Falcon AI's signal engine combines institutional Smart Money Concepts with machine-learning confluence scoring to surface high-probability setups across the major futures markets in real time.
The Falcon AI signal engine operates as a multi-layered analysis pipeline running directly within TradingView on every closed bar. It does not rely on repainting indicators or delayed data — every signal is generated from confirmed price action using institutional market structure methodology.
At its core, the engine applies Smart Money Concepts (SMC) — a framework derived from how institutional participants (banks, prop desks, and market makers) structure their order flow. All detected structures are weighted and passed into the AI confidence layer, which produces a unified score for each candidate signal before it is surfaced on the chart.
The engine outputs three primary signal types, each rendered directly on the TradingView chart with entry, stop-loss, and take-profit levels pre-calculated:
Each signal includes a directional label and the confidence score (0–12).
Every signal carries an AI confidence score from 0 to 12, calculated by a proprietary weighting model that evaluates the strength and number of confluent factors present at the time of signal generation. The score is not a probability of winning — it is a normalized measure of how many institutional confluences align at a given setup.
As a general guideline: scores of 10/12 or higher represent high-confluence setups; scores of 7/12 to 9/12 represent moderate setups worth monitoring; scores of 4/12 to 6/12 are marginal; scores of 2/12 to 3/12 are low but still fire at the default minimum threshold; scores below 2/12 are considered low-confidence and the engine suppresses the signal from display.
Each signal passes through a defined four-stage lifecycle from detection to close. Understanding these stages helps traders know exactly when to engage and when a signal is no longer valid.
A signal is automatically invalidated — and removed from the active queue — if price closes beyond the structural invalidation level (typically the origin of the order block or sweep) before the entry is triggered. This prevents stale signals from lingering on the chart and misleading traders about current conditions.
On TradingView, active signals are displayed as labeled zones on the price chart. The label updates in real time to reflect the current stage. Completed signals remain on the chart as a historical record, color-coded by outcome (target hit vs. stop-loss hit) for performance review.
