that.
The core logic behind the strategy makes plenty of sense. Romo as a reliever is probably better equipped to get outs at the top of the lineup than the second- or third-best starter in a thin Rays rotation. In addition, it means that the pitcher entering in relief of Romo will pitch to the weaker part of the lineup first; that means the new pitcher can be called upon to face more batters without having to expose himself to the most dangerous opposing hitters a third time, likely facing the bottom half of the lineup three times apiece instead. On the whole, the results of this experiment have been positive, which has everyone around baseball talking about the strategy and the Mets in particular considering deploying it on Monday.
It’s hard to imagine that the Sergio Romo Effect won’t have an impact as loud as (or louder than) the Andrew Miller Effect. It seems really unlikely that the strategy will just go away; as we saw with the Andrew Miller Effect, teams might hesitate to try something bold and unusual, but they’ll copy it quickly once they see it working for a rival club.
It’s still possible that MLB will step in at some point and write a new rule that limits this fast evolution of pitching roles. But if that doesn’t happen, we could eventually be looking at a version of baseball in which pitchers are defined by how many outs they’re typically called upon to get rather than in which part of the game they’re called upon to get them. At that point, we might have to entirely reimagine the labels we put on pitchers.
The roles of the truly elite aces like Corey Kluber and Max Scherzer seem unlikely to change very much. There’s little reason to disrupt the role of a guy who stands a solid chance to throw a complete game with brilliant results on any given day. But what if pitchers were used (and valued) based on a combination of the following five factors:
1) How efficiently can the pitcher get outs when throwing fewer pitches at maximum effort?
2) How efficiently can the pitcher get outs when throwing more pitches at an effort level that allows him longevity in the game?
3) At what point should the pitcher be pulled to prevent further exposure to the same hitters?
4) To what extent should the pitcher be shielded from his weak-side platoon?
5) To what extent should the pitcher be shielded from hitters who are particularly good at hitting the types of pitches he throws?
If pitching really is all about getting 27 outs while preventing runs with the highest possible efficiency (and it is), then the way a pitching staff is deployed might continue to become less of a formula and more of a jigsaw puzzle. That means shedding the labels of “starter” and “reliever” in favor of labels that describe hurlers in terms of the above factors. In fact, perhaps labels would end up entirely useless and it would prove a mistake to use them at all. In this hypothetical (future?) environment, it’s likely that pitchers would be valued based on their efficiency in the unique situations they’d be asked to jump into.
Kluber, for instance, is a fairly uncommon asset; he’s an elite ace capable of preventing runs while going deep into games. Taijuan Walker, on the other hand, is a good example of someone who had significant splits last season after facing a lineup twice through. In 2017, Walker owned a 2.68 ERA and .298 opponent’s wOBA for the first two trips through the batting order, making him a very useful pitcher. However, when facing hitters for the third time, Walker’s ERA and wOBA ballooned to 5.97 and .357, respectively. Would he have been more useful to the Diamondbacks if they’d capped his outings at 18 batters faced, perhaps with the added benefit of being able to rest him for fewer days between outings?
Meanwhile, Hader and Green are somewhat of a throwback to the Mariano Rivera-type reliever capable of performing at maximum effort to achieve superhuman results against six to nine hitters. Hader’s done that 12 times so far this year, while Green’s accomplished the feat in seven appearances. Pitchers of this ilk are about as rare as those of Kluber’s, and the ability to get so many outs with such an astonishing level of efficiency is an incredible asset to any pitching staff. Perhaps these players will set a blueprint for others like them in the near future; even pitchers who can perform at 70-80% of Hader’s capabilities for a single trip through the order would be useful pitching every other day or so. There are plenty of starters who’ve had dramatic splits between their first and second trips through the other. Mike Foltynewicz comes to mind as an example, who limited opponents in 2017 to a .233/.302/.348 line the first time through, but allowed an uglier .295/.391/.516 line during his opposition’s second look.
