Velocity is a floor, not a ceiling. The pitchers who survive the jump to the majors are not the ones who throw hardest - they are the ones whose arsenals consistently break what hitters predict. A deep guide to pitch tunneling, IVB, VAA, and what actually separates minor league arms from major league starters.
You have seen this before. A pitching prospect is lighting up Double-A. Mid-90s fastball, strikeout rate pushing 30 percent, ERA in the low twos. The reports say the stuff is electric. The stat line says he's ready. Then he gets to Triple-A, or the majors, and suddenly the fastball nobody could touch is getting fouled off. The breaking ball that buried minor league hitters is getting spit on. The swing-and-miss is gone, even though the radar gun still says 96.
Nothing changed. Except the hitters.
Modern hitters are not losing to velocity. They are losing to prediction errors. The pitchers who succeed at the highest level are the ones whose arsenals consistently break those predictions - not just once, but on every single pitch, across a full at-bat, in a way that compounds over the course of a game. Understanding why requires going deeper than strikeout rates or Stuff+ grades. It requires understanding how the brain processes a baseball in flight.
Every swing starts as a prediction problem. A major league fastball travels from the pitcher's release point to the catcher's mitt in roughly 400 milliseconds. A hitter's swing takes approximately 100 milliseconds to complete. For a league-average fastball, the hitter has roughly 175 milliseconds to commit to a swing. That leaves a decision window of approximately 50 milliseconds - a window so small that true reaction is essentially impossible at high velocity.
What hitters are actually doing is pattern recognition. Within the first 100 to 150 milliseconds after release, the brain extracts information from release height, arm angle, extension, spin direction, and early trajectory. From that data, it constructs a prediction: where will this pitch cross the plate, and when? The swing launches to intersect that predicted location. The hitter is not reacting to the ball at contact - he committed 200 milliseconds earlier.
Pitchers win when the ball does not end up where the prediction said it would.
This is not a new insight, but its implications for how we evaluate pitching prospects are more radical than most analysis acknowledges. A prospect with a 97 mph fastball and a 30 percent strikeout rate in Double-A may be winning primarily because minor league hitters are building poor predictions from inadequate exposure to his release. Against major league hitters with more reps, better pattern libraries, and video rooms full of data, the same arsenal may be immediately readable. The stuff has not changed. The prediction quality has.
A baseball thrown with backspin - the way a four-seam fastball is thrown - experiences an upward Magnus force as it travels toward the plate. Gravity is simultaneously pulling it down. The interaction between these two forces determines how the ball actually moves through space.
Induced Vertical Break measures how much the pitch resists gravity relative to a theoretically spinless ball following a pure ballistic path. For context, league-average IVB on a four-seam fastball is approximately 15 to 16 inches. Elite four-seamers reach 19 to 21 inches. A four-seam fastball with 18 inches of IVB does not literally rise - that would violate physics. What it does is fall significantly less than the hitter's brain expects based on the early flight trajectory. The brain sees the ball leave the hand, calculates a descent curve based on the perceived speed and early path, and generates a prediction. The high-IVB fastball falls off that prediction curve. It arrives higher than expected.
The practical effect is striking. Hitters trained to match their swing path to the expected descent of a typical fastball consistently swing underneath high-IVB pitches at the top of the zone. The contact that does occur is often on the bottom half of the ball - weak pop-ups, harmless flyouts, and the occasional miss entirely.
This is why the swing-and-miss rate on four-seam fastballs is heavily correlated with IVB. A pitcher with 18+ inches of IVB and consistent release height can locate the same fastball to the same zone and generate whiffs repeatedly, not because hitters do not know what is coming, but because the ball repeatedly ends up outside the range of their pre-committed swing path.
| IVB Range | SwStr% on Four-Seam (approx) | Evaluation | Projection |
|---|---|---|---|
| Below 12 inches | 4–7% | Below average | Contact pitch without elite command |
| 12–15 inches | 7–10% | Average | Functional, needs strong secondaries |
| 15–18 inches | 10–13% | Good | Solid strikeout pitch in the upper zone |
| 18+ inches | 13–16%+ | Elite | Bat-missing weapon at the top of the zone |
Vertical Approach Angle describes the angle at which a pitch enters the strike zone. A steep VAA means the pitch is descending sharply as it crosses the plate. A flat VAA means it is arriving on a more horizontal plane. For reference, the MLB average VAA on four-seam fastballs is approximately -4.4 degrees. Elite flat fastballs reach -3.5 to -3.7 degrees. That fraction of a degree represents a meaningful difference in whiff rate at the top of the zone, where the flattest fastballs generate swing-and-miss rates 8 to 9 percentage points higher than average.
