If you felt like relievers were getting special attention this offseason, you weren’t imagining things. In a 2017-2018 winter that featured a free agent freeze the likes of which we haven’t experienced in recent memory, relief pitchers were the one position group that hardly seemed to suffer. They flew off the board remarkably early in comparison to the rest of the free agent crop this season; 17 of the 30 seven-figure major league free agent signings to occur on or before December 15th, 2017 (an admittedly arbitrary date) were relief pitchers.
I’ll follow that statistic up with the obvious disclaimer that one offseason doesn’t necessarily set a trend. But the buyer’s frenzy that took place in regards to relief pitchers this past offseason is in line with a startling trend in today’s baseball climate: reliable relievers are a more valuable commodity today than they’ve ever been before.
Notice the qualifier reliable. I’m not suggesting that the Warwick Saupolds and Alec Ashers of the world are suddenly any more valuable than they would have been five years ago. But the upper echelon of relievers, the ones who can be relied upon to come in the game and consistently get outs in the late innings over the course of a full season, the value of those relievers relative to other positions has increased from what it was in years past.
Of course, baseball is a game of context, and the word reliable doesn’t mean anything without tangible statistics assigned to it. Fortunately, the echelon of relief pitcher I’m talking about seems to have clearly defined itself across the past several seasons.
But before I get too much into those statistics, it’s important to set the context of this analysis by pointing out a clear trend in baseball: starters are pitching fewer innings than ever, leaving relievers to shoulder the remainder of the workload. Below is the number of total innings thrown by the starting pitchers in MLB games in the past three seasons…
2014: 28,992
2015: 28,223 1/3
2016: 27,412 2/3
2017: 26,787 1/3
With the starters getting quicker hooks, MLB relievers have seen their combined workload increase by 735 innings per season since 2014. Because of this, MLB bullpens were forced to handle an average of 73.49 more innings per team in 2017 than they had to in 2014. It looks like we’ll be seeing yet another decrease in total innings pitched by the starters this season; they’re on pace to throw about 26,542 total innings in 2018. With the way things tend to work in September, I’d be willing to bet that innings total will end up being even lower when the season comes to a close.
As one might expect, the number of qualified relievers last season reached its zenith in the modern era (155). But the number of relievers to throw at least 60 innings has remained within the same range across the past decade or so. There were 84 such pitchers in 2017, 85 in 2016, 79 in 2015, 82 in 2014, 93 in 2013 and 88 in 2012. So while we’re seeing bullpens shoulder larger workloads on the whole, we aren’t seeing an increase in the number of workhorse relievers who are able to remain healthy or hold down a job for the bulk of the season.
If an innings threshold doesn’t do it for you, perhaps an overall measure of effectiveness will. WPA, or Win Probability Added, is a measure of how much value a player has provided to a team based on performance in each plate appearance (or batter faced, in this case) in relation to the leverage of those situations. Though there’s been a significant uptick in the number of relief pitchers who accrued a WPA of at least 1 in each of the past several seasons, the number of relief pitchers who’ve managed a WPA of about 2 has remained largely the same. Here’s the breakdown by year (past five years) of pitchers who’ve met that 2.0 WPA mark…
2013: 25
2014: 23
2015: 24
2016: 23
2017: 23
It’s hardly a coincidence that almost every single one of the relievers to accrue 2.0+ WPA in a given season also threw at least 60 innings in that season. So while “reliable reliever” is a somewhat nebulous label, there are clear indications that we’re seeing an increase in the number of reliable relievers needed to make a complete ballclub, but not an increase in the number of reliable relievers in MLB on the whole.
While the above milestones are admittedly somewhat arbitrary, the fact that they’re holding so steady across a period of five years is probably not. The fact that there’s a need for more talented bullpen arms doesn’t necessarily mean that more of them will just suddenly appear. That would likely require a dramatic change in how teams draft and develop players, and it seems unlikely teams would place any additional emphasis on developing pitchers as relievers when the main strategy seems to revolve around turning them into successful starters, and shifting them to the bullpen if that doesn’t work out.
So to recap, bullpens in 2017 were forced to take on an average of 73.49 more innings than they were in 2014, but they aren’t developing any additional high-end arms to compensate for that. It makes sense, then, to think that almost all of those extra innings are likely going to replacement-level or near-replacement level relievers. That works out to nearly an extra out and a half per team game put in the hands of a relief pitcher who may be an up-and-down- or waiver-claim-type arm. Obviously it doesn’t work exactly like that, but the core logic checks out.
Perhaps that’s why a higher percentage of the free agent dollars have been going to relievers on the market lately. According to data pulled from Spotrac.com, total reliever earnings accounted for an average of 28.98% of free agent dollars spent across the past two offseasons. That’s a remarkable upgrade over the four offseasons prior; relievers averaged a 19.54% share of the total free agent spending, topping out at a 21.51% pie slice in 2013. Last year saw 15 different free-agent relievers earn eight-figure guarantees and 21 earn multi-year contracts, both stunningly high numbers in comparison with years past.
