Astros have most dangerous 'pen again

It's really, really hard to pitch in Houston's Minute Maid Park. Despite a deep center field, the short porches in both left field and right field can turn relatively routine fly balls into home runs. Just look at Wednesday night, when Jose Iglesias and Alex Bregman each had 348-foot flies that would never leave any other park but turn into dingers in Houston. If it's not quite Coors Field, it might be closer than you'd think.
That being the case, you'd understand if the Astros had a tough time putting together an effective pitching staff, especially in the bullpen. And even if they did, you'd understand if they had a hard time keeping it together for multiple years, given how volatile relievers can be from one year to the next … yet that's exactly what Houston has done. The Astros have the Majors' deepest bullpen, and for the second year in a row, they may just have baseball's best bullpen.

RP can be pretty volatile year-to-year, which is why this xwOBA chart is all over the place. All the more impressive how good HOU has been. pic.twitter.com/zn9ODbTvsy

— Mike Petriello (@mike_petriello) May 26, 2017

So how do you manage to succeed in a ballpark like that? Step one is "miss bats," since strikeouts don't become homers. Step two is "limit dangerous contact," because weak hits (mostly) don't become homers. The Astros are doing both better than just about anyone.

Baseball's best strikeout bullpen
Obviously, all of baseball is striking out at a record pace, which is something we've been saying every year for the past decade. So while it's technically true that the Astros' bullpen has the highest strikeout rate ever so far this season, it's also pretty misleading, given how much the game has changed over the years. Still, if you just look at 2017, the 31.2 percent strikeout rate Houston's relievers are carrying is the best in the game, ahead of bullpens fronted by Kenley Jansen, Andrew Miller and Dellin Betances.

Not surprising to see LAD/CLE/NYY near top of bullpen K% rankings. A little surprising to see the Astros atop them all. pic.twitter.com/m1ELEdcfG2

— Mike Petriello (@mike_petriello) May 26, 2017

How good a number is that? It's basically what Jacob deGrom (32.3 percent) has done, and the walk percentage of Houston's bullpen (9.7 percent) is identical to deGrom's (9.8), too. Obviously, it's easier to face a few batters at a time than it is to be a starter going deep into the game, but you get the idea. It's an extremely favorable comparison.
Now, it's easy to point to Chris Devenski (43 percent strikeout rate), who has become one of baseball's most dangerous weapons, pitching in any role manager A.J. Hinch needs. Yet while he's arguably one of the 10 best relievers in the game right now, this is really a team effort. Here's what we mean -- so far in 2017, Major League relievers have an average strikeout percentage of 23.2 percent, ranging (with a minimum of 10 innings) from Craig Kimbrel (55.6 percent) to Rob Scahill (3.9 percent). Let's take a 30 percent strikeout rate not just as "above average," but as "excellent." Let's see who's above that. Hint: It's the Astros.
Houston has five relievers who have whiffed 30 percent or more of the hitters they've faced -- James Hoyt (44.4 percent), Devenski (43.0), Brad Peacock (33.3), Michael Feliz (32.5) and Ken Giles (30.8). Only one other team, the Dodgers, has five qualified relievers striking out hitters at at above 30 percent. Nine teams have none.

While these guys are hardly one-pitch pitchers, they've all got at least one elite offering. Devenski, for example, has baseball's best swinging strike rate on his changeup (of 82 pitchers who have thrown 50). Hoyt, Giles and Luke Gregerson all have top-25 swinging-strike rates on their sliders (of 124 who have thrown 50). Feliz, Devenski and Peacock are all top-20 in four-seam swinging strike rates (of 261 who have thrown 50).
When you play in a ballpark like the Astros do, the only sure way to avoid cheap extra bases is to prevent contact. Houston is doing that wonderfully.
Avoiding dangerous contact
Of course, you can't always prevent contact, and we haven't even mentioned Will Harris or Tony Sipp yet. But you might be saying to yourself, "The Astros' relievers are only 15th in average exit velocity allowed, at 85.4 mph." That's league average. That's not impressive.
But as we like to remind, exit velocity matters most when it's combined with launch angle, because hitting the ball really hard straight up is probably still going to be an out. Hitting the ball hard at the correct angle (say, 25-30 degrees) could be a homer, but hitting it hard on the ground right to Jose Altuve is nearly certainly a 4-3 in the box score.
That's the basis of Hit Probability, which takes the combination of exit velocity and launch angle and tells you how often similar batted balls have been hits. As an example, Jake Marisnick's homer on Thursday, hit at 104.5 mph at 28 degrees, falls for a hit 88 percent of the time. Sometimes a great defensive play is made or the wind holds it up, but regardless of the outcome, he deserves credit for the skill it takes to square up a ball like that.
So let's take a look at which bullpens have had the fewest batted balls allowed with at least a 75 percent Hit Probability (or .750 batting average, if you prefer). These are the most dangerous batted balls, the ones good pitchers prevent. The Orioles, for example, have allowed 63 batted balls with a Hit Probability of 75 percent or higher, and 58 of those have become hits, many for extra bases. It's not great. You probably know where the Astros will land...
Fewest batted balls with Hit Probability of 75 percent or higher by bullpens
9.5 percent -- Astros
9.7 percent -- Indians / Braves
9.9 percent -- White Sox
MLB average -- 11.7 percent

So let's tie this all into a bow. The Astros are whiffing more hitters out of their bullpen than anyone, and they're allowing less very dangerous contact than anyone. They don't have the lowest ERA (that'd be the Indians), but that's partially because of their ballpark and partially because of the vagaries of sequencing. (And, partially, because Gregerson's had an early-season home run problem.)
How do we combine good quality of contact and elite strikeout rates? There's two ways, depending on what you prefer. If you prefer batting average, then we have Expected Batting Average (xBA), which combines hit probabilities and whiffs to show a batting average allowed that's independent of defense and ballpark. Houston is first, with a .188 mark that's far better than Cleveland's .209.
If you prefer going a little deeper, we have Expected Weighted On-Base Average, (xwOBA), a more advanced metric which is similar except that it accounts for walks and gives each type of on-base event more weight, rather than treating home runs and singles equally like batting average does. The Astros are first, with a .262 mark that's ahead of the second-place Dodgers at .266.
We knew before the season Houston would have a very deep lineup, and it has. We knew the Astros had a good bullpen last year, but we didn't know for sure if they'd keep it up. They have, and then some. It's why, even before we saw the rebound of Dallas Keuchel, it was easy to pick them as a potential American League champion before the season. So far, so good.