Governed hypotheses and results
Three distinct evidence programs, with their original classifications and limitations. FSI = Fractal Stability Index; CSS = Cohort Similarity Score.
Completed 55-comparison mobility experiment
Loading reviewed value · Loading reviewed value · Loading reviewed value vertical RMS
Mall No-Steps skating versus pain-limited Walking: all 55 matched dates agree for each of these three ACC-clock-independent comparisons. Mall medians: 2.003 miles, 34.1 minutes, 0.0520 g. Walking: 0.141 miles, 5.57 minutes, 0.1382 g.
Exact frozen values, confidence intervals, inferential methods and provenance
Kubios confirms Custom Measurement is designed to request 50 Hz ACC from the Polar H10, while Readiness requests 200 Hz. Kubios reports that the most probable explanation for approximately 200 XYZ records per second in affected Custom KDF files is a software/hardware configuration mismatch in which the H10 may have remained at 200 Hz. Kubios has requested representative Custom and Readiness native KDF files and is investigating; the cause is not yet confirmed. Absolute jerk (g/s), rates, filtered or cropped tails, spectra, timing-dependent dose and dependent FSI/CSS/Tensor outputs remain under forensic review. Historical values and classifications are preserved; they are not validated final physical-time findings. The frozen triplet distance, event-clock duration and whole-record vertical RMS findings do not depend on the unresolved ACC clock.
How quickly do mobility states become distinguishable?
SURF-GEN-01: PARTIALLY_SUPPORTED · validation PASS
With FULL features, Gaussian predictive discrimination converged at 5 triplets; complete geometric maturity required 35: a 30-triplet gap. Removing distance and duration moved predictive convergence to 40 triplets, while geometric convergence remained at 35. Early FULL discrimination therefore depends strongly on exposure information.
The paired experiment asks how measured functional/mechanical outcomes differ. Surfaceology asks when states become predictively and geometrically distinguishable, under which feature assumptions. Neither program replaces the other.
| Profile | M0 linear | M1 MLP | M2 Gaussian | Geometry | Geometry−M2 |
|---|---|---|---|---|---|
| FULL | 44 | 35 | 5 | 35 | 30 |
| STRICT_EXPOSURE_BLIND | 44 | 44 | 40 | 35 | -5 |
Mall: PROVISIONAL; Walk: PROVISIONAL; PT: QUALIFIED. PVC shift: PERSISTENT.
55 accepted triplets; 54 complete model trajectories: 44 development, 5 PVC-01 and 5 continuation. One development triplet lacks PT RMS and is excluded whole, without imputation.
Exact governed hypothesis registry
Statuses are copied from SURF-GEN-01 H1–H6. Triplet endpoint findings and transportation publication gates remain separate; no status is inferred from effect sizes. N: 44 development + 10 forward triplets. All rows: validation PASS; result version 20260920T-LOCAL-SURF-GEN-01-REPRO-V2.
| ID | Hypothesis | Status | Primary evidence |
|---|---|---|---|
| SURF-GEN-01/H1_surface_center | Surface centers converge with repeated observations. | SUPPORTED | {"status": "SUPPORTED", "supporting_n": 30} |
| SURF-GEN-01/H2_surface_covariance | Surface shape and covariance converge. | PARTIALLY_SUPPORTED | {"status": "PARTIALLY_SUPPORTED", "supporting_n": 35} |
| SURF-GEN-01/H3_forward_distance | Distances for later observations converge. | PARTIALLY_SUPPORTED | {"status": "PARTIALLY_SUPPORTED", "supporting_n": 35} |
| SURF-GEN-01/H4_early_prediction | Predictive discrimination converges early. | SUPPORTED | {"status": "SUPPORTED", "supporting_n": 5} |
| SURF-GEN-01/H5_exposure_independent | Convergence persists without distance and duration. | PARTIALLY_SUPPORTED | {"status": "PARTIALLY_SUPPORTED", "supporting_n": 35} |
| SURF-GEN-01/H6_gaussian_before_flexible | Gaussian geometry reaches useful behavior before flexible learned models. | SUPPORTED | {"m1_n": 35, "m2_n": 5, "status": "SUPPORTED"} |
Transportation exposure — historical discovery pass
New separately validated result: TRANSPORT-RMS-01: H-TR-A through H-TR-D, clock-independent whole-record RMS. The discovery status and readiness matrix below describe the prior pass, not the new frozen RMS analysis.
Discovery only. Final timing-dependent inference: BLOCKED_PENDING_MEASUREMENT_AUTHORITY.
The operator reports greater pelvic/SI-region pain and discomfort in Bus, Van, constrained Sedan/Taxi seating and SilverRide Back than SilverRide Front or Motorcycle. This is reported context, not an independently validated causal ranking.
Body coupling, constrained seating, axis structure, repeated transients, duration and load-path organization may matter beyond acceleration magnitude. Mechanism is not established.
Kubios confirms Custom Measurement is designed to request 50 Hz ACC from the Polar H10, while Readiness requests 200 Hz. Kubios reports that the most probable explanation for approximately 200 XYZ records per second in affected Custom KDF files is a software/hardware configuration mismatch in which the H10 may have remained at 200 Hz. Kubios has requested representative Custom and Readiness native KDF files and is investigating; the cause is not yet confirmed. Absolute jerk (g/s), rates, filtered or cropped tails, spectra, timing-dependent dose and dependent FSI/CSS/Tensor outputs remain under forensic review. Historical values and classifications are preserved; they are not validated final physical-time findings. The frozen triplet distance, event-clock duration and whole-record vertical RMS findings do not depend on the unresolved ACC clock.
Whole-record RMS is clock-independent, but no final Bus/Van/Sedan versus Motorcycle/SilverRide Front ranking is authorized by this publication pass. Materialized episodic coverage: {"BUS": 22, "MOTORCYCLE": 8, "SEDAN_TAXI_UNKNOWN": 26, "SILVERRIDE_BACK": 2, "SILVERRIDE_FRONT": 7, "SILVERRIDE_UNKNOWN": 1, "VAN_CUTAWAY": 6}. These are sessions, not independent dates; sedan seating remains unknown.
Publication-readiness matrix
| Endpoint | Clock dependency | Validation | Final publication ready | Reason |
|---|---|---|---|---|
| vertical_dynamic_g_rms | TIMEBASE_INDEPENDENT | DISCOVERY_ONLY | No | Canonical whole-record amplitude; discovery cohort coverage only. |
| horizontal_dynamic_g_rms | TIMEBASE_INDEPENDENT | DISCOVERY_ONLY | No | Canonical whole-record amplitude; sensor axes are not anatomically calibrated. |
| magnitude_dynamic_g_rms | TIMEBASE_INDEPENDENT | DISCOVERY_ONLY | No | Canonical centered raw magnitude; not filtered Method-A resultant. |
| P95 | TIMEBASE_INDEPENDENT | DISCOVERY_ONLY | No | Unfiltered whole-record operator is clock-independent, but no validated canonical transport comparison is frozen. |
| P99 | TIMEBASE_INDEPENDENT | DISCOVERY_ONLY | No | Unfiltered whole-record operator only; stored NES tails are filtered/cropped. |
| P999 | TIMEBASE_INDEPENDENT | DISCOVERY_ONLY | No | Unfiltered whole-record operator only; not the stored NES feature. |
| maximum | TIMEBASE_INDEPENDENT | DISCOVERY_ONLY | No | Unfiltered whole-record operator only; not the stored NES maximum. |
| magnitude_jerk_rms_g_per_s | TIMEBASE_DEPENDENT_AT_RISK | BLOCKED_PENDING_MEASUREMENT_AUTHORITY | No | Absolute derivative requires resolved ACC timing. |
| nes_p99 | TIMEBASE_DEPENDENT_AT_RISK | BLOCKED_PENDING_MEASUREMENT_AUTHORITY | No | Stored filtered/cropped tail depends on clock and sample membership. |
| shock_spike_rate_per_min | TIMEBASE_DEPENDENT_AT_RISK | BLOCKED_PENDING_MEASUREMENT_AUTHORITY | No | Rate denominator depends on ACC clock. |
| nes_shock_cluster_rate_per_hour | TIMEBASE_DEPENDENT_AT_RISK | BLOCKED_PENDING_MEASUREMENT_AUTHORITY | No | Cluster membership and hours depend on timing. |
