Affordance Perception: Reading Ball Trajectories Instantly

PRO CUE: The novice watches a ball and asks, "Where will it be?" The expert watches the same ball and receives the answer to a different question: "What does it invite?" A trajectory is not a set of coordinates to be computed — it is a menu of possibilities to be acted on. Attack this one. Neutralize that one. Survive the next. Perception that must calculate arrives after the ball has; perception that resonates arrives before it does.

Video: Learn to read the flight of the ball more quickly in tennis (Tennis Hacker).

1. Executive Summary & Athletic Intent

Affordance perception is the direct pickup of action possibilities from the structure of light arriving at the eyes — James J. Gibson's ecological-psychology insight that perception is not a computation over raw sensations but a resonance between the environment's structure and the organism's capabilities. An affordance is what the environment offers the animal: a chair affords sitting, a gap affords passing, and an incoming tennis ball affords — depending on its depth, height, pace, and spin, and equally on the athlete's reach and stroke repertoire — an attack, a neutral rally ball, or a survival action. The affordance belongs to neither ball nor athlete alone; it is a property of the athlete-ball-court system, perceived as a unit.

This is the mechanism behind tennis's most mis-described talent: "he reads trajectories instantly" — a phrase suggesting computation of unusual speed. The arithmetic argues otherwise. A rally ball crosses the court in 0.8–1.1 s; a first serve arrives in 0.45–0.55 s. The conscious cortical loop consumes 200–500 ms per cycle (Article 041) before a single muscle is commanded. No serial machine built on those latencies finishes in time. What finishes in time is a perceptual system tuned to relational invariants in the optic flow — variables that specify, without arithmetic, when the ball will arrive, how it will bounce, and what it can be made to do.

The central concept of this article is that the elite athlete does not perceive trajectories as geometry and then decide what to do; he perceives the decision itself, worn on the trajectory — and the bounce is the information event that finalizes the grade. Before the bounce, the flight pattern (launch angle, curvature, seam rotation) specifies a family of possible bounces; at the bounce, the first 100–150 ms of exit flight collapses the family into one; and the graded affordance — attack, neutral, defend — releases the pre-loaded motor program (Article 045) into execution. "Instantly" is not speed of thought. It is the abolition of the think-then-act gap.

The athletic intent is threefold. First, decision speed: the attack-neutral-defend grade is settled within 150–250 ms of the bounce, while the ball is still 0.3–0.5 s from contact — early enough that stance, loading depth, and contact height are chosen, not inherited. Second, selectivity: the grade selects the force recipe (Article 001), the stance, and the court position the shot should earn — the difference between building the point and merely returning the ball. Third, economy and resilience: the accurate affordance reader spends fewer emergency actions per match, preserves the action boundary under fatigue by re-grading honestly (the third-set ball now neutralized instead of attacked), and is deceived less — flight invariants are harder to fake than postures.

This article dissects the mechanism: the ecological claim that meaning is carried in the optic array; the tau mathematics of direct timing; the bounce as an information explosion; the attack-neutral-defend gradient; body-scaled action boundaries; the dorsal-stream substrate and affordance competition; the six-step direct-read cycle; the metrics and dosage; and the errors that betray a read still computing when it should be resonating.

2. Perceptual & Neuro-Cognitive Foundation

2.1 The Ecological Claim: Meaning Is in the Array, Not Added to It

Gibson's objection to the compute-then-perceive picture begins with the input. If the brain received only retinal images — flat, inverted, ambiguous snapshots — then perception would require inference, and inference would require time the athlete does not have. Gibson's answer: the input is not an image but a structured array — light reflected from surfaces arrives at a moving point of observation carrying information about layout, motion, and possibilities for action. Structure that remains invariant under the observer's own movement is information, and the perceptual system evolved to detect it directly, not to reconstruct the world from fragments.

