Pirates pitchers can see better results through changing eye levels

The Pittsburgh Pirates pitching staff in 2017 season was the definition of mediocre, which is actually good news.

The Pittsburgh Pirates’ staff as a whole posted a 4.22 ERA, good for 14th in the league and 0.13 runs better than the league average; doing so with pitching from the likes of Antonio Bastardo, Joaquin Benoit, Tyler Glasnow, and Daniel Hudson. If we’re being kind, we could say that those pitchers had bad seasons.

The main takeaway is this: If 2017 was the floor for this group, then their ceiling is pretty high.

Here at Pirates Breakdown we’ve focused much of our off season coverage on pointing out deficiencies and ways to address them, both for next season and for the long run. One way for the Pirates to do that is to maximize the production from the players they already have; and there’s no place that this can have a greater impact than on the team’s pitching staff.

One potential way of improving pitching performance is to keep batters off balance by changing where pitches are being thrown, known as changing the eye level. The theory goes like this, if a pitcher throws a pitch high when a batter is expecting something low, it’ll force the batter to make a difficult split-second adjustment, which in gives the pitcher a further advantage over the hitter.

It is not a new practice by any means, but if the Pittsburgh Pirates can make changing a hitter’s eye level a point of emphasis, their pitching staff can inch much closer to their ceiling.

How so?

We can test the truthfulness of this idea by looking at the average difference in height that a pitcher threw for each sequential pitch of a plate appearance.

So, for instance, if a pitcher threw a pitch 12 inches off the ground for their first pitch, 24 inches off the ground for their second and 20 inches off the ground for their third, the total difference would be 12+4=16 inches.

We can then take the per pitch average of the total difference to determine what pitchers changed eye levels frequently. Then, comparing that what the outcome of the plate appearance to that change in vertical location, we can see how effective this change in location was.

The question then is how to define what a “successful” and “unsuccessful” batter faced looks like. Perhaps the most obvious successful outcome is a strikeout; however, the definition should also account for a pitcher inducing weak contact. In order to do this, we can utilize the MLB’s xwOBA (Weighted expected on base average) stat for estimating the offensive value of a ball in play.

xwOBA uses the launch angle and exit velocity of a ball in play to determine the average number of runs generated on similar hits; weakly hit grounders and pop flies produce relatively few runs, where as hard hit fly balls often generate one or more runs. The league average wOBA for 2017 was .321, so knocking off some points to get the pitcher to below league average, weak contact can be defined as having an xwOBA at a somewhat arbitrary value of less than .280. So, for our purposes here, a successful batter faced is defined as resulting in either a strikeout or an xwOBA of less than .280.

The Model

Building an effective model to determine the effects of the change in the vertical position has a lot to do with controlling for the other variations that can cause a plate appearance to go one way or the other. In our model we control for the average speed that a pitcher throws at, the average variation in pitch speed, the size of the batter’s strike zone, and the number of pitches in the at bat.

Intuitively there is a negative impact of having no variation in pitch height as batters should be able to more easily square up pitches, but there is also a negative impact of having too much variation, as a pitcher ends up throwing too many balls. To address this reality, we utilize both the average vertical difference as well as its square, in the model.

We can then run these variables through a logistic regression analysis with this form:

PA Success= f (Avg Vertical Diff, Avg Vertical Diff squared, Strike Zone Size, Avg Pitch Speed, Pitches in PA)


This analysis confirms that that changing the vertical positioning of pitches, i.e. changing the batter’s eye level, yields better outcomes for the pitcher, at a high level of significance (>99.99%). However, this bump in improvement does diminish as the eye level change becomes too great.

Under the league average conditions for speed, variation in speed, strike zone size, and pitches per PA, the probability of a pitcher having a successful Plate Appearance is maximized at 61% when a pitcher throws an average of 10.4 inches of vertical difference, or change in the batter’s eye level, during a plate appearance.

Beyond that, there are a few other interesting takeaways.

