Share
SHARE TOX.
IN ONE CLICK.
Open the poster or share toxsystem.it directly.
TOX publishes the evidence trail: historical results, broader multi-asset stress tests, post-freeze results, a same-entry comparison with dynamic protection, a historical C00 €1,000 capital-gate validation, and cryptographic forward chains that keep growing.
Historical and DEMO results do not guarantee future performance. That is exactly why the forward record keeps growing.
The first question was simple: did the result depend on one asset or one market regime?
The stress block widens both the time period and the asset universe. It is not presented as a perfect test; it is there to show how the idea behaves when market conditions change sharply.
The May 2022 Terra/LUNA collapse is not removed to make the chart look better. That period remains in the official result because a serious stress test should keep its worst shocks visible.
Before we treat an edge as interesting, it has to survive tests designed to make self-deception harder.
It should not depend on a single coin.
It has to be observed on data that was not used to build it.
Once frozen, future data cannot be used to rewrite the past.
Shocks are not removed just because they hurt the result.
The core was frozen on August 14, 2026. From that point on, the question was no longer “how good did the past look?” but “what happens once the rules can no longer be changed?”
The period through September 18 was sealed prospectively; the later segment through the start of the daily chain is a temporal reconstruction under the same frozen identity. It is not presented as stronger evidence than it is.
We did not change the entries. We compared the original management with a frozen dynamic-management layer on the same post-freeze trades.
average net bps per trade on the comparable post-freeze sample.
average net bps per trade with the frozen dynamic management.
In the post-freeze comparison, 3,168 trades improved, 1,044 worsened and 7,839 were unchanged. The protection logic remains private; the published evidence covers the results and the frozen identity, not the operating rules.
A result is weak if it assumes infinite capital. We re-ran history under the C00 gate: €1,000 start, 10% initial margin of equity, max 12 opens, nominal 5× leverage, chronological accept-when-affordable, and compounding. When initial margin is already locked, the signal stays out. In the published sample, accepted-subset expectancy stays close to the full stream.
Short windows can be negative. In the published sample the C00 constraint did not erase the observed historical edge. This is not a maintenance-margin or forced-liquidation model, not DEMO fills, and not a guaranteed realizable return.
Historical C00 simulation with equity compounding. Distinct from the prospective €1,000 shadow chain and from operational DEMO.
At the close of each UTC day, the system publishes a new evidence record, advances the cryptographic chain and refreshes the downloadable package. Corrections are append-only: earlier records are not rewritten.
loading…Loaded directly from the current public evidence artifacts.
Download the evidence and ask it to look for errors, cherry-picking, limitations and contradictions. Choose whichever AI you prefer: the review starts from the public files, not from our opinion.
Once you have reviewed and challenged the data, one simple question remains: does the product actually work in your environment? That is what the DEMO is for.
The public build can be verified separately from the statistical evidence.
Open the poster or share toxsystem.it directly.