The affordance is the unit of that detection. It is simultaneously factual and relational — factual because depth and spin are physical; relational because "attackable" has no meaning without this athlete's forehand. Two athletes can look at the same ball and perceive different objects: for one, a mid-court sitter begging to be driven; for the other, a shoulder-high ball exiting a skidding bounce, above his strike zone. Neither is wrong. The affordance lives in the relation, and the coaching question is never "what does the ball look like?" but "what does the ball look like to you, today?"

2.2 Tau: Time-to-Contact Without Arithmetic

As the ball approaches, its retinal image expands. Let $\theta(t)$ be the ball's angular size on the retina; the ratio of current angular size to its instantaneous rate of change specifies time-to-contact:

$$\tau(t) = \frac{\theta(t)}{\dot{\theta}(t)}$$

This is David Lee's tau. For an object closing at constant velocity, $\tau$ equals the true time remaining — no estimate of distance, absolute size, or speed required, because the answer is carried by a single optical ratio. Nervous systems across species — fielders, gulls, catching infants — are tuned to such ratios. Gravity and drag bend the trajectory, so raw $\tau$ drifts from true arrival time late in flight; the calibrated system absorbs the drift, and the athlete never needed absolute accuracy — only the coupling accuracy that lets the swing close on the ball.

Interception research adds the second half: striking actions appear regulated by coupling the tau of the effector to the tau of the ball — holding their difference on a controlled closure course rather than computing a meeting point in advance. The forward swing is not released at a memorized moment; it is regulated by optical expansion until the final approach, by which point the fovea has already left the ball for the contact window (Articles 055, 041) and the program runs on residual specification. The perceptual work is front-loaded: by the last 150 ms, "when?" has been answered and archived.

2.3 The Bounce: An Information Explosion Mid-Flight

Between the opponent's strings and the receiver's strike zone, most balls touch the court once — the richest information event of the flight. The bounce converts spin into trajectory: topspin bites into the surface and rebounds steeper and slower than the incoming angle suggests (the kick); underspin skids and stays low; an angled kick serve exits with lateral deviation no pre-bounce line predicts. Crucially, the conversion is lawful: given incident angle, speed, spin, and surface, the exit trajectory is determined — so the pre-bounce flight, read to the end, specifies the bounce before it happens, and the first 100–150 ms of post-bounce flight specifies everything after it.

The expert's advantage is not faster vision but earlier attunement to the pre-bounce invariants: seam rotation rate disclosing spin class, curvature disclosing the Magnus contribution, incident angle disclosing how much pace the surface confiscates. Occlusion studies show experts predicting bounce character from footage cut before the court touch at levels novices reach only with post-bounce frames. Each surface writes its own grammar — hard courts return pace faithfully, clay confiscates it and lifts the ball, grass skids and shortens the flight — and the read is surface-calibrated. An athlete raised on clay reads a low skidding grass bounce late — his library lacks the page.

2.4 The Affordance Gradient: Attack, Neutral, Defend

The output of the read is not a position but a grade, and each grade carries its own response pattern, court position, and risk budget:

Incoming Ball (post-bounce read)	Affordance Grade	Response Pattern	Court Position Earned
Short, low, slow — inside the service line	ATTACK	Step in, take on the rise, contact above net height, drive or close	Inside baseline; net closure on the next ball
Mid-court, comfortable height, neutral pace	ATTACK-BUILD	Heavier drive, shaped depth, target the open court	On or just inside the baseline
Deep, high, heavy topspin	NEUTRAL	Rally stroke, higher and safer, re-establish position	Behind the baseline, centered
Deep, fast, flat, or heavily angled	NEUTRAL-DEFEND	Compact stroke or redirected block, early recovery	Behind the baseline, shaded to cover
Beyond reach, stretched wide, or at shoe-tops	DEFEND	Slice float, block, lob — buy time at all costs	Recovery sprint

The grade is decided by ratios, not absolutes — depth relative to the court, height relative to the strike zone, pace relative to preparation time, spin relative to the bounce library — and relative to the athlete, the subject of the next subsection.

