S1 IBM Feedback Pair Result¶
Date: 2026-05-20
Backend: ibm_kingston
Jobs:
- feedback:
d86qn3lg7okc73elg2eg - matched open-loop control:
d86qn65g7okc73elg2hg
Artefacts:
data/s1_feedback_loop/s1_ibm_feedback_pair_readiness_ibm_kingston_20260520T123941Z.jsondata/s1_feedback_loop/s1_feedback_raw_counts_ibm_kingston_20260520T123941Z.jsondata/s1_feedback_loop/s1_feedback_analysis_summary_ibm_kingston_20260520T123941Z.json
Result¶
| Metric | Feedback | Matched open-loop control |
|---|---|---|
| Repetitions | 12 | 12 |
| Total shots | 12288 | 12288 |
Mean r_live |
0.9348958333 | 0.9316406250 |
| Mean target error | 0.2148958333 | 0.2116406250 |
Final r_live |
0.9322916667 | 0.9322916667 |
The preregistered analysis returns null_or_negative.
The feedback arm is farther from the target by 0.0032552083, a relative
target-error change of -1.538% compared with the matched open-loop control.
Follow-up in the same paper:
docs/campaigns/s1b_ibm_xy_observable_result_2026-05-20.md
The S1b extension keeps the same feedback/control dynamic bodies but measures direct XY-sector Pauli correlators. It is not a separate paper.
Boundary¶
This is a bounded hardware-control falsification for the tested S1 policy, backend, circuit family, shots, repetitions, and calibration window. It does not establish sub-microsecond feedback, quantum advantage, or backend-general feedback behaviour.