Pitch Tunneling: Is It Real? And How Do Pitchers Actually Pitch?
The concept of pitch tunneling is gaining popularity. The idea is that two different pitches fly down the same trajectory long enough to look nearly identical through the point when a hitter must decide whether, or not, to swing. If pitches thrown back to back travel down this same “tunnel” long enough, a hitter won’t be able to tell them apart until it’s too late. There also seems to be some confusion about the idea itself – is it mechanical, in which pitchers should change their deliveries to improve tunneling? Or, is it simply a matter of pitch sequencing? And, if tunneling is only a sequencing concept, are, and should, pitchers be using it? Let’s find answers to these three questions.
Part 1: Pitch Tunneling Is Not a Mechanical Concept: A Distinction Without a Difference
At Baseball Prospectus, Jeff Long, Jonathan Judge and Harry Pavlidis have identified the “tunnel point” as 23.8 feet before home plate. The goal is that two pitches, thrown back-to-back, would fly through the same narrow tunnel, basically making them appear to be the same pitch.
If a pitcher can’t control his body well enough to precisely repeat his delivery, then his pitch tunneling ability suffers – the hand will be in a noticeably different position at the release of every pitch. And for a pitcher who repeats his delivery exceptionally well, like Jon Lester does, the pitches he throws will look very similar for as long as possible. Again, if you repeat your delivery very well, tunneling will happen on its own if you pair pitches with the same starting points. If you don’t, it won’t. Pitchers don’t need to practice throwing pitches through small hoops 20ish feet from the plate. Rather, they simply need to practice pitching with a focus on doing the exact same thing with their body on every single pitch.
If it is unclear to you whether pitch tunneling is a physical concept worthy of its own term, let me assure you that it is not. The main thing a pitcher can control is the repeatability of delivery, which in turn directly affects release point. Coaches have always emphasized to pitchers: Practice repeating your delivery! If you want to throw more strikes, learn to repeat your delivery!
Repeating ones delivery means that all of the following variables are held constant between pitches. The following are well-controlled by pro pitchers, and, loosely controlled by amateurs.
- Delivery footwork
- Tempo
- Stride length
- Stride direction
- Actual velocity
- Intensity
- Finish and follow through
- Location on the rubber
Having the proper conditioning to prevent fatigue is also a factor.
From a mechanical perspective, we have a distinction without a difference – “repeating the delivery” is all that is required to physically create tunneling. A pitcher who repeats his delivery sets the stage for all pitches to tunnel provided they use the same starting point, the same initial trajectory. And, to shed light which is explained in the next section. Whether he chooses to do this, however, is another story, as this concept hinges on pitch selection.
Part 2: How Pitchers Create Tunneling: Focal Point = Tunnel
If a pitcher repeats his delivery, then the flight of each pitch, to each location, is essentially predetermined by physics. To make them take the same tunnel, then a pitcher would need to pair pitches based on where they start, not where they end. The deviation from a tunnel, for a pitcher who repeats his delivery well, will then only come from deviations in starting location, or focal point (used interchangeably). As a pitcher, one of the most important concepts I learned, and now teach, is how to command my “moving” pitches. For lack of a simpler term, I’ve been calling it pitch vectoring, which stems from a high school physics lesson.
Your goal is to swim straight across a flowing river. If you swim exactly toward point B, from point A, would you make it precisely where you intended? The answer is no: the current would take you with it, so that you would end up on the other side but down-river a bit. Your initial heading, your starting point, the point on the opposite river bank you focused on reaching, would not be where you ended up. Most young pitchers, when asked, have no idea how they locate their curveball or slider. “Where do you look?” I ask. “The mitt.” They reply. If you ask a pro pitcher what happened on a specific pitch that didn’t go well, he’ll often respond, “Well, I tried to start it on the inner half…”
If a pitcher looks at the mitt and wants his curveball to end up at the mitt…how does he account for the 14 inches it breaks? He can’t, and the only way a pitcher can reliably control and locate his moving pitches is if he establishes a deviation point or focal point for each pitch. Sinkers and two-seamers also are thrown within this framework, but aren’t traditionally referred to as breaking pitches.
