The Mets have baseball's second-lowest bullpen ERA, at 1.21. They just won MLB.com's "Bullpen of the Week" honors. It won't last like this, but it doesn't matter. This is a good group, created via international free agency (Jeurys Familia, Hansel Robles), Draft (Paul Sewald), trade (AJ Ramos, Jerry Blevins, Jacob Rhame), Major League free agency (Anthony Swarzak), and, perhaps most importantly: from their own starting rotation.
In Seth Lugo and Robert Gsellman, who had started 55 combined games for the team since 2016, the Mets may have created a pair of above-average relievers in the most modern way possible. They took a pair of talented-yet-inconsistent young starters, stopped asking them to go deep into games, and asked them to throw fewer four-seam fastballs.
Welcome to baseball, 2018-style. This is how it's done now.

The early results have been stunning. Gsellman has allowed a single run in six innings, striking out nearly 40 percent of the batters he's faced. Lugo threw three scoreless innings of relief in the Mets' 6-5 win over Washington on Sunday night, striking out Michael A. Taylor and Pedro Severino with the bases loaded in the bottom of the ninth, and is unscored upon in six innings so far.
This isn't a story entirely about velocity, but it's a little about velocity. As is so often the case, Lugo and Gsellman have been able to throw harder in short stints than they did as starters.
Entering play on Monday, there have been 299 pitchers who threw at least 100 fastballs in 2017 and have also thrown 25 in '18. It's easy to lose velocity, in part because we're comparing cold early-season weather to a full '17 season, and 129 pitchers are down at least 1 mph so far. Only a dozen are up by at least 1 mph Two of the top five are Mets.
Largest velocity increases, 2017-18
+2.1 mph -- Amir Garrett, Reds
+1.8 mph -- Lugo, Mets
+1.7 mph -- Buck Farmer, Tigers
+1.6 mph -- Hector Velazquez, Red Sox
+1.3 mph -- Gsellman, Mets (tied with three others)
(Unsurprisingly, Garrett and Farmer are also working out of the bullpen this year, after having been primarily starting pitchers last season.)
For Lugo, that's up from 91.2 mph to 93 mph. For Gsellman, that's a jump from 92.8 mph to 94.1 mph. Every tick matters, and it's true for this pair, too. Over their careers, when throwing fastballs 94 mph or higher, they allow a .244 average and a .335 slugging percentage. When it's 93 mph or below, that's a .283 average and a .456 slugging percentage. Of course it matters.
As so often happens, a move to relief also brings with it a change in pitch usage, and as we explained before the season began, the trend away from fastballs has been clear for years; just look at Rich Hill, or Lance McCullers It's the same thing here, too.
Lugo is probably best known for his high-spin curveball that made him a sensation in certain circles from the moment he made his Major League debut back in 2016, but he didn't actually throw it that much last year, just 17 percent of the time. After being encouraged to use it more in '18, that's up to nearly a quarter of the time, at the expense of a relatively unimpressive four-seamer.

He's still using his sinking two-seamer 30 percent of the time, just at a higher velocity, and when he uses it and his curveball in the same plate appearance, it can make even a longtime star like Carlos Santana look foolish.
The story is similar but not identical for Gsellman. He's also throwing more breaking pitches, as his curve and slider are up to 35 percent of the time from 27 percent last year. For him, it's more about regaining confidence in his good moving two-seamer, and all but abandoning his own straight, hittable four-seamer.

So now you have two pitchers throwing harder, focusing on their best pitches, and, importantly, not having to worry about turning the lineup over three or four times.
For Lugo, prior to 2017, he'd allowed a .245/.304/.418 line the first two times through a lineup … and a much inflated .285/.347/.425 line the third and fourth time. For Gsellman, prior to '17, he'd allowed a .249/.320/.388 line the first two times through … and a massive .329/.395/.542 line after that. This is the obvious trend that we've seen in baseball for years. Why put non-elite starters in positions where they aren't as likely to succeed? This was always the best path for the Mets.
It's not like it's surprising to see them excel out of the bullpen, anyway, at least in Lugo's case. In 145 1/3 innings as a starter prior to 2018, Lugo had a 4.09 ERA, and had allowed a .264/.319/.438 line. In 20 innings in relief, those numbers had dropped to an ERA of 2.70, and a line of .167/.269/.262.
He was always better suited for this. So was Gsellman.

In retrospect, one of the most important days of the Mets' season may end up having been April 2, a day they didn't even take the field. That day, a week ago Monday, the Mets were supposed to play the Phillies at Citi Field. They got rained out. In response, they pushed scheduled starter Matt Harvey back a day, and skipped Lugo.
Lugo had won the fifth starter job out of camp, beating out longtime Met Zack Wheeler, who has made 66 starts for the team back to 2013, in large part because of how badly Wheeler struggled in both '17 and in Spring Training. But when the No. 5 spot comes up again on Wednesday, it won't be Lugo getting the ball. It'll be Wheeler.
It's not because Wheeler was so impressive in his lone Triple-A start. It's because, as manager Mickey Callaway said, Lugo and Gsellman "add a whole other dynamic to our bullpen, and help us win games."
They do, and they have. A week ago, Lugo was supposed to be in the rotation. He never made that start. The Mets might be better off if he and Gsellman never do.