Baseball has long been a game of tradition, and one rightful criticism of our “national pastime” has been its tendency to be slow to change. One of the most welcome enhancements of some fans’ enjoyment of the game has been the introduction of Statcast in recent years. No, it’s not for everyone, but its existence — and most of all, its availability for free to the populace — adds another avenue of potential involvement for the fan base, while also offering countless opportunities for study of any aspect of our game from a nearly infinite number of perspectives.
And the numbers haven’t just been randomly thrown out for the public without context being provided by those with significant experience with the data. I’ve been writing about this stuff for years, for one. Moreover, props must be given to Mike Petriello, Tom Tango, and Daren Willman (for his indispensable Baseball Savant website) for their contributions.
First, let’s get a bit of big picture perspective here. Below is a comparison of all MLB players’ production by BIP type through May 20, 2017, compared to the full 2016 season:
|Through May 20, 2017||2016|
From a pure overall production standpoint, there’s not much to report. Players recorded a .328 OBP and .536 SLG on BIP in 2016 versus a .324 AVG-.535 SLG through 5/20 of this season. Bear in mind, of course, that we haven’t hit the true warm-weather months yet; the 2017 numbers will only swell from here.
The big news is the disappearance of that pesky “null” category, which prior to this season was fairly well populated. These are all of the batted balls that the hardware installed in each ball park failed to capture, mostly composed of balls hit at very high and very low launch angles. Balls hit at those extreme angles aren’t typically hit very hard, as evidenced by the .161 AVG-.213 SLG line compiled on them.
So what’s the deal? Is the equipment now getting a reading on every batted ball? Not quite. It appears that the exit speed and launch angle of BIP previously categorized as “null” are now being estimated. For instance, there are suddenly a awful lot of pop ups with 80 mph exit speeds and 69 degree launch angles, ground ball singles with 90.3 mph speeds and 17.3 degree launch angles, and ground ball outs with 82.9 mph exit speeds and 20.7 degree launch angles.
What’s the impact of this? I’d say that, for the most part, it’s a positive. You can now say with much certainty that all BIP are being placed in the correct category. (I use 50-plus degrees to define pop ups, 20-50 fly balls, 5-20 liners, and under 5 degrees for grounders.)
The estimation of exit speeds isn’t ideal, however. It won’t affect the measurement of fly-ball or line-drive contact quality at all. It will, however, leave some room for error in the measurement of grounder contact quality. Not all of those singles are hit at 90 mph, and not all of those outs are hit at 80 mph; I’d bet a bunch of those outs are hit much more weakly, and their classification as such might obscure some of the accomplishments of weak grounder generators like Dallas Keuchel, for example. Time will tell.
In any event, the data keeps getting more public, more complete, and more correct with each passing year. And that is a very good thing.
One of the things for which I love to use Statcast is to produce park factors. Take the complete set of batted balls at any park, apply the major-league average production for each BIP type/velocity bucket, incorporate run values, and voila, you’ve got some highly accurate park factors that don’t require multiple years to stabilize.
|Through May 20, 2017||2016||2015||2014||2013|
Color-coding is used above to note significant divergence from league average. Red cells indicate values that are over two full standard deviations above league average. Orange cells are over one STD above, yellow cells over one-half-STD above, blue cells over one-half STD below, and black cells over one STD below league average. Ran out of colors at that point. Variation of over two full STD below league average will be addressed as necessary in the text below.
This table lists the overall single-season park factors for each club going back to 2013, with partial-year data through May 20 included for 2017. One might opine that single-season factors don’t tell you much and might be too volatile. Well, the average year-to-year correlation coefficient for the 2013-16 annual sets of date above is 0.58, indicating a strong correlation. That occurred despite significant modifications of multiple venues over that time frame.
Look at some of the tightly clustered annual park factors for many stadiums. On the hitter-friendly side we have, of course, Colorado (ranging from 123 to 128), Boston (110 to 113), the NY Yankees (106 to 112), Milwaukee (102 to 112), Cincinnati (101 to 110), and Minnesota (100 to 110). On the pitcher-friendly side, we have Miami (81 to 98), Seattle (83 to 97), San Francisco (85 to 96), Oakland (85 to 96), Atlanta (90 to 98), Kansas City (91 to 100), Washington (92 to 96), and the LA Angels (92 to 98). These parks are what they are.
Interestingly enough, the partial season 2017 data has a 0.50 correlation coefficient when compared to 2016. Expect that correlation to increase as the warm weather kicks in, driving park factors up in Baltimore, Boston and Minnesota, to name three.
Here are some park-specific notes gleaned from the 2017 data:
- You might not be able to tell from traditional park-factor data, but the new Atlanta ballpark appears to be much more hitter friendly than its predecessor. The fly balls hit by the Braves and their opponents to data haven’t been hit all that hard (“should have” resulted in .301 AVG-.846 SLG to date), but have actually resulted in much more production (.345 AVG-1.008 SLG, 137.9 Fly Ball Park Factor). In all of MLB, almost exactly 15.0% of fly balls hit between 95-100 mph have gone over the wall this season; 22.0% of such flies have done so in this new yard.
- Great American Ballpark has entrenched itself as one of the most hitter-friendly parks in the game. Almost 49% of all fly balls hit between 100-105 mph have gone for homers this season. In Cincinnati, 32 of 44 (73%) has gone over the wall. If you hit it good here, it’s as good as gone.
- The entire Coors Field effect can be boiled down to one group of batted balls: 95-100 mph fly balls. In all of MLB, such flies have gone for a .290 AVG-.891 SLG thus far in 2017, while in Colorado, hitters have posted a .524 AVG-1.667 SLG mark. Of the flies in that bucket, 29% have gone over the wall, almost twice the MLB rate.
- How much better would Miguel Cabrera’s career numbers be if he played in a more neutral park? All but 103 MLB fly balls hit over 105 mph (492 of 595, 83%) have gone for homers this season. Sixteen of those 103 non-homers were hit in a single yard, Comerica Park.
- Houston is a weird one. Even though Tal’s Hill is gone, this is still your classic long-out/cheap-homer park. Less than 9% of MLB flies hit between 90 and 100 mph have gone over the wall this season. Of those hit in Houston, 23% have. In all of MLB, only 33 homers have been hit on 90-95 mph fly balls; seven of those have been hit in Houston.
- San Diego is not a pitchers’ park. No park swings as wildly from pitchers’ to hitters’ yard as does Petco, and I firmly believe it has much to do with the marine layer. Its 95-100 mph homer rate of 30% this far in 2017 is even higher than the other parks already discussed.
- Then there’s AT&T Park. Batters “should be” hitting .290 AVG-.891 SLG on 95-100 mph fly balls; they’re hitting all of .188 AVG-.375 SLG in San Francisco. Only one of 48 flies hit in that range (2%) has cleared the wall.
- Lastly, let’s have a brief word about doubles on line drives in Milwaukee. Thus far in 2017, about 16% of liners in all of MLB have resulted in doubles. Of the liners hit in Miller Park, 24% have gone for doubles. Let’s drill down and focus on 95-100 mph liners: 19% of all MLB liners in that bucket have resulted in two-base hits, but 31% of them have done so in Milwaukee. MLB hitters have slugged .886 on 95-100 mph liners, but have slugged a whopping 1.306 in Milwaukee.
Lastly, as a supplement, here are the fly-ball park factors for each club from 2013 through May 20th of this year. Again, there’s a strong correlation: an average of 0.61 from 2013 to -16 and 0.66 from 2016 to the partial season 2017 data:
|Through May 20, 2017||2016||2015||2014||2013|