The first is that there is a reduction in likelihood of a successful batter faced when a pitcher increases their average speed. This is a bit counter intuitive since we think of hard throwers as having an easier time striking batters out. There are a few possible causes for this, the first is that it is more difficult for a pitcher to throw accurately when throwing harder than they normally do, leading to more balls or pitches left in bad spots that the batter can hit. The second is that when a pitcher throws harder, any contact that is made would mean the ball traveling out that much faster, increasing the xwOBA. Thirdly, batters could be waiting for fastball, so when they get it, they tend to put it in play more often. Regardless, this is an interesting finding; pitchers may want to look at actually reducing their throwing speed slightly to improve their effectiveness, although, more research is needed.

The second finding is that varying speed is just slightly more effective than varying the eye level. This means that a pitcher throwing two pitches at the same eye level, but with one more MPH in speed difference between pitches, has a slightly better chance of a successful batter faced than a pitcher throwing two pitches at the same speed, but with one more inch of difference. This lends additional evidence to the idea that the Pirates should utilize the curveball more often, as it allows the pitchers to vary both the eye level as well as speed, meaning improvements on both fronts.

How did the Pittsburgh Pirates Stack Up in 2017?

The Pitsburgh Pirates ranked 27th in the league in vertical pitch difference at an average of 7.38 inches per pitch, nowhere near the success maximizing value of 10.4 inches. This lends further evidence to the fact that the 2017 was the floor for this pitching staff, as they were able to be mediocre in ERA while being bad at varying eye level.

The Pirates ranked middle of the pack at 20th for speed difference per pitch, but ranked number one overall in the average speed of their pitches. This highlights their ceiling, as the Pirates were able to throw the ball the hardest of any team in the league in 2017, but were able to effectively control it, ranking in the bottom half of the league for both total number of walks issued and walk rate.

This means that if the Pirates can introduce more variation to their pitch height and speed differences, while maintaining their good control, this pitching staff can greatly increase their effectiveness.

If we go to the graph of the team vertical difference compared to their ERA it looks like this:

A quick takeaway is that as a pitching staff increases their vertical difference, the lower their ERA tends to get.

Upon further inspection, six of the eight playoff teams from this season had an average vertical difference of more than eight inches and all of the playoff teams, except Minnesota, had average vertical differences of more than 7.8 inches; Compare that to the Pittsburgh Pirates 7.38 inches of difference.

If the Pirates moved to that 8 inch difference mark, all else equal, their team ERA would’ve dropped from 4.22 to 3.76, good for 5th best in the league. There’s a similar result for FIP, which would have dropped to 3.91 from their 4.23 figure.

Why does this matter?

This is so profound because it’s a change that the Pittsburgh Pirates can make without any change to player skill or ability; all it takes is a change in game calling.

The Pirates have demonstrated the ability to control the pitches they throw, even while throwing hard, so the only difference would be venturing to different areas of the strike zone to keep batters more off balance.

In the past, I’ve been critical of the Pittsburgh Pirates’ “pitch to contact” method, rather than allowing strikeout pitchers to go for strikeouts, this serves as additional evidence in support of that.

Not encouraging their pitchers to change the batter’s eye level allowed for an estimated 65 additional runs given up by the Pirates this past season. Using the Pythagorean Wins formula and all else being equal, this would have put the Pirates at about 82-80, even with the injury problems and suspensions that the Pirates had this season. This isn’t the type of misstep that small to mid-sized market teams can afford to make, particularly when the rest of the division is weak.

The implementation of this idea will have to come from the top and filter down through but shouldn’t be that hard of a sell. Increasing the Pirates use of eye level changes would work to induce more strikeouts to be sure, but also works to generate a lot of weak contact, a goal of the Pirates pitching coaches. This would be a change to the Pirates pitching philosophy, however slight, but it can take the decent pitching staff that the Pirates had in 2017 and turn them into one of the best in the league, practically overnight.

Image Credit – Daniel Decker Photography

Nate Werner

Nate Werner is a senior at Penn State, where he is studying for his B.S. in Economics. He is a lifelong Pirates fan that uses the tools of statistical analysis to dive deeper into the numbers of baseball. His goal is to take the style of analysis used in front offices across the Major Leagues and bring it to the computer screens of everyday fans. You can read some of Nate’s more general analyses of baseball on goldboxstats.wordpress.com and follow him on Twitter @GoldBoxStats.