It’s not just the free agents, though. Teams have paid handsomely on the trade market for elite bullpen arms in recent years. The Indians gave up a hefty package for Andrew Miller at the 2016 trade deadline that included top prospects Clint Frazier and Justus Sheffield. The second Aroldis Chapman trade brought back a top 10 prospect in Gleyber Torres, and the Cubs parted with the highly-regarded Jorge Soler in order to get just one season of Wade Davis. Extensions for Brad Hand, Kenley Jansen and Felipe Vazquez in the past two offseasons guaranteed significant numbers of years and dollars, too. While no one of these transactions is necessarily an abnormality in and of itself, the general pattern of these reliever valuations and more beyond them are in line with the trend of top-flight bullpen arms being valued more in today’s game than in years past.
At its core, this seems a simple lesson in the laws of supply and demand. With a greater need for relievers that can be depended upon for consistency and high innings totals, contending teams are facing a sense of urgency in pursuing a crop of those relievers that’s remained the same size. Logically, said urgency would figure to drive up the market value of those players in comparison with other positions.
As is always the case in the game of baseball, things could change quickly. For instance, with superstar position players like Manny Machado, Bryce Harper and Josh Donaldson set to hit the market next year, it’s incredibly unlikely that reliever contracts will account for over a quarter of free agent spending, even with Craig Kimbrel, Andrew Miller and Cody Allen set to join the pool. Still, I’m willing to bet that the latter three end up with hefty paydays, with a handful of others surpassing expectations as well.
Bryan Shaw($27M), Jake McGee($27M) Tommy Hunter, ($18M) Anthony Swarzak, ($14M) got more $$ than Moose ($6.5M+ $2.2M incentives)
They considered “reliable” the golden age for Relievers.
This is* the Golden Age for Relievers.
Meanwhile Jared Hughes got like $4m total
And SP Lance Lynn (though he sucked out of the gate) got “only” $12M. (+ $2M incentives for 180 IP)
What does an Andrew Miller contract look like next year, considering he’ll be 33 in a couple of days?
Considering Melancon was given $62 million for his age 32-35 seasons’ (with an Opt out after year 2 ) and Miller is twice the reliever Melancon is.. Min. $62M for Miller.
I’m expecting around the 3 years and $52 million that Wade Davis got (maybe a little more, maybe a little less). I think he could potentially get 4 years, but probably at a little less per season.
With the exception of the Melancon contract, which is full on nutso for the Giants. How it’s even structured is terrible and puts them at a $20m luxury tax hit. “More valuable than ever” is relative and I’m interested to see his value as the trade season begins. Considering the Giant’s will most likely not be a contender, I’m torn between what’s best for them: Find a taker with straight salary dump or try to buy it down to get a prospect or dare I say, one terrible contract swap for another?
Also have to factor in how many innings and appearances relievers make. Are they more valuable than the pinch hitter that appears every couple games in the NL? There are guys that perform year in and year out, and then there are the stock market guys that go up and down. Some GMs may opt to take the low-risk route rather than tie up cash that can be allocated elsewhere.
Developing your own allows a team to have the player for six-plus without the exorbitant rates.
Theoretically, yes, but you didn’t factor this is the most volatile position in baseball. (Probably all Sports) hence the premium price-tag for reliable vets.
Get 12 usable arms to Advanced-A, and you’re likely better off than if you only have four or five.
Developing pitching is a very inexact science, but buying them on the open market tends to be expensive and unrewarding.
loved the article Kyle, keep up the good work!!
Best combination of writing and statistical analysis at this site in some time.
But as you know, Kyle, I’m always looking for holes. The number of quality relievers per season appears steady. But are these pretty much the same people every year or is there a lot of variation? What I’m getting at is, how does the 2.0 WPA gained by the relievers compare with the (presumed) loss of WPA by starters—wouldn’t it have to go somewhere? I recognize the context orientation of WPA and understand that we can’t assume starters and relievers would face the same situational risk. But by shifting IP from starters to relievers aren’t they also almost guaranteed to similarly shift WPA?
I’m trying to distinguish between actual quality relievers and those who 1) weren’t awful, 2) were a bit lucky, and 3) were therefore used enough to compile a bunch of counting stats.
Once again, this was very nice work.
Thanks for the kind words, majorflaw. In my research, I noticed that the number of relievers to accrue at least 1 WPA increased each of the past three years, This would fall in line with the WPA “points” shifting from starters to relievers. A quick look at league stats across the past four years confirms: relievers as a position were worth about 60.34 WPA in 2014, 62.36 WPA in 2015, 69.96 WPA in 2016 and 73.31 WPA in 2017.
If I had to make a hypothesis, I’d say that part of this is increase correlates directly with the increase in reliever innings, as WPA is a cumulative statistic. But the increase seems pretty dramatic as a percentage, which makes me wonder if there are just better relievers in MLB on the whole. More innings means more experience for more relievers, and more opportunity to learn from that experience and improve.