Hitter swing paths are not horizontal - they angle upward through the zone to match the expected descending plane of the pitch. This is optimal for pitches that enter the zone at the angles hitters are expecting. But a fastball with both high IVB and a flat VAA creates a specific kind of problem: the pitch stays on a near-horizontal plane for longer than expected before arriving, and when it does arrive, it is higher in the zone than the swing was designed to intersect.
Flat VAA is primarily a function of extension and release height. Pitchers who release the ball closer to home plate (greater extension) reduce the angle at which the pitch descends before crossing the plate. Pitchers with high release heights generate naturally flatter approach angles at the top of the zone. Overhand arm slots tend to produce steeper VAA; lower three-quarter slots produce flatter angles.
The interaction between IVB and VAA is more important than either metric alone. A pitch with elite IVB but steep VAA still arrives at a predictable angle and can be accounted for. A pitch with moderate IVB but very flat VAA creates a different kind of deception - the ball stays on a plane the hitter is not prepared for. The combination of both - high IVB and flat VAA - is the theoretical ideal for a strikeout fastball at the top of the zone.
| VAA at Plate | Effect | Best Paired With |
|---|---|---|
| Steeper than -5° | Arrives at predictable descent, matches hitter swing plane well | Strong breaking balls below zone to create contrast |
| -4° to -5° | Average approach, functional | Mid-range IVB with plus command |
| -3° to -4° | Flatter than expected, creates swing-plane mismatch | High IVB - both effects compound |
| Flatter than -3° | Elite, stays on plane longest | Any breaking ball with steep VAA for vertical contrast |
IVB and VAA describe individual pitches. Tunneling describes how pitches interact as a set.
The concept is straightforward: pitches that share early flight path before diverging late force the hitter into a specific kind of error. If a fastball and a curveball emerge from the same release point and follow the same initial trajectory for the first 23.8 feet of flight, the hitter cannot distinguish between them until the decision window has essentially closed. Baseball Prospectus research established 23.8 feet as the precise tunnel point, the distance at which a 175-millisecond swing commitment must begin for a league-average fastball. Greg Maddux described this instinctively long before the data existed: his goal, he said, was to make all his pitches look like a column of milk coming toward home plate. Not different types of pitches. A single column. The quantification of that instinct is what pitch tunneling research formalizes. The fastball continues toward the top of the zone. The curveball drops below it. Both looked identical until it was too late to adjust.
Good tunneling is not accidental. It requires matching three things: release height and angle, early trajectory, and perceived velocity window. A curveball thrown with the same arm action as the fastball, from the same release height, at a velocity close enough that the early speed signal does not immediately distinguish it - that pitch will tunnel well. A curveball thrown from a lower release point with dramatically different arm action will telegraph itself regardless of movement quality.
This is where the gap between minor league and major league performance often lives. Minor league hitters have limited exposure, less advanced pitch recognition, and frequently face pitchers whose arsenals work on individual pitch quality alone. A 97 mph fastball and a loopy curveball with dramatically different arm action can still generate strikeouts in Double-A because the hitters simply have not seen enough of the pitcher to build an accurate prediction model. Major league hitters with video rooms, advance scouting, and thousands of reps against similar stuff solve that problem in a fraction of the at-bats a minor league hitter would need.
The question for evaluators is not whether the fastball misses bats in Double-A. It is whether the fastball and the breaking ball share enough early trajectory that the pitch sequence creates decision problems. Whether the changeup mirrors the fastball's initial flight before introducing horizontal or downward movement. Whether the pitcher's release consistency is good enough that multiple pitches can actually share the same tunnel window.
No amount of pitch quality matters if the delivery is inconsistent. Release point variation is one of the clearest signals of a pitcher who will struggle against major league hitters. When the release point moves - even by a fraction of an inch - the early trajectory changes, the tunnel breaks down, and hitters receive more distinguishing information earlier in the flight. A pitcher who tunnels beautifully when fresh and falls off his release point by the fourth inning is effectively showing hitters two different arsenals.