2.5 Body-Scaled Affordances and the Action Boundary

Warren's stair-climbing experiments established the template: perceived climbability tracks the ratio of riser height to leg length, with a critical boundary near $h/L \approx 0.88$ — people of different statures disagree about which absolute steps are climbable while agreeing about the ratio. Perception is scaled to the body, not to the meter.

The tennis equivalents are immediate. Contact height: the strike zone — roughly hip to shoulder, in front of the lead hip — is an affordance of anthropometrics and technique; a ball exiting the bounce at 1.4 m is comfortable to a 1.90 m athlete with a semi-western grip and high to a junior with an eastern grip. Reach and time: a wide ball is gettable only if the time needed to arrive covers inside the time available:

$$R_{\text{action}} = \frac{T_{\text{available}}}{T_{\text{needed}}}$$

With $R_{\text{action}}$ comfortably above 1, the ball affords attack; near 1, neutral; below 1, defense. The boundary is dynamic — fatigue lengthens $T_{\text{needed}}$, pressure distorts it in both directions (Article 072), and the honest athlete re-calibrates continuously. The most common tactical error in tennis is not technical: it is reading set-one affordances with set-three legs.

2.6 The Neural Substrate: Specification and Affordance Competition

The dorsal visual stream — the parietal "where and how" pathway — does not label objects; it specifies relations between environment and potential action: interception points, contact zones, movement parameters. This is the machinery of direct perception: optical structure becomes action-relevant metrics without passing through verbal description — precisely why it outruns the conscious loop.

Cisek's affordance competition hypothesis supplies the architecture. The brain does not first perceive, then decide, then act: fronto-parietal circuits hold multiple potential actions in parallel, specified from the same optical flow, and selection is a competition biased by evidence, priors, and reward history (basal ganglia) — the winning action already partly prepared by premotor cortex before it is "chosen." The serial pipeline is an illusion of introspection, and it explains why the expert's read feels instantaneous: specification and selection occupied the same 150 ms, and the body was already moving while Self 1 still believed it was deciding (Article 045).

2.7 Attunement and Calibration: How the Read Is Educated

Gibson called the learning process the education of attention: skill acquisition is not the addition of computation but the discovery of progressively finer invariants. The beginner sees "a ball coming"; the intermediate sees direction and depth; the expert sees the seam rotation and flight curvature that specify the bounce two events before it happens. Each stage is a finer carving of the same optical array — nothing new is computed; something old is finally noticed.

Calibration is the second half: mapping the perceived invariant onto one's own action capabilities, and re-mapping when capabilities change — after a growth spurt, an injury, or a fatigue threshold crossed. Attunement without calibration produces the athlete who reads perfectly and executes late; calibration without attunement produces fast feet with no idea what the ball is offering. The training corollary is strict: affordance perception is built by high-variability, high-feedback exposure — thousands of trajectories spanning the tour's space of spins, speeds, and surfaces, with the outcome of every read made explicit.

3. Step-by-Step Technical Execution

The direct-read cycle runs in six steps. Times are referenced to the opponent's contact ($t = 0$), the bounce ($t_B$), and the athlete's own contact. Each step is named for what the eyes and the perception-action system do.

Step 1 — Post-Impact Audit & Gaze Handoff (own previous contact +0.1 to +0.4 s)

Hold the contact window ~100 ms after your own impact — the error audit that recalibrates the tau mapping — then hand the gaze to the opponent's trunk while recovering. Late eyes at recovery quietly cause a late preview two shots later (Article 046).

Step 2 — Kinematic Preview (t ≈ −0.5 to 0 s)

Read the opponent's body, not the ball: hip coil, shoulder axis, and takeback depth disclose the family of trajectories being prepared — heavy topspin arc, flat drive, or slice. The preview pre-loads the distribution and pre-opens the two or three motor programs the family implies; the unweighting for the split step is timed from this read (Article 083).

