The 2017 Strike Zone

The strike zone saw its fair share of changes in 2017. (via Arturo Pardavila III)

Spring training is wrapping up, and actual major league games are beginning Thursday for our viewing pleasure. Before the regular season begins, let’s review what happened with respect to the strike zone in 2017.

For many years after the introduction of pitch tracking data in the public sphere, the strike zone was a hot topic. Between 2009 and 2015, the size of the called strike zone increased each season, thanks in large part to the bottom of the zone dropping on the order of three inches—the diameter of a baseball—during that stretch. Armed with a technology-driven feedback loop that was pushing to a more accurately and consistently called strike zone throughout the majors, the zone expanded in the process.

Over the same period, run scoring fell precipitously. At least three analysts independently calculated that the expanded strike zone directly contributed to roughly one third of that reduction in offense. With runs dwindling, MLB began to consider rule changes to the strike zone definition to move the bottom of the zone back up, with the goal of spurring scoring from its lowest level since the early 1980s. Conveniently, after the All-Star break in 2015, home run totals skyrocketed and have not let up since. This new source of offense has helped drive run scoring back to its levels before the strike zone began expanding, without MLB having to interfere with the called strike zone in any official manner.

Here is the recent history of the major league strike zone:

The observations of the strike zone in 2017 suggest what was suspected after the 2016 season was real —that the strike zone size peaked in 2015 and now is slowly shrinking. Here is a visual representation of the typically called strike zone from last season for left-handed hitters and right-handed hitters, from the home plate umpire’s perspective. Cells colored green indicate square inch areas above the front of home plate where more than half of all pitches crossing in the area were called strikes.

PITCHf/x vs. Statcast

An interesting exercise is to compare the strike zone to the previous season, with an eye to seeing where the strike zone is changing. The following strike zone heat map highlights these trends. Blue cells indicate areas where more balls were called in 2017 than 2016, while red cells denote places where more strikes were called than the season prior.

Measurements in the earlier table had already shown a smaller called zone around the knees and below, so it should be no surprise to see more blue than red along the bottom of the zone. The most obvious trend these images uncover is a shift in called strike zone for all hitters toward the left-handed batters box. As will be discussed in more detail below, this trend has been developing for many years for left-handed hitters, but to see this dynamic for right-handed hitters as well is unexpected.

In 2017 in particular, there was a confounding factor in the comparison of the strike zones between seasons. In every season between 2007 and 2016, the source of pitch tracking data was PITCHf/x. Starting in 2017, publicly available pitch tracking data is now produced by Statcast. As the technology is different, the calibration process is also different.

It is possible some of this shift in reported strike zone positions is due to Statcast measuring the location of a pitch as it crosses home plate slightly closer to the left-handed batters box than PITCHf/x would have measured for the exact same pitch. To try to gauge this, consider another heat map showing the average horizontal pitch location from each major league stadium during recent seasons.

It’s looking pretty red in the rightmost column from the 2017 season with Statcast data, with only one stadium showing up as blue in the color map. By comparison, looking at the last year of PITCHf/x data in the column second from the right in 2016, there is a fairly equal mix of blue and red. Of course looking back throughout the years PITCHf/x was employed, it is clear there is yearly variability even without having changed tracking technologies.

Without having an overlapping set of pitches that are measured by both systems, it is not trivial to deduce how much of the shift toward the left-handed batters box is due purely to calibration offset between the two measurement systems and how much is due to adjustments made by home plate umpires. What can be said is that the average reported horizontal pitch location was 0.86 inches closer to the left-handed batters box in 2017 than it was in 2016. That broke down as a 0.97 inches shift for left-handed hitters, and a 0.76 inches shift for right-handed hitters.

In reality, this is relevant only when trying to compare the exact position, or the center of the zone, if you will. Measuring the absolute size of the strike zone should not be affected, provided the measurement error is reasonable for both systems.

Vanishing “Lefty Strike”

One starkly visible change over time is in the lack of strike zone area well beyond the outside edge of home plate for left-handed batters. Contrasting the called strike zone for lefties when pitch tracking was first publicly available to this most recent season demonstrates just how much the strike zone has morphed over a decade.

In this image not only is the baseball-sized addition across the bottom of the zone apparent, but also more or less a baseball-sized subtraction from the outside part of the zone. Of course, a part of the reason there is not even a single green cell beyond one foot left of the center of home plate from this perspective is likely due to a calibration offset between Statcast and PITCHf/x, but the trend to skimming off this “lefty strike” has been in progress for many years.

