Methodology & credibility

The science

Our protocol follows the lineage of the Kolodziejzyk 13-year associative remote viewing program and the broader ARV literature. We treat those results as motivating but unproven at our scale — which is precisely why everything we publish is measured and verifiable rather than asserted.

Integrity safeguards

• Blinding. Viewers and their apps are only ever sent blinded trials — two opaque images, no market, no result. The systems that know the market identity cannot see viewer impressions, and vice-versa (three-zone isolation).
• Commit-reveal. The secret image-to-result pairing is hashed on-chain before viewing and revealed after resolution; predictions are committed before they resolve. Both are verifiable.
• Sentinels & displacement monitoring. Known-answer trials and statistical monitors continuously check for leakage or drift.
• Contamination-safe reporting. Public surfaces show resolved-only aggregates, so studying them confers zero edge on an open question.

Open posture

• Pre-registration. Hypotheses, filters, stopping rules and public kill criteria (the z / P&L thresholds at which we halt) are published in advance and hashed on-chain.
• Open dataset. A resolved-only dataset plus a recompute script let anyone reproduce the headline numbers.
• Red-team bounty. We pay people to find leakage or flaws — the best defense against "it's just leakage."