With more pitching changes per game, lefty or righty specialists could end up being more useful than ever. Maybe that guy with the nasty slider and a batting practice fastball could still find a specialized role getting out opponents who have difficulty hitting breaking balls. The Craig Kimbrels and Corey Knebels who come in to get three or four outs would have their place, too. If the starter/reliever template begins to crumble, the traditional five-man rotation and seven- or eight-man bullpen might crumble with it, leaving behind a roster format in which the number of outs a pitcher is capable of getting might not matter quite so much as long as he’s capable of getting the outs he’s asked to get with a rate of efficiency that justifies his roster spot. Each of the 30 MLB pitching staffs could end up being its own unique cornucopia of pitcher types cleverly assembled by its respective GM and used strategically and creatively by its skipper, the only rule being that it needs to prove adept at getting from zero to 27, game after game.
The question at that point becomes, how do we place a value on each pitcher in these new roles? What is the value of an average 100-pitch guy in comparison to an above-average twice-through-the-order hurler, and how do both compare to a guy like Ryan Dull who needs to be shielded from left-handed hitters but gets righties out nearly 80% of the time? If more teams begin to protect long-appearance pitchers from being exposed to the order a third time through, would the abundance and limited longevity of those pitchers make them less valuable as a group, or would their efficiency and flexibility within the format help elevate their value in comparison with 100-pitch guys and elite short-appearance pitchers?
The cop-out answer is that we’d have to wait to see all this happen in order to know. But it’s probably fair to think that teams would use stats like WPA to find the answer, or create entirely new stats to weight a pitcher’s efficiency against the number of total outs he’s tasked with getting in his particular role. It’s also pretty much a certainty that the market itself would have a say in the value of each class of pitcher. If twice-through-the-order type guys are abundant in a given year, teams may not be willing to pay as much for them. On the flip side, if many teams are in need of a once-through-the-order shutdown guy or a three to four out fireman to bridge the gap between longer-appearance guys, the cost of those players could increase based on supply and demand, much in the same way the value of a good second baseman goes up if more teams are lacking at the position.
The evolution of out-getting won’t simply end with the Rays’ latest experiment. There are clear advantages to be found in the creative deployment of pitchers that contrast heavily with baseball traditions, and with teams becoming more and more data driven, you can bet they’ll continue to search for more effective ways to get from zero to 27. Traditions aren’t rules, after all.
I balked at seeing Romo start but I have since come around to “trying different things” when it comes to winning. If it works….
Maybe 2 starters splitting a start. One goes 5 and the next goes 4 innings. 5 days time they reverse rolls. Limiting your need for extra pitchers on the day. Tandem starters.
Like the Astros used 4 times in the postseason.
Rockies tried it and failed miserably. Doesn’t mean it won’t ever work but there is a precedent of failure
This has to be one of the most well written and thought provoking articles I’ve read on MLBTR. kudos!
Agree. Thought provoking piece.
Agreed. This was a great read and very interesting to think about.
I can’t imagine most starting pitchers are going to like basically being told that they’re not good enough to face a lineup 3 times.
Then they’d better get good enough. There’s no room for huge egos on people that don’t pull their weight.
Yeah well expectations will change too.
If a 5-pitcher rotation has 2 spots reserved for guys who are basically limited to 4-innings, then their mindset will be less “this game belongs to me” and more “I have to give my team the best 4 innings of pitching that I have in me.” People want to be put in a situation where they can succeed and nobody is going to argue if the results are there. The league is only averaging 5 1/3 innings anyway.
Just think about how that can change their approach. They’ll still get chances to go a little deeper, if they’ve only seen 15 hitters through 4 innings and have a lead they’ll go out in the 5th. They’ll still get a chance to show they can be one of the 3 long-starter guys.
I think it had more to do with maximizing the productivity of the high-value pitches than anything else. If a rp has only 1 good pitch and an ok 2nd pitch how much harder is it to stretch him out enough to get 6 outs instead of 3? If a starter only has 2 good pitches and an ok 3rd, why put them in a spot to hold something back early in the game so they arent so predictable in the 5th, 6th, and 7th?
Better off having 3 traditional starters, 3 12-out getters, 2 up-to-9-out guys that are used like Hader and Green, and 5 spots for a mix of specialists and 1-2 inning guys.
A great article in explaining what is happening in MLB.