Step 3 — Early-Flight Sampling (t = 0 to +0.10 s)

The first 100 ms of flight is the richest origin sample the sport offers: launch angle, initial pace, curvature (the Magnus signature), and seam rotation (the spin class). Ride the ball briefly with smooth pursuit; hold the court in the periphery (Articles 052, 055). Call the trajectory class silently and early — deep and heavy, flat and skidding, short and soft — and let the call bias the program competition.

Step 4 — The Bounce Audit (t_B)

The bounce is the decisive information event. Read three things in the first 100–150 ms of exit flight: where it landed (depth class), how it exited (vertical angle, pace retained), and what the spin became (kick, skid, or sit-up). The split step lands here and the first step fires — toward the zone the read is building, not the spot the ball occupies.

Step 5 — Affordance Commitment (t_B + 0.10 s to own contact − 0.15 s)

Grade the ball on the attack-neutral-defend gradient and check it against $R_{\text{action}}$ — honestly, at today's fatigue level. Select stance, contact height, and force recipe; commit fully. Inside the last 150 ms there is no re-litigation: the program runs, and the audit comes after impact.

Step 6 — Anchored Execution & Re-Audit (final 0.15 s to own contact +0.1 s)

The fovea locks the spatial window (Article 041); the pre-committed program executes against that anchor; the post-impact hold (~100 ms) compares predicted contact against actual contact. Every discrepancy is the recalibration signal for the next read — the loop that keeps attunement honest across a career.

4. Performance Metrics & Dosage

4.1 Performance Metrics

Metric	Measurement Method	Elite Benchmark	Recreational Benchmark
Trajectory prediction accuracy (occlusion cut at bounce + 0.10 s)	Video test battery, 40+ trials; call depth and height class	75–85%	45–55%
Time-to-contact estimation error	Temporal judgment task vs. high-speed video	≤ 20 ms	50+ ms
Attack-decision onset (footwork shift after mid-court feed)	Synced video / instrumented insoles	≤ 0.20 s post-bounce	> 0.40 s post-bounce
Strike-zone contact rate (hip-to-shoulder, in front)	Video tagging, 50 strokes	65–75%	35–45%
Short-ball recognition (drop-shot feeds)	First-step direction, video-timed	Correct within 0.30 s on ≥ 80% of feeds	Correct ≤ 50%, late
Bad-bounce clean-strike rate (irregular feeds)	Coach-generated irregular bounces, 20 feeds	65–75%	30–40%
Decision-action consistency (same ball class → same response)	Match video coding	> 80%	Erratic / mood-dependent
Rising-ball attack rate (mid-court balls taken before apex)	Video tagging, seasonal trend	Rising across seasons	Near zero (waits for apex)

4.2 Training Dosage Parameters

Trajectory library (the extraction layer): - Ball machine with graded variability: block 1 — one trajectory, call the class aloud before the bounce; block 2 — two trajectories alternating; block 3 — random, with the call. 3 × 10 balls per block, 2×/week. - Progression gate: ≥ 80% correct class calls at the current variability before raising the grade. The call must precede the bounce or it does not count.

Spin visibility (the invariant layer): - Two-tone or marked-seam balls: juggle, toss, and short-feed; call the spin class (top / flat / slice / kick) from seam rotation before the catch or bounce: 10 min, embedded in every machine session. - Coach-generated bounces: deliberate kick, skid, and sit-up feeds from short range — the surface grammar lesson compressed.

Surface libraries (the calibration layer): - Scheduled exposure blocks on clay and grass, or their closest local analogues; the library must span the surfaces the season will demand. - Irregular-bounce drills — lines, lumps, damp patches: 20 feeds weekly; the goal is not clean conditions but honest reads of unclean ones (Article 079).

Occlusion ladder (the prediction layer): - Video of incoming trajectories cut at the bounce; call landing depth, exit height, and the affordance grade; verify against the full clip: 3 × 8 min, 2–3×/week, film matched to level and surface.

