Transparent signal design.

1) Data inputs

Signals begin with market data and context. We analyze recent price history across multiple timeframes and combine that with volatility and session timing.

• Price action: candles, swings, and structure shifts.
• Volatility: ATR and range expansion to avoid low-quality chop.
• Session timing: filtering for liquid windows.
• Event risk: avoiding major releases when spreads widen.

2) Feature and signal scoring

We transform market data into features that capture trend, momentum, and structure quality. A scoring model estimates directional confidence based on those features.

• Trend bias: higher-timeframe alignment and slope.
• Momentum: impulse vs. correction behavior.
• Structure: break-and-retest vs. midrange noise.
• Consensus: multi-timeframe agreement.

3) Risk and quality filters

A high model score alone does not publish a signal. Additional filters enforce risk and market quality constraints.

• Minimum confidence and edge thresholds.
• Session windows and cooldowns to prevent overtrading.
• Skip rules when volatility is extreme or spreads are abnormal.
• Global limits on open trades across pairs.

4) Signal output

Each published signal includes the essential trade structure so you can make a disciplined decision.

• Pair and direction (buy or sell).
• Entry zone aligned with the model output.
• Stop loss and take profit derived from volatility and structure.
• Confidence score to help rank trades.

5) How to use the signals

1. Check higher-timeframe bias and key levels.
2. Confirm that spreads and volatility are normal.
3. Size your position based on your risk rules.
4. Log the trade and review performance after exit.

6) Limitations and risk

Models can fail during regime shifts, news shocks, or data outages. Spreads, slippage, and execution quality can change real outcomes. Always apply your own judgment and risk limits.