Every KairosQuant model claim must survive: a parameter REGION (not one lucky setting) → deflated Sharpe → Monte-Carlo stress (slippage, missed fills, ruin) → walk-forward out-of-sample → regime stress → repricing on REAL option chains → de-duplication → removal of the top winners. Strategies that fail are archived publicly in our research ledger as null results. We killed our own flagship setup this way — that is the standard.
Every number carries a trust label (LIVE / STALE / MISSING / HISTORICAL / PAPER / SIMULATED / DATA_BLOCKED). Stale marks are hidden, never displayed as current. Dealer-gamma and positioning values are ESTIMATES from public open interest — actual dealer books are not observable, and anyone claiming otherwise is guessing. Receipts are append-only; corrections are separate timestamped events.
No personalized investment advice. No guaranteed-profit claims. No directive buy/sell calls. No performance marketing built on cherry-picked winners. Education and research tooling only.