Geant4 Simulation and Reconstruction: Supplementary Material for "A Monte Carlo decomposition of angular and invariant-mass resolution in a segmented π0 spectrometer: granularity, acceptance, and calorimetric response

Version 3.1.1 (2026-11-03). Adds the ntuple and the analysis of a Geant4 production with a production cut of 0.1 mm, made to test whether the Cerenkov-signal estimate of the manuscript depends on the cut. Code, geometry, physics list and estimator are unchanged with respect to version 3.1.0; the documentation of 3.1.0 is corrected (seeds, manifest). New in 3.1.1:- output_M4_cut01.root (141,898,462 bytes, SHA-256 a90f820cb5a1757cb078a04869a375817e06160923c54db57fa55796f34d3077): 1,000,000 events, /run/setCut 0.1 mm, seed 22222. Same ntuple structure as output_Cherenkov_2M.root.- analysis/m4_cherenkov/results/risultati_M4_cut01_geom.json: output of analyze_cherenkov.py --geom (photons associated to the real polar/azimuthal windows of the module faces).- Corrections: the 2x10^6-event production of version 3.1.0 was run with seed 22221; the release note of 3.1.0 had a placeholder and macros/RUN_M4.md a wrong seed. The manifest of 3.1.0 had a last line not valid for sha256sum -c. No code was changed. Result of the cut test: the Gaussian-core width of the ratio estimator/true photon energy changes by -0.7 to 0 percentage points for the Frank-Tamm-weighted track length and by -0.5 to +0.9 points for the deposited energy, with respect to the default cut of 0.7 mm. With these widths the fast Monte Carlo gives sigma_M = 7.61 +/- 0.03 and 5.37 +/- 0.02 MeV/c^2 (weighted track length and deposited energy, module-centre directions), against 7.70 and 5.07 for the default cut. Bootstrap errors were not computed for this production. The execution logs are not part of the record. Version 3.1.0 (2026-10-31): Cerenkov track-length estimator (no optical photons), output_Cherenkov_2M.root (2,000,000 events, seed 22221), analysis scripts, macros, unit test of the weight function. Version 3.0.0 (2026-10-15): complete package of the productions of 2x10^6 and 7x10^6 events, chunked reanalysis and diagnostics; all its files remain in this record. The estimator is a lower bound on the width of a Cerenkov counter: it has no light production, absorption or collection, and no photoelectron statistics. The refractive indices (F2 1.62, SF5 1.67) are nominal catalogue values. The fast Monte Carlo that uses the widths is archived in 10.5281/zenodo.22226331. Companion paper: A. Fallica, "A Monte Carlo decomposition of angular and invariant-mass resolution in a segmented pi0 spectrometer: granularity, acceptance, and calorimetric response" (under review). Licenses: MIT (code and data), CC-BY-4.0 (documentation). This deposit accompanies: A. Fallica, "A Monte Carlo decomposition of angular and invariant-mass resolution in a segmented π0 spectrometer: granularity, acceptance, and calorimetric response" (in preparation). It documents the complementary Geant4-based simulation and offline reconstruction chain described in Sec. VIII and Appendix B of that paper — a same-author, parallel study to the paper's main fast Monte Carlo, not an independent third-party simulation.