Extension adds a separate dimension. Every additional inch of extension toward home plate has two effects: it reduces the time the hitter has to process the pitch, and it flattens the VAA. For a pitcher releasing at 6.5 feet of extension versus 6.0 feet, the difference in perceived velocity is roughly 1 to 1.5 mph - not from actually throwing harder, but from delivering the same velocity from a point closer to the plate. That compression is meaningful at the major league level where fractions of a mph in perceived velocity affect swing timing decisions.
The combination of elite extension, consistent release point, and pitches that actually share a tunnel window is rarer than the prospect industry typically acknowledges. Finding a pitcher who has all three is finding a genuine major league arm - regardless of what the radar gun says.
A 30 percent strikeout rate in Double-A is meaningful. It is not proof of a major league arm. The distinction matters enormously for prospect evaluation, and collapsing the two is one of the most common analytical errors in amateur scouting.
Minor league hitters are solving a different problem than major league hitters. They have less exposure, less advance information, and are frequently facing pitchers in their first sustained run through a level. A pitcher who is winning those at-bats on raw stuff - velocity alone, or a single elite pitch without an effective tunnel partner - may generate impressive strikeout totals against opponents who simply have not seen him enough to build an accurate prediction model.
The more useful question is: does this pitcher's arsenal create genuine decision problems, or is it creating recency problems? The difference is whether the pitch combinations force hitters into prediction errors on the physics of the pitches, or whether hitters are simply failing to recognize a pitch they have not seen much of. The first is a structural advantage that scales upward. The second is a temporary advantage that disappears.
Metrics that help answer this distinction: SwStr% on individual pitches (is the swing-and-miss coming from genuinely good pitch shapes or from hitters swinging at pitches they cannot see?), release point consistency over the course of outings (does the pitcher maintain his tunnel window deep in games?), and pitch mix distribution (does he use his secondary pitches enough to establish them as threats, or does he lean on one offering and dare hitters to lay off everything else?).
Abstract principles are only useful when they connect to real players. Here is how the tunneling and IVB framework applies to specific pitching prospects currently tracked in the model or relevant to the 2026 class.
Modern MLB pitchers throw 11 distinct pitch types. Each works through a different physical mechanism. Understanding what each pitch does, how it creates deception, and what arsenal role it fills is the foundation for evaluating any pitching prospect.
The baseline pitch in every arsenal. Thrown with all four seams rotating through the fingers, generating backspin and the Magnus force that produces IVB. A four-seam fastball does not rise, it falls less than the hitter expects. Elite four-seamers above 18 inches of IVB generate whiffs at the top of the zone because the brain predicts a descent that does not happen.
The pitch lives or dies by IVB, VAA, and location. A 97 mph four-seamer thrown down the middle is less effective than a 93 mph four-seamer with 20 inches of IVB at the letters.
Where a four-seamer fights gravity, a sinker works with it. Thrown with the fingers aligned along the two seams, generating sidespin and topspin that produces both downward and arm-side horizontal movement. The result is a pitch that arrives lower and further inside than a four-seam from the same release point.
Sinkers generate groundballs through contact, not swings and misses. The best sinkers in baseball add seam-shifted wake, where seam orientation forces asymmetric airflow separation, adding 6-9 inches of extra drop beyond what spin alone predicts. This is what makes certain sinkers feel "heavy" to hitters.
The cutter sits between a fastball and a slider in both velocity and movement. Thrown with slight off-center pressure on the fingertips, it generates just enough gyroscopic axis shift to move glove-side with less IVB than a four-seamer. The pitch arrives looking like a fastball before cutting late across the plate.
The cutter's value is its location against opposite-handed hitters. A right-handed pitcher's cutter runs into the hands of left-handed hitters, breaking bats and producing weak contact. Mariano Rivera turned this into the most dominant single-pitch career in baseball history. Corbin Burnes built a cutter-primary arsenal to become a Cy Young winner.
The oldest breaking ball in baseball. Thrown with 12-to-6 topspin, the curveball drops sharply due to downward Magnus force. A good curveball from a high release point can drop 14-16 inches below where a same-velocity pitch without spin would land. The best curveballs have both vertical drop and late horizontal break, making them difficult to track in two planes simultaneously.
The curveball's size is its advantage and disadvantage. The dramatic drop generates swings and misses below the zone, but the long arc from release to contact means hitters have more time to identify it. Release point and tunnel quality determine whether a curveball is elite or hittable.
