Batted-Ball Data: Not All Fly Balls Are Created Equal

This is the second in a series of pieces on the emergence of batted-ball data into the mainstream. Earlier this week, we covered the basics in a fair amount of detail. Today, we’re going to drill down a bit into arguably the most important piece of information discussed in that article; the fine line between the most and least productive fly balls, where we will quickly discover that not all hard fly balls are created equal.

To reiterate that fine line: in 2014, MLB batters hit .560 AVG-1.884 SLG on fly balls hit at 92.5 MPH or higher, but only .077 AVG-.148 SLG on fly balls hit between 75 and 90 MPH. Such a stark difference in production, despite a gap of only 2.5 MPH between the upper boundary of the less productive and the lower boundary of the more productive group. One might define the 92.5+ MPH fly balls as the “hot zone”, and the 75-90 MPH group as the “donut hole”. As stated in the previous article, all 2014 MPH data comes from a different provider than the one MLB utilizes for StatCast in 2015. The StatCast data, based on early returns, runs a few MPH faster.

Of course, it’s a major oversimplification to simply snatch those numbers out of a vacuum. The first, arguably most important factor to discuss is horizontal angle. While the vertical exit angle determines whether a batted ball is a popup (> 50 degrees), fly ball (between 20 and 50), a line drive (between 5 and 20) or ground ball (below 5), the horizontal angle determines the field sector to which a ball is hit. The field sectors are defined as follows:

LF = > 110 degrees
LCF = between 95 and 110
CF = between 85 and 95
RCF = between 70 and 85
RF = < 70 While spin of the ball off of the bat might cause a ball to hook or tail out of a particular sector before reaching the ground or a fielder's glove, the horizontal exit angle is used for analytical purposes. Let's now revisit the fly ball "hot zone" and "donut hole" on a field sector-specific basis.

The above table breaks down the actual production on MLB fly balls in 2014 by exit speed range and field sector. The two right-most columns aggregate the data for all field sectors combined. Closer scrutiny shows how the fly ball “hot zone” and “donut hole” shifts by field sector.

In LF and RF, you can start doing significant fly ball damage at 87.5 MPH and up. To those sectors, the “hot zone” starts somewhere between 87.5 and 90 MPH. The LF and RF “donut holes” run roughly from 75-87.5 MPH. There is a very sharp delineation between the hot and cold zones; on one hand, it might seem bizarre to see such a sudden jump in production at a specific MPH marker, but on the other, it makes perfect sense, as there is a huge difference between a warning track fly ball and a just-enough homer. One counts for .000 AVG-.000 SLG, the other for 1.000 AVG-4.000 SLG. This is the fine line between, say, Raul Ibanez 2013 and 2014.

In LCF and RCF, the hot zone appears to kick in just north of 92.5 MPH, while the donut hole begins just south of 90 MPH. In CF, the hot zone doesn’t kick in until you reach about midway between 97.5 and 100 MPH, while the donut hole runs from about 75 MPH to over 95 MPH. Horizontal angle clearly makes a big difference production-wise.

In all field sectors, you’ll notice the high production levels for fly balls between 60-70 MPH, extending down to 55 MPH in LF and RF. These would be your bloop singles, which are not a function of talent, but of luck.

All of this is pretty much unsurprising when you think about it; you don’t have to hit the ball as hard down the lines to do damage as you would in the middle of the field. What it does accomplish, however, is the identification of some MPH benchmarks for success in each field sector. Bottom line: fly balls at 97.5 MPH and up, to any field sector, give you a good chance of doing damage. Fly balls at 75-87.5 MPH are usually outs, to any field sector. Depending on the field sector, as indicated above, the hot zone kicks in somewhere between say, 88.8 and 98.8 MPH, while the donut hole upper boundary can range from anywhere between 87.5 to 96.0 MPH. This a healthy amount of MPH overlap between the “good” and “bad” zones. Yes, a dead-out, donut hole 96 MPH fly to CF is hit much harder than a just enough, 88 MPH homer down the line. Exit speed alone can only tell you so much.

A few words about hitter fly ball pull tendencies are in order. The topic of pulling, both in the air and on the ground, and resulting overshifting are worthy of an article in their own right, and they will get one. Here are some brief parameters specific to fly balls, however. I use a simple stat called “pull ratio” when evaluating hitters. For each BIP type, It is simply (BIP to LF + LCF)/(RF + RCF) for righty hitters, and (RF + RCF)/(LF + LCF) for lefties. AL lefty hitters had an average fly ball pull ratio of 1.22, NL lefties, 1.18. AL righty hitters had an average fly ball ratio of 1.43, NL righties, 1.10.

