In the never-ending battle to figure out how to be a pitcher and survive at Coors Field, perhaps the Rockies have given us a sneak peek into their thinking, at least when it comes to the types of pitchers they want to have in their bullpen.
Lots of sliders. Lots of fastballs. Few curveballs. Almost no changeups.
We had a good idea this was the case last winter, when the Rockies acquired Jake McGee, Chad Qualls and Jason Motte, some of the most fastball-heavy relievers. Now that Colorado has reportedly signed Greg Holland to go along with Mike Dunn, who inked a three-year deal in December, we have a much better idea.
The Rockies' past five relief additions have all followed a similar template.
Just check out the 2015-16 pitch mixtures of the seven most likely Colorado relievers, adding Adam Ottavino and Carlos Estevez to Dunn, Holland, Motte, McGee and Qualls. (We'll group different kinds of fastballs, like four-seam and sinker, together for these purposes. If the numbers don't quite add up to 100 percent, that's because we're not showing pitches used less than 5 percent of the time here.)
2015-16 pitch usage by percentage, from Rockies' projected 2017 bullpen
Dunn -- 62 fastball / 37 slider
Estevez -- 66 fastball / 19 slider / 9 changeup
Holland -- 48 fastball / 46 slider / 4 curveball
McGee -- 78 fastball / 11 slider / 11 curveball
Motte -- 94 fastball
Ottavino -- 57 fastball / 43 slider 
Qualls -- 58 fastball / 40 slider
See a trend? That's a group that is overwhelmingly fastball/slider. Combined, we're looking at 69 percent fastballs, 26 percent sliders, 3 percent curves and 2 percent changeups. If you consider both fastballs and sliders to be "hard" pitches, then that's 95 percent hard stuff -- and even Estevez's "changeup" averaged 90 mph, making it one of the hardest-thrown changes in baseball, and faster than the four-seam fastballs of some pitchers.
Taken by itself, that does not seem to be a coincidence, particularly because six of the seven names were acquired from outside the organization. (Ottavino was claimed from St. Louis back in 2012, while Estevez was signed as an 18-year-old international free agent in '11). Five have arrived in merely the past calendar year. Still, in order to figure out if that's extreme, we'll have to compare it to other bullpens around the game.
Gif: Ottavino throws 2 nasty sliders
Here's how we're going to do this. Let's take the seven names above, along with their pitch usage, and see, if they'd functioned as a unit last year, how their pitch repertoires would have looked as opposed to other relief groups. This is obviously not inclusive of every Rockies reliever you'll see -- surely, guys like Miguel Castro, Chris Rusin, Jordan Lyles, Gonzalez Germen or Scott Oberg will appear at some point -- but we had to draw the line somewhere, and our group of seven could easily be the Opening Day bullpen.
So is it extreme? Uh, yes. While our hypothetical 2017 Rockies bullpen would have only been sixth in fastball usage and fourth in slider usage, the point isn't that they're leaning heavily on a single pitch, it's that they're specifically targeting a combination of fastballs and sliders, more than any other bullpen.
Highest fastball/slider percentage by bullpens (using 2016 stats)
94.6 -- projected 2017 Rockies
90.4 -- Orioles
88.6 -- Reds
86.8 -- Mets
86.5 -- Nationals
86.0 -- 2016 Rockies
If you're throwing that much hard stuff, there's not much room for the soft stuff, and so you can probably already imagine where our hypothetical 2017 Rockies bullpen appears on the list of curveballs and changeups.
Low. Extremely low.
Lowest curveball/changeup percentage by bullpens (using 2016 stats)
4.6 -- projected 2017 Rockies
9.6 -- Orioles
11.5 -- Reds
13.3 -- Mets
13.7 -- Nationals
13.9 -- 2016 Rockies
Now, does it work? That's the trick, of course, because all the theories matter only so far as they actually turn into on-field success, and in 2016, it didn't really work. McGee got hurt. So did Motte. And Qualls. Ottavino missed the early part of the season recovering from Tommy John surgery, while Estevez was back-and-forth between the Majors and Triple-A as a rookie. Plenty of other relievers simply didn't produce. It's not easy to pitch at Coors Field, no matter what you throw.
Gif: Greg Holland Strikeout
There's long been evidence, however, that sliders may be the most effective secondary pitch to use at altitude, as Colorado starter Jon Gray has proven. If we look at 2016 pitch outcomes at Coors Field -- at a high level anyway, as we're just looking at all pitches, uncorrected for pitcher or hitter skill -- sliders had the lowest average against (.226), lowest slugging (.394) and highest whiff rate (33.7 percent) of the seven main pitch types. So that's something, even if by itself that's not conclusive.
This isn't even a new strategy, necessarily. It's been more than four years since Jhoulys Chacin admitted he ditched his curveball and began using his slider more at home, and even then the majority of pitches at Coors were four-seamers and sliders. If there's a difference now, perhaps it's less about forcing pitchers to adapt to the environment, and more about simply getting pitchers best-equipped to succeed there in the first place. That doesn't mean it'll work, and it didn't last year. It can't just be business as usual at altitude, though. We know that won't work.
This might.