Results and reproducibility archive for the preregistered midpoint-tracking test HS-1.0

Results of the prospectively registered one-qubit midpoint-tracking experiment HS-1.0, acquired on ibm_kingston qubit 103 on 29 October 2026. The archive contains all 162 circuits' counts (8192 shots each; 1,327,104 total), the frozen analysis code, noiseless digital predictions, logical and compiled circuits, execution records, diagnostics, checksums and a result report. The prespecified dip-shape criterion was met: both shoulders exceed the fixed centre in all 20,000 conditional posterior samples. The left and right contrasts are 0.012988 [0.010424, 0.015520] and 0.010654 [0.008143, 0.013222]. The corrected centre error is 2.554%, above the noiseless digital prediction of 0.707878%; the data support the registered dip shape, not precise location of the minimizer or the ideal absolute error. Intervals condition on the registered binomial/readout-calibration model. No new hardware acquisition or post-hoc adjustment of the primary criterion was performed for this archive. One account-specific instance identifier was replaced by a hash, with a transparent redaction log. The original preregistration remains a separate record.