One might say that hitters should focus on pulling the ball in the air, as less exit velocity would be needed to do damage. It’s not nearly that simple, unfortunately. A higher fly ball pull ratio equals a higher grounder pull ratio, for almost all hitters, leading to infield overshifts, and the opening up of the outer half of the plate for pitchers to attack. A successful power hitter looks for opportunities to selectively pull the ball in the air, without totally selling out to pull at all times.

Let’s look a couple of players with fairly similar fly ball exit speed portfolios. Austin Jackson had an average fly ball exit of speed of 83.5 MPH in 2014, just above the AL average of 83.2 MPH. 17.3% of his fly balls were hit at 92.5 MPH and higher. Jackson’s overall fly ball production was very low, however, at .195 AVG-.398 SLG. Brian Dozier’s average fly ball velocity was slightly higher, but still squarely in the average range, at 84.9 MPH. 24.3% of his fly balls were hit at 92.5 MPH or higher, and his fly ball production was markedly higher at .243 AVG-.764 SLG.

Why was there such a huge difference in these two players’ fly ball production? Jackson only pulled three of those 92.5+ MPH fly balls to the LF field sector, while Dozier pulled a whopping 22. Now does this make Dozier a better hitter than Jackson? Not necessarily; Dozier does have an excessive ground ball pulling issue that creates problems of its own. He is certainly better at selectively pulling in the air than Jackson, however.

Over half of Jackson’s hard fly balls were hit to CF, RCF and RF, and weren’t hit hard enough to do much damage. Fully three quarters of Dozier’s were hit to LF and LCF, and did plenty of harm to opposing pitchers. Jackson actually hit significantly more 97.5+ MPH flies compared to Dozier, a 9 to 3 margin, but hit 7 of them to the middle three field sectors, while Dozier focused on the extreme pull side. Exit speed is indeed only a piece of the puzzle.

Now that we’ve taken a look at exit speed within the context of field sectors, let’s take the next step and place both of those within the context of different ballparks. If a 95 MPH fly ball is a homer to LF, but a routine out to CF in a neutral park, what happens when that fly ball is hit in Coors Field as compared to Safeco Field?

The two above grids detail the production on fly balls at Coors and Safeco Fields, by exit velocity and field sector. As you can tell by the bottom right corner fields on both grids, a lot more damage was done on fly balls at Coors than at Safeco. Interestingly enough, this raw data actually understates that difference, as there were many more hard-hit fly balls hit at Safeco last season than there were at Coors. We’re not here to calculate park factors right now, however. Let’s instead focus on the “hot” and “donut-hole” zones at both parks.

At Coors, you only have to hit fly balls at 87.5 MPH down both lines to start doing real damage. At Safeco, you have to get north of 92.5 MPH to accomplish much. At Coors, you only have to hit the ball at 92.5 MPH to start doing material damage in both gaps, while at Safeco, you have to get all the way up to 97.5 MPH in LCF, and 95 MPH in RCF. To CF, the magic number is around 95 MPH at Coors, but a hefty 100 MPH in Safeco. In a baseball world where tenths of one MPH of batted ball speed can make a difference, these 2.5 to 5 MPH differences in production thresholds are absolutely massive.

The donut hole MPH zones are also substantially larger at Safeco compared to Coors. Down the lines, the dead zones at Coors are quite small; say, 75-80 MPH to LF and 80-85 MPH to RF; in Safeco, the LF dead zone is much larger, from 70-87.5 MPH. At Coors, the dead zone in the LCF gap runs from 80-90 MPH, and from 75-90 in RCF. This compares to an amazing 70-95 MPH in LCF and 75-92.5 MPH in RCF in Safeco. Interestingly the dead zones to CF in both parks are quite large; from 70-95 MPH in Coors, and from 80-100 MPH in Safeco.

If you want to wash down the comparison of Coors to Safeco into a singular morsel, consider the overall production on fly balls between 97.5-100 MPH; .701 AVG-2.368 SLG in Coors, .360 AVG-1.124 SLG in Safeco. Or if you need a second one, how about 90-92.5 MPH fly balls? .259 AVG-.741 SLG at Coors, .081 AVG-.216 SLG at Safeco.

That about sums it up; no two 92.5 MPH fly balls are created equal. Horizontal angle, ball park, weather conditions, etc., all have their say. The same fly ball, with minor contextual shifts, can either be the most harmless of outs or the most productive offensive play in the game. Introducing granular batted-ball data into the evaluation process enables us to peel away the biases and the noise, and combined with traditional scouting methods, better evaluate the true talent within.