Workflow overview
Kineologist|PITCH
Baseball pitching analysis from motion files to phase-locked evidence.
PITCH turns pitching motion capture files into P1–P4 phase markers, an average baseline, baseline-relative telemetry, 3D review, reports, plots and threshold-controlled rule outputs. It is designed for coaches and analysts who need to inspect how a throw differs from a reference movement.
PITCH workflow
PITCH is easy to run.
Load baseline and subject files, choose the pipeline options, run the pipeline, then review the outputs. PITCH keeps the user workflow simple while the backend handles P1–P4 phase detection, baseline creation, phase-aware crop windows, telemetry, reports, plots and baseball-specific review artifacts.
Load baseline files
Add one or more reference pitching files so PITCH can build an average reference movement.
Load subject file
Add the athlete file that will be reviewed against the generated pitching baseline.
Choose options
Select handedness, skeleton/remap options, crop policy, report scope and analysis profile, then run the pipeline.
Detect phases
PITCH identifies P1–P4 phase windows and sport-specific marker evidence for the pitching sequence.
Create baseline
Reference clips are phase-cropped, normalised, averaged and prepared as comparison authority.
Pipeline steps
The nine-stage automated PITCH pipeline.
PITCH separates operator work from backend evidence generation. The reviewer loads files and chooses the workflow; the software performs the phase detection, baseline creation, rule evaluation and artifact export.
01
Load baseline files
Add one or more reference pitching files so PITCH can build an average reference movement.
02
Load subject file
Add the athlete file that will be reviewed against the generated pitching baseline.
03
Choose options
Select handedness, skeleton/remap options, crop policy, report scope and analysis profile, then run the pipeline.
04
Detect phases
PITCH identifies P1–P4 phase windows and sport-specific marker evidence for the pitching sequence.
05
Create baseline
Reference clips are phase-cropped, normalised, averaged and prepared as comparison authority.
06
Analyse subject
The same phase-aware logic is applied to the subject file so comparison is fair and repeatable.
07
Compare evidence
Baseline and subject are aligned for timing, magnitude, local rotation and telemetry deltas.
08
Evaluate rules
Cocking and valgus rule families detect persistent patterns and grade severity where appropriate.
09
Generate outputs
Coach, Master and Muscle reports, plots, thresholds, CSVs and review surfaces are written as artifacts.
Average pitching baseline
Create a phase-aware pitching baseline before reviewing the athlete.
The baseline can be created from one or multiple reference files. P markers define comparable windows, each reference clip is cropped and normalised, the average movement is generated, and the baseline is then aligned to the subject file for phase-locked comparison of timing and rotations.
Average reference
Multiple reference files can be cropped, normalised and averaged into one comparison authority.
Shared P windows
Baseline and subject use the same P1–P4 grammar so timing and shape comparisons stay fair.
Original and stretched
Review native timing differences and phase-locked magnitude differences separately.
Artifact authority
Baseline files, markers and comparison outputs are written as inspectable artifacts.
3D and curve review
Review phase, timing and magnitude against the movement timeline.
TrackView lets the reviewer inspect selected telemetry channels against the motion timeline. The coaching layer shows baseline-relative timing and magnitude differences so the reviewer can ask where the player diverges and by how much.
P1–P4 phase markers
Meaningful pitching moments are marked for consistent review.
Phase-locked baseline
Subject and baseline are compared in aligned movement time.
Telemetry curves
Channel curves expose timing, speed, rotation and magnitude patterns.
3D context
Visual motion remains connected to the same generated evidence.
Cocking analysis
Cocking analysis compares technique and performance against the baseline.
The cocking analysis layer focuses on performance and technique comparison. It reviews persistent phase-relative patterns such as hand path, elbow/shoulder relationships, trunk separation, arm timing, readiness and baseline deltas. Detected patterns are graded by severity so coaches can see whether a finding is small, moderate or more pronounced relative to the selected review context.
Phase-relative review
Findings are interpreted inside the pitching sequence rather than as isolated frames.
Technique comparison
The athlete is compared against the selected reference or average baseline.
Persistent patterns
Repeated timing and magnitude relationships support severity grading.
Coach language
Outputs translate the rule findings into reviewable coaching language.
Valgus analysis
Valgus analysis highlights heightened-risk pitching patterns.
The valgus analysis is a rule-based review layer for identifying heightened-risk movement patterns in the pitching sequence. It is designed to support coaches with player injury-prevention strategies by identifying proxy indicators such as arm timing, forearm lag, elbow velocity, trunk timing and inverted-W style patterns. It is not a medical diagnosis or direct torque measurement.
