# Experiment report
- Question and measurable criterion:
- Prediction and physical/model assumptions:
- Robot, firmware, source revision and configuration:
- Evidence type: physical / simulation / replay / calculation:
- Method: changed factor, response, fixed conditions, start/reset and stop rules:
- Pilot versus development versus held-out validation:
- Raw records: relative CSV links; every attempted trial and outcome:
- Plots: units, conditions, captions and dataset source:
- Results: successes/attempts, signed/absolute errors and conditional metrics:
- Observations:
- Interpretation and alternative explanations:
- Uncertainty, confounding and operating envelope:
- Decision: retain / revert / inconclusive; supporting evidence:
- Unresolved faults and physical-validation status:
- Next experiment with a measurable decision criterion:
- Reproduction instructions, sources and AI assistance:
