Version: Evidence report · 11 September 2026 Author: Arthur Gardner Entity: NanŌmics™ PBC Contact: info@nanomics.io Status: Experiment · not a Demonstrated result · not a proof Patents: 8 U.S. provisionals filed · 11 additional draft-ready Companions: Distributed Thermodynamic Security PUBLISH-4.0 · Physical Reality Sentry whitepaper v1.1
This report is the board and the figures for the 11 September 2026 experiment cited in Distributed Thermodynamic Security PUBLISH-4.0. Forty-seven attack-surface cases were simulated against the five-class accept rule: physical, cryptographic, logistical, statistical, and temporal, evaluated together. Forty-five held: the case was rejected or never became a certified leaf. Two honest controls were accepted, as expected. The residual count is zero.
A certified leaf is stored only when the five-class accept succeeds and the meter moves for that accept. A software-hash miss is a notice inside an existing class. It is not a sixth door. The 26 May 2026 production entropy result remains the Demonstrated five-of-five on that path. This board does not change that label. Chance-genuine values on the May figure are the same rounded figures already published in the IACR evaluation table. The accept threshold is omitted.
The category paper and the product whitepaper keep the short sentence and the pointer. This file holds the pictures and the row list. Hover a colored cell or a point for the case note.
Label for this board: experiment. Label for the 26 May six-case table: Demonstrated on the production entropy posterior of that day. The Gardner Conjecture is not proved here.
| Item | Count | Meaning |
|---|---|---|
| Simulated | 47 | Attack-surface cases run on the enrolled-write path |
| Held | 45 | Rejected, or never stored as a certified leaf |
| Honest controls | 2 | Accept expected; certified leaf; meter moved |
| Residual | 0 | No leftover simulated case |
Green = the defense held. Blue = honest control, accept expected. Red = no certified leaf, or the meter did not move. The software-hash notice case (S12) is held as a class result and still stores a certified row, because a hash miss is a notice, not a sixth class.
HeldControlNo certified row / meter did not move
Y is chance-genuine, rounded as in the published IACR table (A1 0.005 · A2 0.174 · A3 0.120 · A4 no score, plotted at 0 · A5 0.120 · control 0.984). Green accepted. Red rejected. The accept threshold is implementation-defined and is not printed. This figure does not add rows to the May five-of-five.
Same chance-genuine axis. These three are not in the May five-of-five. Digital replay and a second copy of genuine numbers still score high (0.984). The temporal class is what must reject a used epoch. Fake audio scores 0.054 and is rejected on score.
This report publishes the 47 cases that were run. Forty-five held. Two honest controls were accepted. The residual count is zero.
The combined machine receipt for the same day is a companion artifact in the NanŌmics resources section. This HTML is the public board.
Gardner, A. (2026). Distributed Thermodynamic Security (DTS) — Pre-NDA Overview. PUBLISH-4.0. NanŌmics PBC. https://nanomics.io/resources.
Gardner, A. (2026). Physical Reality Sentry Platform — Technical Whitepaper. Version 1.1. NanŌmics PBC. https://nanomics.io/resources.
Gardner, A. (2026). IACR DTS Security Model. Version 1.2. May evaluation table (chance-genuine figures used in Figure 2).
NanŌmics PBC · Public Benefit Corporation · Delaware, USA
Contact: info@nanomics.io