Fatigue recalibration (the honesty layer): - Final 10 minutes of physical sessions: graded feeds at RPE ≥ 8; the athlete must re-grade — defend balls he would attack fresh — because the boundary shrinks and the map must follow (Articles 072, 077).

Constraint games (the transfer layer): - "Every ball landing inside the service line must be attacked" — the grade is scored, not suggested. - Wrong-grade penalty: an attacked defend-grade ball ending in error concedes the point outright.

Diversity quota: the weekly trajectory diet must include lefties, slice artists, heavy topspin, flat hitters, and at least one age group outside the athlete's own — affordance libraries overfit to a narrow diet, and the tour does not serve one.

5. Errors, Causes & Corrections

Perceptual Error	Underlying Cause	Performance Consequence	Corrective Strategy
Spectator tracking	Watches the ball as scenery; no invariant extraction, no call.	Decisions arrive with the ball rather than before it; perpetual time pressure.	Call-aloud protocol on every machine ball; the class call must precede the bounce to count.
Spin blindness	Seam rotation never educated; every bounce treated as a surprise.	Chronically late on kick serves and heavy topspin.	Two-tone ball spin-calling; short-range coach bounces (kick / skid / sit-up) until the call is automatic.
Monocular depth failure	Depth judged from image size alone; binocular and texture cues degraded under flat light or dusk.	Misgrades short balls deep and deep balls short.	Depth-audit video sets under match lighting; position-judgment drills with feedback on every call (Article 075).
Hero-ball overgrading	Action boundary read from set-one legs all match; ego inflates $R_{\text{action}}$.	Attacks defend-grade balls; errors at the worst tactical moments; third-set collapse.	Fatigue recalibration block at RPE ≥ 8; coach audits that grades are re-issued by set three.
Passive undergrading	Boundary never tested; every ball neutralized by default.	Mid-court sitters returned meekly; opponent granted time on every short ball.	Constraint game: every ball inside the service line must be attacked; rising-ball counts tracked weekly.
Static calibration	One map for all days; no re-mapping after fatigue, injury, or equipment change.	Reads correct but executions late; contact drifts backward across matches.	Pre-match boundary check (5 graded feeds); honest re-grade rules by set and by RPE.
Single-hypothesis lock	Commits to one bounce prediction pre-bounce; ignores disconfirming early exit flight.	Wrong-footed by unexpected skid or kick; no secondary program primed.	Two-hypothesis protocol: primary and secondary call before the bounce; confirm-or-switch inside the first 100 ms post-bounce.
Apex-waiting	Lets every ball reach apex and fall; discards the rising-ball affordance.	Contacts below the strike zone; surrenders position and time on every mid-court ball.	Take-it-early progressions: half-rise, three-quarter-rise, full rising contact; strike-zone contact rate audited weekly.

6. Diagnostic Diagram

PANEL A — THE DIRECT-READ TIMELINE (reference: opponent contact, t = 0)

 time ────────────────────────────────────────────────────────────────────→
 t=0         +0.10 s          t_B (bounce)     t_B+0.15 s         own contact
  |             |                 |                 |                  |
 [KINEMATIC ][EARLY-FLIGHT  ][BOUNCE     ][AFFORDANCE      ][ANCHORED     ]
 [PREVIEW   ][TAU SAMPLING   ][AUDIT      ][COMMITMENT      ][EXECUTION    ]
 body read    angle, pace,    depth,      grade + program   window lock,
 family of    curvature,      exit        select; no        program runs,
 trajectories seam rotation   angle       re-litigation     +100 ms audit

 information:  family (vague) ──→ ┌─ spike: resolves ─┐ ──→ one ball ──→ locked
                 early bias        │  to one bounce   │     specified     execution
                                   └──────────────────┘