I quit watching MLB 2 years ago, and came back this year as I have some time. Signed up for the MLB Internet package, and found that I don’t recognize most of the teams from what they were 2 years ago. It’s fantasy league for real. There aren’t 4 teams worth watching. Most teams don’t have a dozen decent major league players on their roster.
I stopped watching the NFL 8-10 years ago. It was difficult to watch a team sport when 2-6 players on each team are running off the field each play and being replaced by others. Totally impossible to keep track of who all the players are, let alone who is matching up with whom.
Now MLB is being micro-managed by a bunch of geeks that were not good enough to play the sport. K’s are epidemic as those in power want the pitcher to K more and don’t care if the batter K’s. Then there are the shifts. Batting averages are abysmal. Balls are not being put in play.
So now we read that the game is about to become even more scientific and less of an art form. Players will continue to be programmed robots as so few can do more then one or two things well; hence rosters changing even more every single day. Games will continue to be longer, the percentage of balls in play will continue to go down, and the only ones loving what’s happening are those that play baseball on a computer. (Fundamental play at the MLB level is on the level of 10-12 year-old little league, and getting worse each year).
Sorry to see a sport I played from sunup to sundown slowly fading away as it becomes as enjoyable as listening to music composed on a computer. Baseball used to be about challenging players; seeing who made smart plays at the opportune times; who could do something out of the ordinary. As the spontaneity leaves and the bureaucratic complexity continues to take hold, the audience will leave as it’s impossible to follow what is going on. Just empty calories.
I, too, enjoy writing letters instead of email, motorless reel lawnmowers, kerosene lamps, and treating illnesses with leeches.
That’s basically what i got out of his rant.
Rant?
If you don’t think MLB knows that there is a major problem with the lack of balls being put in play, you haven’t been watching. Even this site has noted the problem numerous times.
10/10 old man rant
those damn millennials am i right
I hear and understand your frustration Carrington.
Understood. but things can change you never know
The analytics trend generally started with smaller market teams that could never afford to compete on an even footing with the big money teams. So, they had to find new ways to build their rosters and utilize their players in ways that weren’t expensive. “Small ball” or “money ball” or “pitching & defense” teams that we were seeing pop up 15-20 years ago were responses to big expensive powerhouses with killer lineups.
Nowadays what’s changed is that the big money teams have been paying attention to how lower payroll teams have found success. They’ve gotten smart and figured out that if the bargain basement teams can find ways to get lots of value out of their dollars, so can big market teams, and they have more dollars to play with. That’s why we’re seeing these dramatic tear down/rebuild cycles in big markets, why we see teams seemingly waste money eating other teams albatross contracts… it’s a game within a game.
So now, that’s the paradigm – you have to use analytically constructed teams to compete or you’re not going anywhere. Whether that’s a good thing or a bad thing, or a bit of both… that’s the debate. I’ll agree with you that there’s a lot about it that isn’t always very fun to watch, and isn’t really in the “spirit” of the game.
There aren’t 4 teams worth watching? Are you serious? It sounds to me like you just don’t like baseball.
Sounds like that to me too. Especially since he said it
When are teams gonna start adjusting their lineup to The Opener? When TB is gonna start Romo, bat your 6-7-8 guys first, then the top of your regular order.
Not so fast. All the manager has to do is remove the opener pitcher after a batter but the top 3 hitters cannot change order or pinch hit for or else they are gone from the game. Oddly genius.
That is exactly what is going to happen.
Dueling front offices, as opposed to athletes competing. My bean counters vs. your bean counters.
The amazing thing about this computer generation is that they do not understand that life is dynamic, not static. For every action, there are multiple reactions.
Watching professional sports is supposed to be about watching the players competing by allowing them to play. With the obscene salaries, those players will happily do anything they’re told to. And why not? I sure would.
To many of us, watching dueling front offices is also entertaining.
A good bean counter understands these things are dynamic.
Teams won’t bat their poor hitters at the top of the lineup as those players become most likely to get an additional at bat on the last turn through the order.
There already is an understanding of the reactions, but I’m glad you’re having fun beating up that straw man.
