WL4 MOTION INTELLIGENCE — PUBLIC EVIDENCE SUMMARY Collection review: 22 September 2026 DEVELOPMENT RECORDS 4,742 archived result records across 13 selected historical datasets: 4,282 arm-result records; 360 cascade/comparison records; 100 paired scenario records. Includes WL4, baseline and proxy results and failures. Copies, aggregate/batch duplicates and derived comparisons were not added again. This does not establish 4,742 unique executions, scenarios or successful tests. All 13 selected result files matched their existing SHA-256 manifest entries. Hash agreement is not independent route validation or proof of execution. One historical v0.8.4 study contains 2,400 arm results over 100 synthetic scenarios, three OMPL planners, four budgets and two arms. A separate historical comparison contains 400 results over 100 scenarios and four arms. These are included in the historical total, not additional totals. Historical internal validity claims were not independently revalidated during collection. The internal bend-radius validator has a documented limitation. PUBLIC REPEATABILITY — SEPARATE SET 30/30 completed runs; 120/120 cases reported internally valid. Four cases repeated, not 120 distinct scenarios. Report and raw-output hashes checked during collection. SEPARATE DISCRETE XYZ ASSESSMENT — ONE FOUR-SCENARIO TRIAL OMPL + WL4: 4/4 PASS. VAMP + WL4: 4/4 PASS. Raw OMPL: 0/4 PASS. Raw VAMP: 0/4 PASS. All eight raw routes failed the discrete bend-radius assessment. False-positive internal VAMP claims were identified for DIST_HARD_1600 / vamp_raw and DIST_EXTREME_2600 / vamp_raw. The original reassessment remained rejected with exit code 2 because of these discrepancies. The supplemental geometry classification preserves that rejection; it does not erase or approve the original full procedure. SCOPE The XYZ validator is a separate implementation, not third-party reproduction. The trial assesses discrete XYZ polylines, scenario obstacles/clearance, continuity and discrete bend radius. It does not prove a smooth physical robot trajectory, joint limits, kinematics, self-collision or hardware safety. 250 ms is the configured planner budget; 275 ms is the historical acceptance tolerance. Neither is a website response-time promise or speed-advantage claim. No general superiority or validity of the full historical benchmark is claimed. PUBLIC DEMO The scene is illustrative playback, not a live WL4 engine run. The browser route tool performs local descriptive geometry analysis only. The WL4 engine and proprietary source are not distributed by this website. PROTECTED TECHNICAL EVALUATION / LICENSING / ACQUISITION smeetsdevelopments@icloud.com