PANEL B — THE AFFORDANCE GRADIENT (post-bounce ball, side view)

 court depth:  net ──────────────── service line ─────────── baseline ──→
               |  ATTACK   |  ATTACK-BUILD | NEUTRAL | NEUT-DEFEND | DEFEND |
 ball height:  high ──┐         strike zone (hip–shoulder) governs the
                        │         horizontal read; deep + high + heavy →
                 low ──┘         neutral-defend; beyond reach → full defense

 grade = f(depth, height, pace, spin) ÷ the athlete's CURRENT capability

PANEL C — THE ACTION BOUNDARY UNDER FATIGUE (same incoming ball)

  R_action = T_available / T_needed

  fresh (set 1):   T_needed 0.55 s  |████████████|       R ≈ 1.50  → ATTACK
  tired (set 2):   T_needed 0.70 s  |████████████████|   R ≈ 1.15  → NEUTRAL
  spent (set 3):   T_needed 0.85 s  |███████████████████ R ≈ 0.95  → DEFEND

  the ball did not change; the athlete did — the honest read re-grades

Reading the diagram: Panel A shows where information lives across the flight: a vague family before the bounce, a spike of resolution at the bounce, and a single specified ball after it — the read must be biased early and confirmed late, never begun late. Panel B maps the grade onto court geometry: the same ball is attack, neutral, or defense depending on where it lands, how it exits, and who is asking. Panel C is the honesty test of the whole system: the action boundary is a ratio, and the ratio belongs to today's body, not last month's.

7. Video Demonstration

[SwLx2q6qc3I]

Placeholder: insert a two-part film. Part one: incoming rally trajectories filmed from the receiver's perspective at 240 fps, frozen at four frames — opponent contact, +0.10 s, the bounce, and bounce + 0.15 s — with the affordance class call overlaid at each freeze and the full flight revealed only after each call. Part two: split-screen of an elite and a recreational athlete receiving the same three mid-court balls, with decision-onset markers (first attacking footwork) and strike-zone contact tags — the elite marker fires within 0.20 s of the bounce, the recreational marker after the ball has passed its apex. Pair with the Article 041 Quiet Eye footage for the anchor and the Article 083 split-step sequence for the timing chain.

8. Cross-Domain Synergy

Article 001 (Ground Reaction Forces — Vertical vs. Horizontal Vectors): downstream of the grade — the affordance read selects the force recipe; attack, neutral, and defend are, mechanically, different vector mixes.
Article 041 (Quiet Eye (QE) — Fixing Visual Anchors Before Impact): the destination of every read — the contact window the fovea locks once the affordance is committed.
Article 045 (Subconscious Motor Program Recruitment — Self 1 vs. Self 2): direct perception is Self 2's native language; affordances are what the subcortical engine is built to see, and verbal computation is what it must never be asked to do mid-flight.
Article 046 (Spatial Awareness and Anticipatory Visual Cues): the upstream preview — the kinematic read of the opponent that pre-loads the trajectory family before the ball exists.
Article 055 (Saccadic Eye Movement vs. Smooth Pursuit in Ball Tracking): the oculomotor mechanics of early-flight sampling — where pursuit ends, why the ball is abandoned before contact, and what the eyes do instead.
Article 063 (Visual Anchor Shifts from Ball Seams to Opponent Hip Cues): the seam as the spin invariant — this article reads seams in flight; Article 063 shifts the anchor to the opponent before flight begins.
Article 068 (Sensory Deprivation and Visual Goggle Training Methods): the stress test of attunement — degraded arrays force the system onto coarser invariants, exposing which variables the athlete actually uses.
Article 072 (Stress Resilience and Cortisol Buffering during Tiebreakers): pressure distorts the action boundary in both directions — the hero ball and the passive push are the same error wearing opposite signs.
Article 077 (Executive Function Maintenance into 5th Set Conditions): the late-match recalibration problem — the boundary shrinks and the map must shrink with it, honestly.
Article 079 (Subconscious Micro-Adjustments in Unpredictable Bounces): the correction layer downstream — when the read is wrong or the bounce lies, micro-adjustment salvages the stroke.
Article 083 (The Split Step: Neural Activation and Timing): the timing machinery the read triggers — unweighting timed from the preview, landing synchronized with the bounce.

