Independent weighted educational practice

Educational AI-assisted review is not clinician certification, an official examination result, a diagnosis, or patient-care advice. Clinical approval remains false. The source recording is unchanged.

Five objectives, ten possible points

Recording 3 uses the retained editorial weights 1, 2, 2, 3 and 2 for artifact, premature atrial complexes, premature ventricular complexes, left ventricular hypertrophy and one repolarization objective. Either the descriptive ischemic ST/T choice (64) or hypertrophy-associated ST/T choice (67) earns the same repolarization weight of 2. Both selected still earn 2.

Choices 6, 54 and 68 are explicitly neutral: zero credit and zero penalty, with their diagnostic uncertainties displayed. Every selected incorrect option deducts 1. The total is the larger of zero and objective credit minus incorrect-selection deductions. Warnings never cancel credit or add another penalty.

Separate scoring contracts

The new scale is independent-weighted-practice-v1, key version 2-code64-consistency. It is not the legacy scoring method and is not the independent equal-objective scale. Session exports retain the scale and key version; there is no cross-scale aggregate. Superseded recording-3 keys are rejected by the current application rather than silently reused.

What reviewed means

A draft, incomplete or unresolved v2 key cannot produce a numerical score. independent-ai-reviewed means the independent AI-assisted educational content and implementation checks have been completed for that version. It does not mean a clinician approved the key or that the weights are official. The clinical-approval field remains false. A wholly old cached application cannot be remotely revoked by this contract; reload to use the current release.

Measurement transparency

The default tracing uses sequential 2.5-second lead panels plus a 10-second II strip. The notes identify visible lead and time windows; endpoints from different displayed beats must not be combined. Local baselines and endpoint ranges are retained rather than presenting false precision. The V6 baseline windows are half-open: include samples at the starting time and exclude the ending time, so the QRS onset at 9.140 s is not included in the pre-QRS baseline.

• Recording 3: V6 T inversion near 9.410 s is 0.293–0.332 mV against its local pre-QRS baseline. This supports a descriptive repolarization finding, not coronary causation. The beat follows a PVC, an explicit causal caveat.
• Recording 5: the actual same-visible-beat early V1 atrial rise at 6.836–6.840 s has conservative P-onset bounds of 6.834–6.840 s. Paired with V1/II ventricular onset at 7.042–7.050 s, those bounds support a PR range of 202–216 ms. Its 89 classifications and maximum of 8 are retained.
• Recording 5: visible QRS intervals are bounded at 90–110 and 98–110 ms; no unsupported precise multibeat mean is claimed. Visible V4 QT is 410–426 ms and Bazett QTc approximately 468–486 ms.

Download original measurement and 89-row recording-3 notes · Download the reproducibility module

To recalculate from the public waveform bytes in your browser console:

(await import('/reports/reproduce-weighted-measurements.mjs'))
  .reproduceWeightedMeasurements()

The module verifies the waveform payload hashes and recalculates sample/baseline values and interval arithmetic. Endpoint attribution remains a manual interpretive decision; this is not an automated diagnostic validator. The educational evidence references are symbolic identifiers with release-byte integrity pins, not an independent runtime clinical-approval gate.