That said, the fact that 2.0+ WPA relievers has remained so steady over the last five years means that there’s probably somewhat of a talent component involved. So while there are more relievers getting a shot at that 1.0 and 1.5 WPA mark, much of those performances could simply be flukey. 2.0 is much harder to achieve, though, so over a large sample size it’s a much more difficult benchmark for a pitcher to simply luck into.
As far as the variation aspect, there are definitely some repeat names, as you’d expect. But plenty of new names make the list each year (Josh Hader, Shane Greene and Archie Bradley were new additions to last year’s list). Part of it also could be based purely on the leverage situations these players are pitching in; I wouldn’t have expected to see Alex Claudio’s name on any list of the top MLB relievers, but he achieved 3.33 WPA in 2017.
“More innings means experience for more relievers, and more opportunity to learn from that experience and improve.”
Also more opportunity to blow their arm out, no. Is there any statistical basis which distinguishes this ‘improvement due to experience’ or is it just to be assumed? I’d like to see that improvement charted right next to the player’s normal aging pattern and see if the latter is being misread as the former.
“2.0 is much harder to achieve . . . It’s a much more difficult benchmark for a pitcher to simply luck into.”
It can’t all be luck, obviously. But how do you explain the 2017 version of Matt Albers? There was some luck involved in him even making the team. (BTW, what was his WPA last year—bb-ref either doesn’t have it or it isn’t obvious). I’m looking for a way to distinguish between your reliable relievers and Matt Albers, if you have one.
Thanks for the thoughtful response.
The improvement due to experience is just assumed. I wouldn’t even say assumed, it’s more of a theory without any real statistics to back it up. I’m actually at a loss to explain why exactly 2.0 WPA seems to be such a magic number, while 1.0 WPA seems to be obtainable simply by a percentage of the total relievers.
Matt Albers’ WPA last year was 0.67, by the way. His career high is 1.24.. 2017 Matt Albers is largely a product of BABIP and soft contact.
“The improvement due to experience is just assumed. . . . it’s more of a theory without any real statistics to back it up.”
Which means it could be real or could be illusory. Not sure what to look for, perhaps some Tanner Roark types who accrue that experience while riding the downside of the aging curve; do they improve as they gain experience or does their performance roughly follow the same aging patterns as pitchers who entered the league at a much younger age?
“I’m actually at a loss to explain why exactly 2.0 WPA seems to be such a magic number . . . “
That’s OK. Finding a signal amidst all the noise created by new metrics is an ongoing process. Starting with questions rather than (presumed) answers is a great first step.
“Matt Albers . . . 2017 is largely a product of BABIP and soft contact.”
Thanks. Actually your system was better at distinguishing between Albers and an actual quality reliever than anticipated. Wasn’t expecting that result. Good job once again.
Reliever innings increased from roughly 33% to 38% of all I innings from 2014 to 2017, or about 15%. WPA actually increased less than that. So there is no indication of an increase in average relief quality.
‘Extensions for Brad Hand, Kenley Jansen and Felipe Vazquez.. ‘
Technically, Jansen didn’t get an extension but got signed back to his former team from free agency.
I just hope that the concept of a ‘starter’ doesn’t go out of baseball entirely. 6 pitching changes per game makes it hard to watch. Luckily for us, the 3-inning-pitchers experiment by the rockies a few years ago didn’t really go that well.
I think we will see much more of this, honestly – 4 and 5-inning starters. Sports evolve – look at the NBA and how teams have moved away from the mid-range jumper because analytics showed conclusively it’s the least efficient way to score in the game. Or even the NFL – in the beginning of organized football the forward pass was outlawed. For a very long time, in fact.. Sports change with new information.
I’m NOT saying it will happen, but definitely a possibility that Kimbrel becomes the first reliever to sign a $100 million contract. I think though it is more likely in the $75-85 million range and not of a long enough term to justify the bigger number.
If you account for baseball inflation, I’d say the BJ Ryan contract way back in 2006 ($47/5yrs) would be much above the $100mm mark today.
Familia and Blevins should pull in a good prospect each. Too bad Ramos blows goats.
Is mlbtr just realizing this? Watch playoff baseball. Games are won and lost on relievers these days.
Sir Anthony Dominguez with the six out save for the Phillies in St. Louis today.
The Phillies are all done on pitching with more prospects on the way.
Manny Machado and Bryce Harper suiting up for the Phillies in 2019.
“Sir Anthony Dominguez”? Never heard of him!
Just kidding. His real name is Seranthony Dominguez. Two strikeouts, no hits, and zero walks in that six out SV Not bad at all, Impressive numbers Casey.
The Padres will be spreading the wealth at some point before the trade deadline……Brad Hand,Craig Stammen,and Kirby Yates are all worth more in trade then they are to a 90 loss team.
Kyle, every time I see more of your work I’m even more impressed. I’m always honored to have been on the same staff as you several years ago.
This is what the angels need to do next offseason, Rebuild the bullpen and spend money on it. Line up is set , I think the rotation is underrated but good enough to keep them in games. If they could add a lights out closer and set up guy, they would be dangerous.