The record, including the failures.
Every result stays here. A failed test gets the same space and the same publication deadline as a passing one.
| Experiment | Registered | Question | Pass criteria | Outcome |
|---|---|---|---|---|
| HOMOLOGATION · ML turn panel | 11 Sep 2026 19:23 UTC | Does a diverse frozen ML team improve one-hour payoff over the original pair and matched controls? | All six predeclared tests; read the criteria | Pending · starts 14 Sep 2026 UTC |
| One-hour forest · archived development | Not pre-registered; inspected development data | Can a forest identify cost-positive moves? | Follow-up checks on activity, concentration, and constant-direction controls | Not evidence of a durable edge. Frozen later-data test: no trades. |
HOMOLOGATION · The next test
Loading the recorded experiment status…
Window: 14 September 00:00 UTC to 19 September 00:00 UTC, 2026. Results are scheduled for publication after 19 September 01:10 UTC, within 48 hours of close. Missing data or failed criteria count as not passed; the window will not be extended.
Read the standalone pre-registration →
One-hour fixed-stop simulations, not the live trailing-stop account. Account sizing: N/A; no capital was allocated. Pip results cannot be read as account returns.
An earlier positive result did not survive scrutiny.
Evidence snapshot: 9 September 2026. This is the superseded fixed 3×ATR-stop forest configuration, with a one-hour horizon; it does not describe the current SMA trailing-stop system.
- Net result
- +145.0 pips · simulation; account sizing N/A
- Without the top three winners
- −26.2 pips · simulation; account sizing N/A
- Active days
- 14 days over four development periods; 20 trades, all buys.
- Constant-direction control
- Always-buy at the same opportunities matched the model. Simulation; account sizing N/A.
- Frozen newer-data test
- 0 trades from 1 July to 9 September 2026. No evidence of profitable execution.
Longest interval between recorded historical trades: 68.2 calendar days, including calendar and data gaps. It is not a claim about continuously observed market inactivity.
Source: versioned forest-development-audit and forest-external research records. Read the dated manuscript and limitations.