For a four-seam fastball, where we look is where the pitch ends up – it becomes the pitcher’s visual set point. So, for a curveball to hit the catcher’s target, a pitcher will typically “start” it at the catcher’s forehead. He’ll then focus his eyes, point his front shoulder, and finally move his chest, then hand, toward that focal point. That, in a sense, is how pitchers actually locate their pitches. The focal point changes depending on the break of the pitch. This sounds like pitching 101, but it’s not; I wasn’t taught this until I was 21 years old, three years into my Division-I baseball career. The illustration below shows the pitcher’s point of view and how his entire arsenal would break out from one fastball starting point, or tunnel. Each pitcher’s breakout would vary depending on what they throw and how much each pitch breaks. The takeaway here is that a pitcher can’t possibly tunnel two different pitches that end up in the location – he can only start them at the same focal point, intending them to end at different points.
Part 3: Is Tunneling Considered in Pitch Calling?
Starting a pitch at the same focal point is the same as starting it down the same tunnel. This is the only thing a pitcher truly controls, the action item of this whole pitch tunneling concept, which leaves a pitcher with only predetermined options. To throw a high fastball and then follow it up with a “tunneled” slider leaves us with, well, a slider that will end up mid-thigh on the outer half (shown in the previous illustration). In that example, we must remember that good tunneling is not the goal: getting outs is the goal. Throwing mid-thigh sliders won’t, over the long haul, help that goal, even if they look like fastballs at first. The right pitch to choose depends on the pitcher’s arsenal, strengths and weaknesses, situational context, and more.
You might be at home thinking, plan the fastball better, so that then the next pitch fits into that tunnel more effectively! That’s a good thought, but if you were to plan out each pitch before seeing how the hitter reacts, you’d be very ineffective. High-level pitchers do start each at-bat with a plan, but it evolves with each pitch according to the following:
- Did I hit my spot?
- Ball or strike?
- Swing or no swing?
- Swing and miss or foul?
- How did he foul it off?
- How did he miss? Early or late?
- How did his swing look?
- How did he stride?
- If an off-speed pitch, was he on his front foot?
- How was the quality of the pitch I threw?
Pitches Are Chosen for Their Expected Outcome
Let’s look at a common example. Think of a typical, hard-throwing late-innings pitcher with an upper-90s fastball and slider. These hard-throwing pitchers will usually manufacture strikeouts with two pitch variations:
- A buried (in the dirt) slider
- A high fastball
If the game is on the line and he has two strikes on a hitter, good money is on one of those pitches being thrown. The problem, though, is that these pitches will never be in the same tunnel. Consider the following two diagrams that show the different focal points, or tunnels:
The high, strikeout fastball (letter-high) will never look like the low, strikeout slider that breaks into the dirt – a pitcher has no way to get them in the same tunnel. Yet, both pitches get high percentages of swings and misses because when thrown in those locations, hitters are often left with a zero-sum outcome – take the pitch for a ball, or swing at it and almost definitely miss. A pitcher with two strikes on the hitter would never choose to follow up a high fastball with a mid-thigh slider; he’d always follow it up with a slider down, where it would get the desired result. The two wouldn’t tunnel, but it wouldn’t matter because the hitter still wouldn’t feel comfortable taking either in a two-strike count, nor would be find it easy to lay off. The pitcher would choose either pitch, even without a tunneling effect, because he likes the expected outcome.
So, Does Tunneling Work?
How effective are tunneled pitches? I suggest reading this excellent article by Jon Roegele here at THT. His analysis shows how pairs of tunneled pitches, which he refers to as “in the band,” get a higher-than-average swing-and-miss percentage.
The closer consecutive pitches are to one another at the swing decision point, the less distance apart they need to arrive at home plate to generate higher than normal swing and miss rates.
There’s no doubt that there’s legitimacy to this concept; when we pair pitches that share a tunnel, hitters get fooled, and whiff more often. The problem, though, is in implementation. Pitchers tend to choose pitches, hedging risk, on expected outcome. Sometimes tunneling aligns with this, sometimes it doesn’t.