The slider is the most-thrown breaking ball in modern baseball. Thrown with a mix of side spin and some topspin, it produces both downward and glove-side movement at higher velocity than a curveball. The slider's key advantage is its combination of movement and speed, making it harder to identify than a curveball and harder to hit squarely than a fastball.
The slider sits in the middle of the velocity spectrum, which creates its best tunneling opportunities. Off a fastball, the hitter sees the same arm speed and early path before the ball breaks. The most effective sliders generate 8+ inches of horizontal movement with 4-6 inches of vertical drop, creating movement in two planes that is difficult to track.
The sweeper is the signature pitch of the modern pitching revolution. Thrown with a nearly pure gyroscopic spin axis, it generates extreme horizontal movement with minimal vertical drop, essentially sweeping across the zone from the pitcher's arm side to glove side. Average sweeper horizontal break is 16-20 inches. Elite sweepers reach 22-24 inches.
The sweeper works differently from a traditional slider. A slider has a mix of side and top spin, producing both horizontal and vertical movement. A sweeper has almost pure side spin, producing almost purely horizontal sweep. Against same-sided hitters, the sweeper exits the strike zone entirely, chasing hitters into a pitch they cannot make contact with. Against opposite-handed hitters, it breaks across the plate as a strike. The Yankees called it the "whirly" before "sweeper" took hold around 2022. Shohei Ohtani, Adam Ottavino, Yu Darvish, and Nestor Cortes are among the most prominent practitioners. Adam Ottavino's sweeper is widely considered the league's best, breaking up to 20 inches horizontally.
The slurve sits between a slider and a curveball in both movement profile and mechanics. It carries more horizontal sweep than a traditional curveball and more vertical drop than a true slider. The result is a pitch that moves in both planes simultaneously, making it difficult to track. Pitches with two-plane movement are most effective when the arm angle and release point match the fastball perfectly. Clayton Kershaw is the canonical example of a pitcher whose breaking pitches share their fastball's arm slot, and his slider has been described as a slurve at times in his career.
The slurve is harder to command than a pure slider or pure curveball because the arm angle must be precisely maintained to reproduce the two-plane movement. Pitchers who can repeat it, though, have the most deceptive breaking ball in baseball.
The changeup is the pitch most dependent on tunnel quality to be effective. Thrown to match the arm speed and early trajectory of a fastball at 8-12 mph slower, it forces a timing error rather than a tracking error. The hitter commits based on fastball-speed prediction and the bat arrives early. The late arm-side fade adds a second deception layer after the initial timing disruption.
There are three main grip types, each with different movement profiles. Circle changeup: maximum arm-side fade and depth. Palmball: similar to circle with lower spin. Vulcan changeup: split between middle and ring fingers for a tumbling action. All three work by reducing backspin relative to a fastball, letting gravity take over late in the flight.
The split-finger fastball is thrown with fingers spread wide on either side of the ball, reducing backspin to near zero. The result is a pitch that starts on a fastball plane before dropping sharply below the zone, often 6-10 inches more than a fastball from the same release point. The splitter's primary weapon is the late drop after the tunnel point, arriving where the fastball suggested before diving out of the zone.
Splitters are more physically demanding than most pitches and are associated with higher arm injury rates. They are also among the most effective put-away pitches when healthy. Paul Skenes throws a splitter-sinker hybrid called the splinker that combines the dive of a splitter with the arm-side fade of a sinker, producing movement that generates the highest whiff rate of any pitch in his arsenal.
The knuckleball is the only pitch in baseball that works by eliminating spin rather than creating it. Thrown with fingernails or fingertips digging into the seams to prevent rotation, the ball tumbles through air currents with no Magnus force to stabilize it. The result is an unpredictable, wobbling trajectory that neither the hitter nor the catcher can track reliably.
The knuckleball is not part of modern prospect development. No current Top 100 prospect throws one as a primary pitch. It requires years of specialized development and creates serious challenges for catchers and pitch-framing. R.A. Dickey won a Cy Young with it in 2012. It is mentioned here for completeness, not as an arsenal building block.