Left is Right?

The astute reader will notice in the visual representation of the 2017 strike zone that the called strike zone for right-handed batters is larger than for left-handed batters. From the images, it is clear the largest difference in pitch calling between batters on opposite sides of the plate is on the inside edge of the zone. This remains an area where umpires do not call strikes as often for left-handed hitters as they do for right-handed hitters.

Here is a recent history of the gap in strike zone size between right-handed hitters and left-handed hitters:

Since the gap in strike zone size jumped from a consistent six or so inches to double digits in 2014, it has not looked back. Coincident with this widening zone difference has been a notable change in right-handed hitter strikeout rate relative to left-handed hitters. What used to be handedness neutral has developed into an advantage for left-handed hitters in recent seasons.

The Shift

Defensive shifts have become more prevalent over the past several seasons. One set of studies found the number of called balls in plate appearances in which the shift was enacted was greater than expected based on the number of called balls for the same hitters and pitchers without any shifting. There were signs of fewer strikes being thrown by pitchers with the shift on, as well. A worthwhile exercise may be to compare the average strike zone with the shift to the strike zone without a shift, to see if there is any visible difference.

Since the number of pitches with the shift on is still relatively small compared to the overall pitch total for a season, all pitches from 2015 through 2017 were combined for this analysis. The typical strike zone when either a traditional or non-traditional defensive shift was in place then was determined, as well as the zone for all called pitches with no shift.

The measured strike zone with the shift in place does appear to be slightly smaller than when the defense is in standard positions. This result is consistent with the previously mentioned set of studies that detected more balls and fewer strikes than expected when teams employed the shift. These strike zone sizes are the major league average, and no attempt to control for batter height was performed.

That said, it would be expected that batters who are shifted against if anything would tend to be a group with more power hitters than the non-shifted group, so the players in that group probably would be at least as tall as the non-shifted group. Running the numbers, this turns out to be the case, with the average height of the shifted group of batters being about half an inch taller than the non-shifted group. If anything, this suggests the calculated reduction in strike zone size with the shift employed is slightly understated.

There are no doubt several reasons this gap could exist. There are clear examples of home plate umpires tending to, if anything, work to even the odds when either the offensive or defensive team has an advantage. When the count is in the hitter’s favor, strike calls tend to be more generous. Conversely, when the pitcher is way ahead in the count, the strike zone tightens. Studies have shown other ways umpire tendencies sway in this manner, whether this is happening subconsciously or not.

It is easy to believe when an umpire watches a team load up on the pull side of a hitter, this could be construed as an advantage taken by the defensive team, which if anything would suggest we might expect a called strike zone that marginally favors the offense. Of course with the shift employed, pitchers may deviate from their normal pitch sequencing as far as pitch types and pitch locations, which could perhaps lead to worse pitch command, which could lead to harder pitches to receive for the catcher, which could then take a certain number of borderline pitches and flip them from called strikes to balls as a result of poor framing.

Whatever the cause, when teams shift, the called strike zone shrinks slightly, and the ball-to-strike ratio rises. The fact that this happens is not obvious, but it is a subtle counterproductive consequence of implementing a defensive shift.

While the strike zone is no longer expanding wildly and dramatically influencing the major league run environment annually, there are always new things to be learned about this primary battleground in the game of baseball.

References and Resources

• Noah Davis and Michael Lopez, FiveThirtyEight, “Umpires Are Less Blind Than They Used To Be”
• Etan Green and David Daniels, “Impact Aversion: Agency Failure and Decision Bias at High Stakes”
• Ben Lindbergh, Grantland, “The Strike Zone Time Travel Test: How Much of Baseball’s Anemic Offense Really Stems From the Swollen Zone?”
• Ben Lindbergh, The Ringer, “The Juiced Ball is Back”
• Brian Mills, Managerial and Decision Economics, “Social Pressure at the Plate: Inequality Aversion, Status, and Mere Exposure”
• Jeff Passan, Yahoo! Sports, “Sources: MLB could alter strike zone as response to declining offense”
• Jon Roegele, The Hardball Times, “The Strike Zone Expansion is Out of Control”
• Jon Roegele, The Hardball Times,“The 2015 Strike Zone”
• Jon Roegele, The Hardball Times, “The Strike Zone During the PITCHf/x Era”
• Gerald Schifman, The Hardball Times, “The Lurking Error in Statcast Data”