Endpoint-anchored reporting of protocol-dependent sphere-packing fractions in a finite container

Abstract

This record contains the preprint and reproducibility materials for "Endpoint-anchored reporting of protocol-dependent sphere-packing fractions in a finite container" by Jun Tsuzurugi. The study reports a finite-container experiment comparing gently poured aluminum spheres with spheres subjected to 200 counted manual taps. Packing fractions are recalculated from recorded dry and saturated masses under the stated water-density and aluminum-density conventions. Six matched trials per condition give packing fractions of 0.570 +/- 0.008 (sample SD) for the poured preparation and 0.607 +/- 0.012 (sample SD) after 200 taps. The mean paired increase is 0.037, with an approximate 95% confidence interval from 0.023 to 0.051. Established endpoint normalization and Hill-type inversion are used to report the condition means relative to declared reference levels. The inverse coordinate is a diagnostic re-expression of packing fraction, not an independently measured predictor or an independent validation of a packing law. The report separates trial-to-trial variability, statistical confidence intervals, assigned absolute-scale sensitivity, and endpoint/exponent sensitivity. The baseline does not identify friction, fabric, wall, or protocol-history contributions. The accompanying reproducibility archive contains the manuscript LaTeX source, the original baseline CSV, deterministic Python scripts for numerical recalculation and figure generation, Figure 1 in EPS and PDF formats, calculation summaries, and reproducibility instructions. The recorded mass fields are retained unchanged, and calculations use the masses rather than the rounded packing-fraction column. This is a preprint prepared for submission to Particulate Science and Technology, together with its reproducibility materials. It is not the journal's published version of record.