| Pitch | Family | Velo (mph) | IVB (in) | H-Break (in) | Spin (RPM) | Primary Weapon |
|---|---|---|---|---|---|---|
| Four-Seam Fastball | Fastball | 92-101 | +15 to +21 | 0-4 (arm) | 2200+ | IVB + VAA at top of zone |
| Sinker / Two-Seam | Fastball | 90-97 | +4 to +10 | 8-14 (arm) | 2000+ | Groundballs + SSW drop |
| Cutter | Fastball | 88-95 | +8 to +14 | 2-6 (glove) | 2400+ | Late glove-side cut |
| Curveball | Breaking | 73-84 | -10 to -16 | 6-12 (glove) | 2500+ | Vertical drop below zone |
| Slider | Breaking | 82-90 | -2 to -8 | 5-10 (glove) | 2600+ | Horizontal break + drop |
| Sweeper | Breaking | 80-88 | 0 to -4 | 16-22 (glove) | 2800+ | Extreme horizontal sweep |
| Slurve | Breaking | 79-86 | -6 to -12 | 8-14 (glove) | 2500+ | Two-plane movement |
| Changeup | Offspeed | 82-90 | +6 to +14 | 10-16 (arm) | 1500-1800 | Timing disruption + fade |
| Splitter | Offspeed | 85-91 | +2 to +8 | 4-8 (arm) | 1400-1700 | Late dive off fastball tunnel |
| Knuckleball | Specialty | 65-80 | Unpredictable | Unpredictable | <25 | No-spin turbulence |
IVB positive = pitch resists gravity (rises relative to spinless). IVB negative = pitch drops faster than spinless. H-Break: arm-side = moves toward throwing arm side, glove-side = moves toward glove side. All ranges represent MLB pitch-tracking averages across 2024-2025 Statcast data.
Even a perfectly tunneling two-pitch combination has a fundamental problem: with only two pitches, hitters face binary prediction. They are either right or wrong. In a binary system, even partial information is useful - a hitter who eliminates one pitch from the possibility space has a 50/50 guess on every pitch, plus whatever directional information the partial signal provides.
A three-pitch arsenal reduces that to a third-probability problem even before the tunnel starts providing information. A four-pitch arsenal with strong tunnel relationships between multiple pairs - fastball and changeup tunneling in one direction, fastball and curveball tunneling in another - creates a prediction problem that cannot be solved with available information. There are simply too many branches in the decision tree for the available processing time.
This is why pitch mix distribution matters as much as individual pitch quality in prospect evaluation. A pitcher who throws his fastball 70 percent of the time has told hitters that the slider or curveball appears roughly 30 percent of the time. Hitters can sit fastball, look for the breaking ball as a change-of-plan, and function at a reasonable success rate. A pitcher who distributes pitches more evenly - 45 percent fastball, 30 percent slider, 25 percent changeup - forces hitters into a much harder prediction problem before any single pitch has even been released.
This is one of the reasons Corbin Burnes' reinvention was so effective. By moving away from the fastball as a primary pitch and building a cutter-heavy attack with multiple complementary offerings, he transformed a binary prediction problem into a multi-branch one. The individual pitches are excellent. But the distribution and sequencing compound that excellence into something that resists adjustment in a way that raw stuff alone cannot achieve.
Induced vertical break measures how much a pitch resists gravity compared to a theoretically spinless ball thrown at the same release point. League-average IVB on a four-seam fastball is 15 to 16 inches. Elite four-seamers reach 19 to 21 inches. High IVB means the ball arrives higher than the hitter's brain predicted, generating swing-and-miss at the top of the zone.
Read more in our baseball metrics handbook.
Vertical approach angle describes the angle at which a pitch enters the strike zone. The MLB average VAA on four-seam fastballs is approximately -4.4 degrees. Flatter VAA closer to -3.7 means the pitch stays on a horizontal plane longer, creating swing-plane mismatches. Combined with high IVB, flat VAA produces the highest whiff rates at the top of the zone.
Pitch tunneling is when two pitches share the same early flight path before diverging late. The tunnel point, established by Baseball Prospectus research, is 23.8 feet from release. That is the distance at which a hitter must begin their swing commitment. Pitches that look identical through that point force the hitter to guess, regardless of their individual movement quality.
The tunnel is the deception itself. Two pitches with identical movement that come from different release points are easier to identify than two pitches with different movement that share the same release path.
Modern MLB pitchers throw 10 distinct pitch types. The fastball family includes the four-seam, sinker (two-seam), and cutter. The breaking ball family includes the curveball, slider, sweeper, and slurve. The offspeed family includes the changeup and splitter. The knuckleball stands alone as a specialty pitch.