If a hitter has a slow bat, opposite-field approach, or a long swing, a smart pitcher might throw him fastballs inside the entire at-bat, until the hitter is out or proves he can adjust. This doesn’t run counter to the effectiveness of pitch tunneling, just illustrates that trying to choose pitches for tunnel effect often comes secondary to pitching to an expected outcome. For this opposite-field hitter, every inside fastball has an expected outcome of either weak contact or no contact. Deception isn’t required – all pitches will look the same because they are the same, but it doesn’t matter. As a teammate of mine once put it, “hitters often know what pitch is coming but still can’t do anything about it.”
Additionally, pitchers choose pitches that may appear to be chosen based on tunneling, when in fact they’re paired for expected outcome.
For example, fastball-curveball pitchers don’t necessarily pair high fastballs with curveballs because they look the same. Rather, they pair them because both pitches are swing-and-miss pitches. This is an important distinction. The only fastball on which a pitcher will reliably get swings and misses is a high fastball. All other fastball trajectories either look like a ball out of the hand, or are in the strike zone just enough to be put in play. The one location where a pitcher feels safe with a straight fastball is up. Everything else is left to chance. With two strikes, I’ve thrown five fastballs in a row at the letters, knowing that if I hit that spot, I’ll get a strikeout, or at worst a pop-up, sooner or later. No other location has that power.
The letter-high fastball is often paired with a curveball that bounces. But, even if we assume that the high fastball and the bouncing curveball have the same tunnel (they likely don’t, as pitchers usually start a bounced curveball at the catcher’s chest), those two in the pair are both chosen for their zero-sum outcome – ball, or strikeout. They both have that same purpose, regardless of what deception they might produce. If I was hunting for a curveball strikeout, I’d never throw it in the strike zone intentionally to pair with a high-fastball’s tunnel. Rather, I’d throw it so it bounced on the point of the plate, because that gives me the best expected outcome. Most high fastballs will share a tunnel with a curveball that breaks into, not out of, the strike zone. So, the pair doesn’t go both ways.
It’s also advantageous to not chase swings and misses at all times. In many cases, pitchers will go out of a tunnel to induce contact and end an at-bat quicker. An example a fastball that ends up low and away, followed by a slider that ends up the same spot. A slider that tunnels would break off the plate, which would more likely induce a swing and miss. But, sliders that break to the outside corner are difficult to hit hard, and a pitcher might, in this case, go out of the tunnel intentionally to induce weak contact to keep his pitch count low.
Conclusion
Assuming a pitcher can repeat his delivery well, pitch tunneling is most relevant as a concept of pitch sequencing. Though a pitcher’s ability to use tunneled sequences may increase deception, pitchers select sequences based on expected outcomes. If the expected outcome of the pitch is the best of all considered, then that is the right pitch, regardless of whether it follows the previous pitch tunnel, or even if the hitter is expecting it. Though pitch tunnels are intriguing and can certainly aid pitchers in making good pitch-calling decisions, expected outcomes will likely remain the gold standard.
References & Resources
- Baseball Prospectus, Pitch Tunnel Pair Statistics
- Jeff Long, Jonathan Judge and Harry Pavlidis, Baseball Prospectus, “Prospectus Feature: Introducing Pitch Tunnels”
- Jon Roegele, The Hardball Times, “The Effects of Pitch Sequencing”
- Darius Austin, Baseball Prospectus Wrigleyville, “Jon Lester and Pitch Tunnels”
- Daniel Schwartz, FanGraphs, “The Importance of Release Point Consistency”
- Eno Sarris, FanGraphs, “What Can Hitters Actually See Out of a Pitcher’s Hand?”
Very interesting article! Being a Giants fan, this made me think of Madison Bumgarner. I think I remember hearing about him making an effort to have all of his pitches essentially look the same. I’m wondering if his repertoire is a case where the tunneling concept makes sense. Just based off of watching, he seems to have 4 basic pitch type/location combinations to a right-handed batter: high fastball, curveball in the zone, cutter/slider inside, and curveball in the dirt. Could the first three exist within the same tunnel? If so it seems like he works within the tunneling concept until he gets to two strikes, when he basically just pounds high fastballs or buries curveballs in the dirt. Thanks!
Hey Michael,
Man, Bumgarner is good. Tons of deception in addition to a great repertoire. His high-fastball/curve for a strike combo probably tunnel, but there’s no way the slider or cutter are coming out of that same window…that would leave the cutter middle-up and away, almost ditto with the slider.
