Directional intelligence
A statistical compass.
Your execution.
Find directional bias across markets, with a defined time window and the historical evidence behind it.
Forex · Metals · Indices · Energy · Crypto
The evidence
Did price finish in the bias direction? And did it travel further with that bias than against it?
At the end of the defined window.
Maximum favorable move exceeded the adverse move.
Jan 2024–Sep 20, 2026
Both models, weighted by resolved windows.
Historical frequencies, not a calibrated probability for any individual signal or a trading win rate. Results vary by year and market; past performance does not guarantee future results.
Inside a window
A fixed reference. A clear horizon. Real candles from start to finish, including the windows that got it wrong.
Most recent resolved hit and miss in each model at the evidence date. Examples are illustrative, not a representative sample. Moving averages provide chart context, not validated entry instructions.
Open windows identify a bullish or bearish bias with an immutable starting price.
Window progress separates an active bias from a completed historical result.
Use the direction alongside your analysis. Tradient does not supply a proven entry, stop or profit target.
The models
Multi-timeframe price and moving-average context identify directional windows across markets. The horizon is 20 observed daily bars.
Swing includes bullish and bearish windows; the historical sample is predominantly bullish.
A separate multi-timeframe price-context model identifies bullish windows across nine indices. The horizon is 12 observed hourly bars.
Short-term currently produces bullish biases only. Twelve observed hourly bars can span more than twelve clock hours.
Directional accuracy: the final price is above the starting reference for a bullish bias, or below it for a bearish bias. Flat results are not hits.
Favorable movement: the furthest price move in the bias direction exceeds the furthest move against it during the window. This does not imply that an entry was available or profitable.
Historical research windows from January 2024 through Sep 20, 2026, not audited live trade returns. History has been reused in research and is not a pristine holdout. Overlapping windows and related markets are correlated. Neither model maintains 60% in every year.
Swing excursions use M5-derived historical extremes; Short-term uses its 12 complete H1 bars. Horizons count observed bars, not fixed calendar durations. Incomplete windows are excluded. OANDA mid prices exclude execution costs and are not executable broker quotes.
Context before conviction
Read-only market research. No automated orders.