Pre-Committed Falsification Criteria: Glycerol Titration Test for Coherence Window Detection in Magnetic Tweezers Hopping Data

This document records the pre-committed classification thresholds for the glycerol titration falsification test of the Markov assumption in DNA torsional hopping dynamics. All thresholds were derived and archived on 17 February 2026 (internal record DNA-CH-02) before the Walker, Vanderlinden & Lipfert (2018) benchmark dataset was examined. No threshold was modified after data examination. Primary classification thresholds: ΔCV < −0.300 and Spearman ρ < −0.900 (both required simultaneously) for Non-Markovian confirmed; |ΔCV| < 0.15 and |ρ| < 0.30 for Markov consistent; neither for Inconclusive. The conjunction requirement prevents single-outlier false positives. The document includes the formal statement of the Kramers Rate Invariance theorem (the mathematical basis for the test), the three null classifications with interpretations, the list of forbidden post-hoc actions, and the full glycerol protocol parameters (0%, 5%, 10%, 15%, 20% v/v; 100 mM NaCl; 3 pN and 4 pN force conditions; minimum N = 75 events per condition). This record constitutes the human-readable pre-registration. The executable record is the analysis script (glycerol_script.py), which has these thresholds hardcoded and does not accept runtime modification. Both records agree. The statistical framework and discriminator theorem are developed in Broomhead (2026b). The full experimental protocol is specified in Broomhead (2026c).