| nes_spectral_centroid_hz | TIMEBASE_DEPENDENT_AT_RISK | BLOCKED_PENDING_MEASUREMENT_AUTHORITY | No | Frequency axis/filtering depends on timing. |
| cumulative_dynamic_shock | TIMEBASE_DEPENDENT_AT_RISK | BLOCKED_PENDING_MEASUREMENT_AUTHORITY | No | Time integral depends on timing. |
| fsi_tensor_v0_3_score | TIMEBASE_DEPENDENT_AT_RISK | BLOCKED_PENDING_MEASUREMENT_AUTHORITY | No | Current Tensor inputs inherit impact-rate and duration dependencies; 65 legacy rows lack exact current session identity, so no current-cohort coverage is asserted. |
| css_cohort_review | TIMEBASE_DEPENDENT_AT_RISK | BLOCKED_PENDING_MEASUREMENT_AUTHORITY | No | Current canonical inputs include timing-dependent outputs. |
| fsi_score | TIMEBASE_DEPENDENT_AT_RISK | BLOCKED_PENDING_MEASUREMENT_AUTHORITY | No | Current compatibility scalar inherits Tensor timing dependencies. |
Propose a new frozen descriptive session/date-aware comparison of existing whole-record vertical, horizontal and magnitude RMS, separately by governed vehicle/seat cohort, with exact source hashes, missingness and no pooled pain ranking. No filtered tails, new formulas, Phase 2, or entropy execution. Independent-day sensitivity and seat-unknown exclusions must be preregistered.
Continuing research
FNS/SNS and pre/post ride protocols remain separate continuing experiments. Entropy/complexity is a research program; no new entropy or FSI/CSS claim is published here.
Limitations and provenance
- Repeated N-of-1 observations do not establish population-wide or causal effects.
- Surfaceology tests state distinguishability and maturation; it does not independently validate every paired biomechanical hypothesis.
- A frozen-reference maturity threshold is not a universal minimum sample size.
- Canonical FSI/CSS/Tensor retain their existing definitions. No new scalar, clinical outcome, entropy result or replacement model is introduced.
- Transportation seat labels are context, not measured pain. Motorcycle may show high amplitude despite greater reported tolerance; this motivates a hypothesis, not a proven mechanism.
Source versions and hashes
Canonical public SURF-GEN-01: 20260920T-LOCAL-SURF-GEN-01-REPRO-V2. REPRO-V1 remains noncanonical historical provenance.
{
"clock_classification_sha256": "eb7c291a3258ca58b47f63278d281b09bd9862c485fa6e24d0d0f81e0101b5ea",
"publication_scope": "aggregate summaries only; original scientific artifacts remain immutable",
"source_hashes": {
"HYPOTHESIS_EVALUATION.json": "b96546ad57ae50d41e6624b89c0b008d46da44d6267513e92d4db5d6656ea4e9",
"MANIFEST.json": "be64eddc3da05cacf1936b273ff30da628c225765a4fc660826c7cd0017056b5",
"REPORT.md": "c15581af33468604641335abc15b3703f9cef4c8d477e4e6838aee0405162042",
"RESULTS.json": "0bedb7757a82e7b4d3e1e01e37f4628ae01ecb347ba0d3cf8c54ccbd4a312e8b",
"VALIDATION.json": "5fe2bcec6cff9685ec776b1c9684c5c199142f218a7dc159ad3e0fba68bb41f8"
},
"surf_gen_run": "20260920T-LOCAL-SURF-GEN-01-REPRO-V2",
"transport_coverage_sha256": "6fcbe47b62628d3c7b24a5d05dc140a2766b22893a13c70c65663081398b56e5",
"triplet_results_file_sha256": "1df720024412974af6e93f8384b43e7d2dbb220a004e2b6d5e768765e28d6671"
}Machine-readable hypotheses, exact results and readiness matrixTransportation: localized exposure evidence
Explore SURF-TRANS transportation states and the V8 absolute localized result. These are separate from the ambulation hypotheses and do not change their classifications.