5.6%

Anonymous, verifiable

The operator is intentionally anonymous — not nameless. Three cryptographically and operationally checkable signals back this position, so subscribers do not have to take the anonymity claim on trust alone.

By the numbers

Phantom-data purge 2026-05-30 (P3). Previous figures aggregated paper TP/SL exits that closed at peak_price=0.5 instead of true market resolution, inflating win rate to 65% and PnL to +$1,634. The clean ledger filters to truly-resolved trades (exit_reason=resolved AND exit_price ∈ {0.0, 1.0}, i.e. market-settled binary outcome) and shows the real 30-day picture: n=36, WR 5.6%, −$252 cumulative. System is on a profitability-proven gate — no live capital deploys until N≥30 truly-resolved trades clear WR LCB ≥ 0.55, 30d PnL > 0, drawdown < 15%, Sharpe > 1.0. Current state fails 4 of 6 gates; system stays in paper/shadow until edge surfaces.

Statistical calibration

Performance metrics are reported against Bayesian posterior bounds with explicit lower-confidence-bound floors. Calibration methodology, sample construction, and full statistical workings are detailed in the subscriber methodology brief.

Eight protection layers

Every order passes through eight independent kill criteria before reaching the chain. Any layer can veto. All currently armed.

Last 10 resolved trades

Anonymized sample. Market labels obfuscated. Sizes capped per position-limit layer. Full ledger is private — this is a representative slice.

30-day equity profile

Cumulative paper-shadow P&L across the trailing 30 days on the truly-resolved ledger, finishing at −$252.18. Curve is a placeholder pending the daily-journal wiring — the per-day series ships once the live execution daemon writes its own journal. The prior +$1,634 curve aggregated phantom take-profit exits at peak_price=0.5 and was retracted on 2026-05-30 (P3 phantom-data purge).

12-week rolling win rate

Win rate by week across the trailing 12 weeks. Bars retracted — pending re-validation. The prior chart was generated on phantom-inflated win rate (65% paper-TP); the truly-resolved 30d sample (n=36, WR 5.6%, Wilson LCB 0.02) is too small to render a meaningful 12-week breakdown. Bars will be regenerated once the system accumulates ≥100 truly-resolved trades and clears the profitability-proven gate (LCB ≥ 0.55, PnL > 0).

Sample-size confidence

The 5.6% win rate is computed over n=36 truly-resolved trades (30-day window). Sample size is far below the n≥100 floor required for a stable posterior, so the published Wilson 95% lower bound sits near zero. The shaded distribution below is intentionally retained as a calibration-methodology illustration only — not a forward-looking claim. System remains in shadow-paper validation until n clears 100 and edge surfaces.