9. Self-Assessment Matrix

#	Assessment Question	Yes	No
1	On mid-court balls, do you know your grade (attack / neutral / defend) before or as the ball bounces, rather than after?	☐	☐
2	Can you call the bounce character of a heavy topspin ball (kick, skid, sit-up) before it lands, at better than 7 of 10?	☐	☐
3	Do you take more than half your mid-court balls on the rise, before the apex?	☐	☐
4	Against a lefty's slice or a heavy kick serve, do your first two steps still go to the right place?	☐	☐
5	In the third set, do you find yourself playing honestly deeper, safer positions on balls you attacked in set one?	☐	☐
6	When fatigued, does your strike-zone contact rate stay within ~10% of your fresh rate?	☐	☐
7	On bad bounces (lines, lumps, damp patches), do you still put the racket cleanly on more than two of three?	☐	☐
8	On match video, does the same class of ball produce the same response — same stance, same contact height, same intent — throughout the match?	☐	☐

Scoring: 7–8 "Yes" — the read is functioning as direct perception; proceed to surface-library expansion and deception dosing. 4–6 — mixed profile; prioritize the failed prerequisites (items 1, 2, 5 — early grading, spin visibility, honest recalibration) with 3–4 weeks of call-aloud and occlusion work. 0–3 — the game is being played on late, computed information; run the full corrective block in Section 5 and re-test with the occlusion battery in 4 weeks. The eyes are rarely the problem; the education of them is.

10. Printable Practice Card: The Direct-Read Protocol

Session goal: grade the ball before the bounce confirms it; execute the grade without re-litigation; audit every read.

Block	Drill	Dose	Cue
A — Prime (5 min)	Two-tone ball spin calls: toss, juggle, short feeds; call spin class from seam before the catch	3 × 20 tosses	"Read the seam, not the sphere."
A — Prime	Peripheral court-hold: fovea on feeder's ball, periphery calls hand signals at court edges	3 × 1 min	"Ball in the center, court in the margins."
B — Extract (10 min)	Machine, graded variability: one trajectory → two → random; class call must precede the bounce	3 × 10 balls	"Call it before it lands."
B — Extract	Occlusion video: cut at the bounce; call depth, exit height, grade; verify against the full clip	3 × 8 min (2–3×/week)	"The bounce is the answer key."
C — Transfer (15 min)	Constraint game: every ball inside the service line must be attacked; wrong grade loses the point	1 set to 4	"The grade is the shot."
C — Transfer	Disguised feeds: coach fakes heavy and feeds flat (and vice versa); athlete holds two hypotheses until bounce + 0.10 s	20 feeds	"Primary and secondary — confirm, don't assume."
D — Audit (5 min)	Video 10 balls: mark decision onset vs. bounce; tag strike-zone contact; check response consistency	10 clips	Check: onset ≤ 0.20 s post-bounce, contact in the zone, same ball → same response.

Weekly frequency: 2 machine/occlusion sessions (Blocks A–B) + Blocks A, C, D embedded in every court session. Progression gate: do not raise machine variability until class calls are ≥ 80% correct before the bounce at the current grade; do not add disguise until honest-feed grading is ≥ 8 of 10. Durability rule: one grading block per week at RPE ≥ 8 and one under consequence (wrong grade = sprint) — a read that only works fresh and calm is a souvenir, not a skill.

Cross-references: ART-001, ART-041, ART-045, ART-046, ART-055, ART-063, ART-068, ART-072, ART-077, ART-079, ART-083. Pillar II: Neuro-Athletics & Cognition — TennisKB 200-Article Knowledge Base.