Dealing with Dead Zone: How to Optimize Fastball Performance
Recently, there has been some attention on pitchers who have a “dead zone” fastball. There are certain MLB teams who will hesitate drafting a pitcher whose fastball profiles like this. In this piece, we’ll examine exactly what a dead zone fastball is, how it performs in a pitcher’s arsenal, and determine whether or not organizations should stray away from players with this type of fastball.
A dead zone fastball is defined as a fastball with equal amount induced vertical break (IVB) and horizontal break (HB). For the rest of the article, when we talk about vertical break, we are using induced vertical break. A fastball with an inch or two difference between IVB and HB can still live in the dead zone, such as a pitch with 16 inches IVB and 14 inches HB. On a pitch movement plot, this would typically be someone who’s fastball falls on the grey diagonal line in the plot below (the line would be mirrored for a left-handed pitcher).
In this piece, we will be utilizing NCAA Trackman data to analyze the dead zone, which is quantified by the difference between vertical movement and horizontal movement. Therefore, dead zone fastballs have a movement difference near zero.
Grouping pitches by movement difference, we use four different performance metrics for each group: Whiff Percentage, Swing Percentage, Chase Percentage, and Weighted On Base Average (wOBA).
Looking Into All Fastballs
Plotting the movement differential against each of these six statistics, we get a visualization that can be examined to look for patterns, especially near the dead zone. Fastballs with more “run”, traditionally known as two-seam fastballs and sinkers, have a negative movement differential, which places them on the left side of the graph. On the other hand, the fastballs with more “ride”, usually four-seam fastballs, have a positive movement differential, placing them on the right side of the plot. A general idea for the dead zone range is constrained by the dotted vertical lines in each visualization.
The smoothed lines in the plot represent a rolling average for the KPI at each movement differential value. Looking at the results, there are clear relationships that can be seen among some of the stats and movement differential. Looking at wOBA, the graph shows that as the movement differential gets closer to the dead zone area, the wOBA increases towards .400 and goes back down as the trend lines go towards “ride” fastballs. Both whiff and chase percentage hold similar patterns, suggesting that hitters are swinging and missing less, and not chasing as much when fastballs are thrown in the dead zone.
However, the concern with including all fastballs is clear from the graph: the variance is wide near the extremes, so the results may be a bit biased. To achieve more accurate results, we must control for velocity and location.
Controlling For Velocity
My first attempt at controlling for velocity looked at “average” velocity fastballs. I filtered my original dataset to only include fastballs in the range from 88 to 91 MPH. The average fastball in our dataset was a little below 90 MPH, so this range is an accurate representation of how hard the majority of the pitchers would throw.
We can start to see some impacts of the dead zone in this result. Whiff and wOBA start to show signs of declined pitcher performance in the dead zone. The wOBA line has higher values overall, implying that average velocity performs worse than above-average velocity. So far, there is not significant evidence to suggest dead zone fastballs are less effective than other fastballs.
With no strong relationship among all KPIs even after controlling for velocity, I wanted to control for location to see how these results could change.
Controlling For Location
I first wanted to examine pitches high in the zone. It’s common for pitchers with more ride on their fastballs to throw up in the zone to generate more swings and misses, so I expect to see that trend in the figure.
The new figure shows fastballs that were located above the average vertical height of all pitches in the database. The first thing I noticed was the whiff percentage held true to what I discussed earlier. On the whiff line, there is a clear positive linear relationship. Looking at other metrics, there is still no strong evidence suggesting dead zone pitches cause a decrease in performance according to wOBA. There is still a slight relationship that can be seen with chase percentage, similar to the previous two visualizations.
Lastly, I wanted to look at pitches low in the zone to see if a relationship would emerge controlling for location. Since high in the zone did not show a strong relationship, I did not expect low pitches to either. This time, I took any pitch that was lower than the average vertical height of all pitches.
To my surprise, pitches located below the average height show a very strong indication that dead zone fastballs perform worse than other fastballs in many metrics. First looking at wOBA, we can see a clear curve that increases as the movement differential gets closer to the dead zone area and goes back down once it goes away from it. The trend line shows roughly a 100-point swing (.300 to .400) in pitches in the dead zone versus ones that aren’t.
Looking at the whiff percentage, a similar relationship can be seen. Batters are not swinging and missing as much on pitches in the dead zone.
How to Deal with Dead Zone
So, what does this all mean? In general, a pitcher who throws a dead zone fastball does not necessarily need to overhaul his entire process to change the shape of his fastball. There are ways to work around it such as throw harder than average or avoid throwing low in the zone as often. A pitcher whose fastball classifies as dead zone can focus on secondary pitches that can play well off each other through tunneling, mirroring pitches, etc. It is up to the player and pitching coach to find creative ways to make a dead zone pitcher’s fastball effective. Moving from a two inch differential to a four inch differential could make a significant impact, so no change is too small to make.
One of the most interesting findings from this study was looking at chase percentage throughout the whole process. I think another conclusion can be drawn from looking at each red line. No matter what the constraint was, or what I was trying to control for, the dead zone fastballs consistently had lower chase percentages. I believe this means that hitters can see the ball better when it is in the dead zone. This could be due to several factors including BP pitching the hitters see, machines producing similar spin direction and movement as dead zone, or just because dead zone pitchers are very common (especially at early levels of baseball), so hitters get used to the movement of a dead zone fastball.
Overall, there is not a major difference in dead zone fastballs compared to ride or run fastballs. Pitches in the dead zone are less effective than those that have more ride or run when thrown low in the zone, as well as when they are thrown at average velocity. However, there are ways to work with a dead zone fastball, and still be effective. The main conclusion to take away is that hitters gain a great advantage by being able to see dead zone fastballs, therefore forcing pitchers to locate pitches well, and execute their secondary stuff.
In the future, I would like to explore how changing the definition of dead zone impacts my results from this analysis. Creating an expected spin direction model, I could compare the expected spin direction with the actual spin direction. The actual spin direction would be considered the “dead zone” for each pitcher. This is the profile that a batter would expect the pitcher’s fastball to show. The closer a pitcher is to his expected spin direction, the more they are classified as having a dead zone fastball.