But, that’s OK – because even if those two don’t match up in the same tunnel, if they break to the inside corner of the plate, the expected outcome he assumes he’s getting is a jam-job or swing and miss.
Thing to remember, just like with Zach Britton and Mariano Rivera…they can know it’s coming and still not be able to hit it.
You make a strong case. Can we quantify it by run value?
Take 0-0 count and a particular pitcher and batter. Each ‘call’ (pitch + location) has an estimated run value. A mix of two calls may have a “tunneling bonus” on top, from unpredictability. How big is the tunneling bonus in run value, compared to the base values?
Fastball/curve is the pairing with the biggest difference in movement, so it can best be placed onto the zone without one prong landing in a fat location.
That’s an interesting idea.
Expected outcome is really a factor of a lot of different things, like the hitter’s tendencies, swings on previous pitches, his previous at-bats, previous week at the plate, who’s on deck, etc. All the info a pitcher gathers on the fly gives him an idea of what he thinks will give him the best chance on the next pitch.
But if we layered run value as the metric for expected outcome with pitch tunneling, maybe that puts it all together, as you suggest. Reminds me a little of Zach Greinke’s thoughts on pitching away versus inside…that, at worst, most hitters hit a 300 foot rocket to the opposite field outfielder. Thanks for reading!
1) How wide is the “good enough” tunnel? Tunneling through a point is the max bonus, but tunneling through a four-inch circle might also be enough to mess with a batter’s swing decision. Or a larger circle to a lesser degree — still harder to decide on than two wildly different locations.
2) Does tunneling make more sense when a pitcher is okay with weak contact, versus needing a whiff? Location obviously matters for quality of contact, so I’m not sure it works out, but it might have less need to hit the absolute corners of the zone.
3) You can’t tunnel a high fastball with a buried slider, but nobody has perfect command anyway. The hitter is also looking for hung sliders and centered fastballs. A buried slider tunnels with a centered fastball. … Eh, I think this is true but not useful, in that it doesn’t affect how to pitch, in terms of pitch and location. You still try to hit good spots, this “bad/good tunneling” is just a fact about what happens when you miss. Says “it’s good to mix your pitches and leave uncertainty” I guess.
Looks like we’re reading this at the same time!
1. I think “good enough” is larger than most would assume, my guess is the size of a volleyball. I didn’t see much correlation with pitcher release point consistency and their ERA, WHIP, etc. among big leaguers. Among amateurs, hitters may still pick up the large differences in amateur pitcher release points, but even then…just because a hitter detects a difference doesn’t mean he’ll choose to not swing at it if it’s a ball, or make solid contact. Hitting is hard.
2. Per weak contact, really I think it makes a lot of sense going out of the tunnel looking for weak contact…because we want the hitter to swing. most of those are going to originate middle and up, and break toward the corners, but not off the plate.
3. I think we’re on the same page!
Dan I think you Blewett, your exapmle of a hard throwing closer with the wipeout slider is a good example of tunneling because the slider comes out in the same tunnel the batter reacts fastball and gets slider. His above average fastball and slider work in the same tunnel which makes him able to get outs with two pitches. You are trying to put his up and in fastball into a tunnel but it doesn’t belong there. It’s a purpose pitch to keep a batter out of rhythm, the tunnels are important because it forces the batter to guess and reduce his odds of getting a hit. If you did the numbers on number of pitches that were tunneled to the lower corners of the plate I would guess that the most successful pitchers have three pitches they can tunnel in both those zones allowing them to beat the platoon split. Of course you would have to include all types of fastballs.
Thanks for taking the time to mock my last name. Reminds me of my playing days, when the people who paid to sit in the stands and watch me pitch, overweight and 20 years removed from their little league glory days, would heckle me. Have a great week.
saw it in action last night in a different way, my son (high school age) got 2 balls on very close 2 seamers ump said they were just low to a lefty hitter, so on 3-1 count the hitter swung at 2 balls low and away in the dirt for the k, looked like the same pitch the batter wasnt interested in the previous 2 pitches. I was puzzled and after the inning asked. Both changeups that even fooled me as the observer. Not to mention my son has very low confidence in that pitch perhaps until now.
Thanks for sharing please do share such post