Risk-pattern review
The system flags motion patterns that may deserve coaching attention.
Proxy indicators
Forearm lag, elbow velocity and trunk timing help contextualise findings.
Severity language
Persistent patterns are graded into readable severity bands.
Evidence retained
Reports remain tied to generated telemetry and artifact outputs.
Motion-derived proxy layer
Muscle proxy sequencing shows contribution patterns in context.
The system infers muscle proxy activity from relative joint rotations and velocities, then creates proxy groups for major muscle regions. This allows sequencing and contribution to be visualised as percentage-style proxy activity with direct 3D visual feedback and a dedicated Muscle Report table. It is not EMG or direct measured activation.
Joint rotation and velocity inference
Proxy activity is inferred from relative joint rotations and movement velocities.
Full-body muscle mapping
Motion-derived proxy channels map across major body regions.
Flexion and extension groups
Flexion and extension families make contribution patterns easier to inspect.
Muscle activity HUD
Proxy percentages can be reviewed beside the 3D motion display.
Sequencing and contribution reports
Dedicated report surfaces summarise ordering, contribution and comparison patterns.
Coach report
The PITCH Coach Report turns generated findings into a readable review.
The Coach Report is the fast applied review surface. It summarises baseball-specific findings, phase timing, baseline-relative differences and rule-triggered observations in language that can support coaching review without requiring the coach to open raw telemetry, debug files or rule snapshots.
Applied language
Plain-language findings can be discussed with coaches and athletes.
Phase context
Observations remain tied to P1–P4 movement windows.
Baseline comparison
The report explains where the athlete differs from the reference.
Rule-triggered findings
Persistent detected patterns can be converted into review language.
Muscle report
The PITCH Muscle Report tables motion-derived proxy sequencing.
The Muscle Report is the structured proxy-contribution surface. It shows sequencing and contribution in table form so reviewers can inspect how major motion-derived proxy groups appear through the pitching sequence. It is not EMG and does not claim direct measured muscle activation.
Proxy contribution
Contribution is inferred from movement relationships, not sensors on the muscle.
Sequencing table
Tables help review relative order and contribution.
Visual link
Proxy views can be reviewed alongside 3D motion.
Stable artifact
The report remains part of the generated output package.
Master report
The PITCH Master Report preserves deeper technical context.
The Master Report is the deeper review surface for analysts, educators and validation partners. It keeps more of the technical interpretation, generated evidence and rule context visible so a run can be inspected beyond a short coaching summary.
Detailed evidence
The report keeps generated context available for deeper review.
Long-form interpretation
Findings are explained with more technical detail than the Coach Report.
Telemetry context
Phase markers, curves and deltas remain connected to the run.
Validation support
Useful for facilities and education partners testing the workflow.
Plots
MATLAB-style plots support technical inspection.
Generated plots preserve channel evidence outside the GUI. They support biomechanics-style inspection, documentation, audit and deeper review of selected telemetry channels.
Channel evidence
Selected telemetry channels can be inspected outside the main interface.
Marker timing
Phase lines help reviewers connect plots back to P1–P4 movement moments.
Audit trail
Generated plot files support documentation, sharing and deeper review.
Curve shape
Reviewers can inspect timing, peaks, ramps and magnitude changes directly.
Thresholds
Threshold controls support expert review rather than one universal judgement.
Thresholds allow reviewers to tune overlays, markers, rule sensitivity and severity bands for context, demographic group, coaching opinion or validation work.
Rule sensitivity
Thresholds help tune how persistent patterns are classified for review.
Severity bands
Findings can be shaped into clearer caution or coaching levels.
Context control
Facilities and beta partners can adapt settings to their review protocol.
Transparent settings
The control layer remains visible instead of hiding assumptions inside the output.
Viewport review
Flexible viewport controls keep 3D inspection readable.
Perspective, front and side views, camera-local pan, orbit/centered modes and compare-offset controls help reviewers inspect subject and baseline motion without fighting the camera.
Orbit / centered
Switch camera behaviour to match the review task.
Perspective / front / side
Use standard viewpoints for motion inspection.
Camera-local pan
Move around the subject without losing context.
Compare offset
Separate subject and baseline for side-by-side inspection.
Muscle & labelling options
Reviewers can switch muscle overlays and labels to inspect the motion without losing spatial context.
Next step
Review pitching motion through deterministic evidence.
Book a walkthrough of Kineologist|PITCH, test the workflow with motion data, or discuss a beta validation partnership.