Two results we publish because they went against us
Dealer positioning does not predict next-day direction
We tested whether dealer positioning predicts next-day direction across 888 sessions. It does not. So the platform publishes no directional view, and the interface has no bullish or bearish label anywhere.
Our features do not improve a volatility model
Adding our features to a standard volatility model made it worse out of sample. We do not claim to forecast volatility levels.
Intraday test, frozen and not yet evaluated
A pre-registered intraday test was frozen in August 2026 and will not be evaluated until November 2026. Until it reports, no intraday-predictive claim appears anywhere on this site or in the product. If it reports positively, the claim will be made once, with its sample and its p-value. This is also listed among the things Optwise is not, so a compliance reader meets it there too.
How an expiry call is settled
Every expiry carries one of two calls. We either publish a pin zone we stand behind, or we state that the book gives no usable anchor and the zone may break.
We freeze both kinds of call at the open and grade them against the exchange's official close. A zone counts as holding when the close lands within 20 points of it. We grade a may-break call on the same terms, and it counts against us when the zone holds anyway. Each call is sealed when it is frozen, so it cannot be edited.
Reference
- How it works — the screens, and when a desk uses each.
Structural analytics for education and research · not a recommendation