DOSSIER / PROGRAM 01 / VERDICT · 29 SEPTEMBER 2026
The alert thesis did not hold.
Program 01 was built to alert when curated wallets converged on a token. We measured whether the alert could reach someone while those wallets still held the positions it described. The result did not support that claim, so the convergence-alert product is retired.
This page summarizes the published measurement in correction #021 and the dated retirement decision in #026. It is not a fresh independent computation.
Measured finding
Of the measured wallet positions that triggered an alert, as reported in #021.
Median hold among positions for which a sell was recorded. This is not the median of all positions.
Median from last qualifying buy to message, down from roughly 23 hours. Speed alone did not resolve the miss.
The unknowns belong beside the headline.
54.8% of positions had no recorded sell; we cannot distinguish continued holding from remaining capture gaps. About 18% of triggering positions could not be matched to a buy in their own convergence window and were excluded from that analysis. The public #021 entry does not give a verified sample size, so this page does not present one.
These limits make the result narrower than a claim about every wallet or every trade. They do not rescue the original promise that a fast convergence alert puts the reader in front of the observed positions.
Program 01 is a historical record now.
The original Signals and Handler convergence-alert product will not resume under its old rules or claim. Ingestion and detection remain stopped. There is no subscriber alerting, waitlist intake, paid tier or reopening date attached to it. The original method and measurement remain public; #021 has not been rewritten to pretend this decision was made in August. The recorded position for prior signups in #023 is also preserved.
Any future convergence product would need a new specification, fresh measurement and a separately recorded authorization. The old detector is not a standing offer waiting for a switch to be turned back on.
Inspect structure. Preserve evidence. Follow supported movement.
The Token Structure Lens is a separate read-only public beta: paste a Solana mint, inspect recorded controls and up to 20 largest token accounts, and see source slots and explicit unknowns. It does not identify people or rate a token's safety.
A bounded Capital Path trail and opt-in later-purchase watch are in development behind evidence, database, provider-budget and sender gates. They are not operating alerts today. Where supported, the claim will be that the same recorded controller sold one token and later bought another, never that the exact proceeds funded the purchase.
Source and revision
- #020 — parser gap and corrected trade capture.
- #021 — original measurement and caveats, published 7 August 2026.
- #022 — halted ingestion and operator notification failure.
- #026 — decision to retire Program 01, dated 29 September 2026.
Verdict revision 1 · 29 September 2026. If the interpretation changes, Dossier will append a dated correction and keep these earlier entries readable.