LOL
This reminds me of when the crummy Cubs had their coaches rotating to manage one year.
The Rays are a terribly run franchise whose front office is so inept that they have 2 starting pitchers on their roster and not much of anything in their farm system.
Like anything else in life, sports have a balance. The balance can be tipped temporarily and show gains, but sooner rather then later it will prove to be short-lived. This reminds me of the screwball diets that come out every Spring – eat nothing but magpies and Cherrios and 88% of people are guaranteed to lose weight. Look – here are the statistics! And don’t call us in 2 years if you are walking around sick.
I, too, think numbers are the opposite of facts.
No, the Rays are a smart organization who aren’t allowed to spend much money by their cheap owners, so they have to resort to innovation.
Ya it’s easy to fool and just destroys the game a bit more
probably with this is that most of the reason modern teams bat their best players at the top of the order is because they are more likely to get an extra at bat on the last time through the order.
They’re not going to reconstruct their lineup (in this way) just to deal with an opener, they would just be playing further into the opposition’s hand
Tough to see anything like this taking hold league wide in the near future. Interesting read for sure, but teams are not going to buy in to such a radical change for a long time, if ever
If these pitching strategies are proving more effective than the status quo, then teams will buy into them quick enough or find themselves at a disadvantage.
Maybe it’s just me, but I feel like the whole pitching usage thing turns back to the 2014/2015 KC Royals. With Herrera, Davis, and Holland, they showed in the end of games, they almost have the game completely wrapped up. And that’s something teams started replicating.
They definitely weren’t the first team to have setup guys as good as their closer but there was something about it, how obvious they were about using them in that formula.
The Rockies did something similar to this in 2012ish, except that their experiment ended in dumpster fire. I suppose kudos to Dan O’Down and co.
I think the biggest obstacle to this are the pitchers themselves. If Romo and whatever starter don’t buy in, the Rays can’t do it at all. And a lot of pitchers are not going to appreciate having “real” starts taken away, or closing opportunities taken away.
It will happen eventually but I think the players will slow things down, cause most pitchers are old school.
Next step would be if this happens alot, start rearranging the lineups.
Won’t work. Lineups can only be rearranged prior to the game, and the manager can simply yank one opener in favor of another in response.
But once the lineup is turned in, that pitcher is bound to start and face at least one batter.
Yep, and then that pitcher can be pulled after facing one below-average hitter, and another can come in and face the others. Meanwhile a team is stuck with its 6-7-8 hitters atop the order gaining more at-bats while the best hitters see fewer at-bats in that game. Advantage: pitching
Zack Cozart criticized the Rays for starting a reliever, saying it was bad for baseball. I wonder how he feels about defensive shifts 😛
the same I would think. He is the kind of player who would have been alot better without them
This third time through the order I am not buying all that much. For one thing these batters have faced the pitcher dozens of times. I think the major reason results are worse the third time through the order is the SP is just more tired. It is very hard in any physical activity to pace yourself so you are as effective in the third part as in the first two parts.
I’m sure that’s an element of it, but stuff like getting zeroed in on a guy’s release point is not something that generally carries over the next time you face a pitcher. That’s been an observable effect going back to when we had 8-team leagues and everybody faced everybody over and over and over again.
You don’t have to buy it. The numbers are there regardless, and teams will build based around those numbers.
I know I’m not alone in this… for those of us who were into saber stuff long before it was cool find this stuff rewarding after all those years of complaining about how silly it is to rigidly use your pitchers based on what inning it is, rather than the scenario, losing games in the 7th because you just gotta save your most dominant pitcher with the bases empty in the 9th.
put down the computer and watch the game. it’s about fans and people. not about robotic performance and striking out every at bat. I am sure you don’t notice much from stat sheets and box scores
I wonder if a team could try going back to a 4 man rotation when only asking the pitchers to go through the order twice.
I think this will be especially useful for smaller market/revenue teams, who don’t have five good starters. I know in the minors they do tandem starters and always wondered why it wasn’t done in the majors.