Each pitch generates whiffs through a different mechanism. Fastballs use IVB and VAA. Breaking balls use late horizontal or vertical movement. Offspeed pitches use timing disruption against a fastball tunnel. The complete reference is in the 2026 Pitch Atlas above.
A sweeper is a breaking ball thrown with a near-pure gyroscopic spin axis that produces extreme horizontal movement with minimal vertical drop. Average sweeper horizontal break is 16 to 20 inches. Elite sweepers reach 22 to 24 inches.
The pitch differs from a traditional slider by prioritizing side-to-side sweep over downward bite. The Yankees called it the "whirly" before "sweeper" took hold around 2022. Adam Ottavino, Shohei Ohtani, Yu Darvish, and Nestor Cortes are among the most prominent practitioners.
A splitter, or split-finger fastball, is thrown with the fingers spread wide across the seams, reducing backspin to near zero. The pitch tunnels off the fastball's early trajectory before dropping sharply below the zone. Elite splitters drop 6 to 10 inches more than a fastball from the same release point.
Paul Skenes uses a splitter-sinker hybrid called the splinker as his primary put-away pitch, generating one of the highest whiff rates of any pitch in baseball.
Pitch design quality is one of three primary inputs into our prospect CUP model. Pitchers with elite IVB, repeatable arsenals, and strong tunneling profiles tend to be promoted more aggressively because their stuff translates immediately to MLB hitters.
Track current arsenal-driven CUP changes in the weekly calibration log.
A slider has a mix of side spin and topspin, producing both horizontal and vertical movement. A sweeper has almost pure side spin, producing almost purely horizontal sweep with minimal drop. Sliders typically sit 82 to 90 mph with 5 to 10 inches of horizontal movement. Sweepers sit 80 to 88 mph with 16 to 22 inches of horizontal movement.
Sliders are versatile against both-handed hitters. Sweepers are most effective against same-handed hitters who chase them out of the zone.
The velocity obsession in prospect evaluation is not irrational - velocity correlates with success because it compresses reaction time and makes individual pitches harder to identify. But velocity is a floor, not a ceiling. A 97 mph fastball that tunnels poorly, moves predictably, and is released from the same clear window as a curveball with dramatically different arm action can be solved by good major league hitters in under a full game's worth of at-bats.
The traditional scouting framework grades individual pitches on a 20-80 scale. A 60-grade fastball, a 55-grade slider, a 50-grade changeup. The grades describe each pitch in isolation. They do not capture whether the 60-grade fastball and the 55-grade slider actually share enough early trajectory to create a tunnel. A pitcher with a 70-grade fastball and a 45-grade curveball with poor tunnel relationships may project worse than a pitcher with a 60-grade fastball and a 55-grade curveball that shares the first 25 feet of flight perfectly.
The analytical frontier in pitcher evaluation is moving toward pitch-level trajectory data that measures tunnel quality directly. How similar are the first 15 feet of flight between the primary fastball and the primary breaking ball? How consistent is the release point within and across outings? What is the decision-point location for each pitch pair, and how much movement has occurred at that point relative to the total movement?
None of those questions are answered by a radar gun reading or a raw strikeout rate. But they are increasingly answerable from the Statcast and TrackMan data that player development departments have access to - and that the most analytically sophisticated organizations are using to make development decisions that look unintuitive from the outside.
The battle is won before the ball leaves the hand because by the time the ball is in flight, the pitcher's entire arsenal has already either set up or failed to set up the deception. Release point consistency has already determined whether multiple pitches can share the same tunnel window. IVB and VAA have already determined whether the fastball will arrive where the hitter's brain expects it to. Extension has already compressed or expanded the window the hitter has to distinguish between pitch types.
Evaluating pitching prospects through this lens changes the conversation materially. It shifts focus from what any single pitch does in isolation to what the arsenal does as a system. It demands asking whether the strikeout rate reflects genuine prediction problems or temporary exposure gaps. It requires distinguishing between pitchers who will scale because their arsenal creates structural deception and pitchers who are generating stats against underprepared competition.
That distinction is the core of accurate pitcher projection. And it is consistently underweighted in how the prospect industry talks about arms - because individual pitch grades are easier to communicate than pitch interaction quality, and because minor league strikeout numbers are immediately satisfying in a way that tunneling percentages are not.
The pitchers who survive the jump to the majors are usually not the ones who were the most impressive in Double-A. They are the ones whose arsenals kept creating the same prediction problem at every level - because the problem was structural, not situational. Track which arms have it in the weekly update.