Frequently asked questions.
If your question isn’t here, email founder@omniquant.trade — we answer everything.
General
Is OmniQuant financial advice?+
Is OmniQuant legal in my country?+
Is OmniQuant a trading bot? Can it trade for me?+
How is OmniQuant different from other AI finance tools?+
Predictions
Why do you abstain from predictions sometimes?+
What happens when the model is wrong?+
How current are the predictions?+
Accuracy
Can OmniQuant guarantee accuracy?+
How do you measure accuracy?+
Directionally, on the real close-to-close move. A call is right if the price moves the way we said by its horizon, measured on official closes — not intraday ticks. Every prediction is locked before its outcome window opens, so it’s out-of-sample every time.
Two exclusions are worth knowing about, because they change the denominator. Moves too small to be a direction — smaller than the symbol’s own noise band — are thrown out of both the numerator and the denominator rather than being counted as wins; the band rule in force is: move smaller than the symbol's noise band, resolved at validation time: 0.5× the symbol's 20-day daily-return stdev clamped 0.1–5% where a volatility snapshot exists (about half of settled rows), otherwise a flat 0.3% default. The band is not scaled by horizon. That currently removes 17.4% of settled calls. And a confidence floor may exclude the calls the model itself scored lowest. No confidence floor is applied — every scored prediction is in that cohort, including the ones we were least sure of.
The published cohort definition, exactly as the engine reports it: Strict directional accuracy — a prediction is scored correct only when the symbol moved beyond its noise band and the model called that direction. Band-indeterminate outcomes (see indeterminate.band_description for the exact rule in force) are excluded from both the numerator and the denominator — not counted for or against. Cohort also excludes archived, abstained, out-of-universe, and shadow-challenger predictions. No confidence floor is applied: every scored prediction is in the cohort, including the ones the model was least sure of. From 26 October 2026 OQ publishes and scores both up and down calls. Between 22 August 2026 and 26 October 2026 only up calls were published and scored, so the down calls made then are not counted; calls made before 22 August 2026 are scored in both directions, as they were published.
The full breakdown — per horizon, per market, direction split, and the misses — is on the track record.