When Aroldis Chapman was traded to the Yankees from the Reds (for mere pennies if you ask me) there was a lot of talk (including on this very site) about how the new Yankees bullpen stacked up against the dominant Royals bullpen of the past few years. In my view, the new Yankees bullpen projects to be better than the Royals bullpen was in either 2014 or 2015..
From 2014-2015
Wade Davis has been better than Andrew Miller, but Chapman has been at least as good as Davis, while Betances has been better than them all. When you account in the difference between the Yankees third best reliever (Miller) and the Royals second best reliever, you're talking about a 1+ win difference. Oh, and the second-best Royals pitcher on that list just had Tommy John surgery and will miss the entire season.
2016 Steamer projections:
As for the 2016 season, all three of the Yankees relievers project to be better than Wade Davis and of course better than Kelvin Herrera and Joakim Soria. Now you can quibble about Steamer being a bit low in Davis, an argument I would perhaps agree with, but I think at best you can say that he's as good as the Yankees "trihedra". The Yankees essentially have three Wade Davises.
The Yankees acquisition is a microcosm of the new macro trend that has been bandied about ever since the Royals won back-to-back AL championships and the 2015 World Series - overwhelm opponents with a dominant bullpen. This makes sense, regardless of the Royals success. You want every aspect of your team to be good. However, the Royals success brought this idea to the zeitgeist.
One thing though that stuck out to me when I saw teams trying to build a bullpen like the Royals - the Royals did not design to build such a dominant bullpen.
The Royals don't have a long history of having a dominant bullpen. When you remove 2014 and 2015 and look at the best bullpens (by fWAR) from 2004-2013 (prior 10 seasons) the Royals appear just twice.
Dayton Moore may or may not have made the bullpen a point of emphasis when he first arrived to Kansas City, but he had not succeeded in that plan until very recently. Considering that among Dayton's early acquisitions were guys like Yasuhika Yabuta, John Bale, and Juan Cruz, with guys like Blake Wood, Robinson Tejada, and Dusty Hughes throwing 60+ innings in a year, it is clear the Royals early pens under Moore were a far cry from H-D-H.
The Royals may have even backed into their dominant bullpen. Wade Davis was a former top 20 prospect with the Rays who failed as a starter. Then he was turned into a good reliever, but turned back into a failed starter (by Dayton Moore) before taking a reliever role again and becoming one of the most dominant relief arms in the game. Kelvin Herrera got a small signing bonus out of the Dominican, but was a failed starter turned good reliever.
Greg Holland was a tenth round pick who wasn't on Baseball America's top 200 pre-draft list and never made a top 100 list. He was really not very good in the minors and the Royals even tried giving him some starts in Wilmington his first season. Holland played SS/3B in high school and was a walk on to his college team. It was hard for anyone to predict he'd become a two-time All-Star closer. Ryan Madson had Tommy John surgery then literally didn't throw a big league pitcher for three and a half years before his rebound with the Royals. Luke Hochevar went from first overall pick to poor starter to solid reliever.
There's no way that Dayton Moore knew that all of this would come together. The only common theme among the above guys is really one thing - failed starter.
Let's look at the top 30 relievers in 2015 by fWAR and where they came from:
I thought it was going to be a little more one-sided, but almost 70% of the top 30 relievers in 2015 were once starters. Whether they are recent converts or only lasted a year or so in the minors, the majority of them were once seen as starting pitchers by their clubs.
When the Astros traded away Vince Velasquez, Mark Appel, and other in exchange for the Phillies young relief stud Ken Giles I wondered if the Astros would have just been better off converting Velasquez to a reliever and keeping all the prospects. Of course, Giles is a really good reliever but Velasquez is a decent starter. He's likely a better starter than Wade Davis or Luke Hochevar ever was. Steamer projects Velasquez to be a reliever for 2016 and gives him a 3.24/3.39 ERA/FIP. That would roughly put him in the top 70 or so relievers in baseball, and I owould gladly take the over to beat that ERA/FIP projection.
This isn't new information of course. Starters normally do better once they are banished to the bullpen.
I think there are really two strategies to build a bullpen:
1. Trade/sign the established elite relievers
Relievers are highly fungible. However the elite relievers - Craig Kimbrel, Aroldis Chapman, Wade Davis - usually remain elite relievers for several years. For example, let's take five elite level relievers: Wade Davis, Aroldis Chapman, Greg Holland, Craig Kimbrel, and Kenley Jansen. If you average out their FIP- and ERA- you get:
2. Get lucky