It strikes me as a bit hysterical to chalk up the direction of the way baseball strategies are changing to the incorporation of sabermetrics. Most of these changes are hardly new: I remember old timey baseball people talking dismissively about pull-happy hitters and specialized closers in 1990s, claiming how, in the old days, they would do X and X to counter these things. In many ways, a lot of these changes are exactly X and X, claimed by old timey baseball people. The difference is that these changes are backed up by numbers and carry a greater sense of “authority” and “certainty.” But are these changes eternal “right answers”? If one is facing a Tony Gwynn esque hitter, skilled at hitting singles in the opposite direction frequently, defensive shifts become a liability. There were good reasons why set closer role became dominant at one time. I would imagine that many of these wheels would be rediscovered, except, this time, these rediscoveries would be backed up by data.
As I see it, the only thing that use of data achieves is that the exploitation of the apparent weaknesses (and thus the cycle of adaptation) would be followed much more ruthlessly and systematically. In a sense, it’s a self-licking ice cream cone that keeps going in circles: a lot easier for hitters to focus on being a pull happy one trick pony (or opposite field slap hitter–people who can do multiple things well are rare), so they are countered with defensive shifts; pretty soon, everyone will be going the opposite field making shifts obsolete,, then being pull-happy will be fashionable again, then shifts will make their triumphant return, and so forth. It’s been done before:…just that evolutions that took decades will now only take years.
You’re speaking in overly broad terms. Most players are not being shifted against to any extreme, only players who show extreme tendencies to hit the ball one way. You cite Tony Gwynn as being some sort of knock against shifting when obviously teams wouldn’t be shifting against him (or any other players like him) in the first place.
If the majority of those pull-happy hitters who do get shifted against were actually willing and/or capable of going against the shift, then they would’ve done it long ago. It doesn’t take years to adjust to seeing someone shift against you.
Most pull hitters do not want to have to deal with adjusting their approach and giving up their power for the sake of defeating infield shifts so most of them will continue to be shifted against. There isn’t going to be some broad trend of every pull hitter doing nothing but going oppo to defeat shifts to the point that shifting becomes ineffective and teams stop using it.
No, that’s not what I meant at all. What data achieves is that the tendencies of the hitters are captured much more precisely and the countermeasures against these can be deployed much more readily. One does not need to deploy an extreme shift if the hitter reliably (say 80% of the time) hits the ball at 10-15 feet to the right of 2b bag (and that batter can be LH, for good measure.) If that’s the tendency, there is no reason for an “extreme shift,” unless one were to mean by that a configuration where SS, 2b, and 1b are crowding the area between 5-20 feet to the right of 2b. Cartoonish, yes, but again, if the players are being predictable and most of their tendencies can be put in neat boxes, why the heck not, with regards coming up with “unconventional” approaches to take advantage of them.
In the days before data explosion, these tendencies were not captured with such precision. The better managers had a better approximate understanding of the tendencies which they put to use. But information was not precise enough to pull off the trick, for the most part. Now, the information is easy to come by. It is only rocket science (i.e. ug level physics) to take advantage of them.
But the real trick of people like Gwynn (at least the somewhat mythologized version) is that he was not quite so easily predictable. With enough data, you could come up with the best counter for Gwynn, but even that best counter would be wrong too often to be used too regularly: put 3 infielders to the right of 2b, he’ll hit the ball to opposite field with very high probability, etc., to use the overused example.–even if the “best” counter is to put the fielders to the right of 2b, at very particular positions. But the art of being “unpredictable” is lost. Too many players (both hitters and pitchers) can thus be put in too neat boxes to which a counter can be found that are hardly ever (relatively speaking) wrong, and this is partly driven, in turn, by the quest for “right answers” that players can stick to all the time (which rules out unpredictability). But most “right answers” can be countered. Every revolution runs out of steam in a few years, and we are losing something along the way.
No
A lot of interesting things in this article and can’t break it all down in a comment but the teams that don’t have good enough starters will attempt bullpen games. Bottom line is it all depends on the quality of pitchers each team has to how they handle their pitching staff. Having quality Starters as well as quality Relievers will always have the best chance to win regardless of what gimmicks are used to mask deficiencies of average pitching.
Not quality starters and quality relievers, nbresnak. Quality pitchers. That’s all a team needs to get outs.