1. Build the eligible futures market universe
The scan begins with Binance USDⓈ M Futures exchange information. Only actively trading USDT-quoted perpetual contracts with valid symbol filters are eligible. Inactive, expired, malformed and stable-asset base pairs are excluded before any technical work begins.
2. Rank markets by real trading liquidity
Eligible markets are matched with their 24-hour ticker statistics. Quote volume is validated as a positive finite number, sorted from highest to lowest and used to select up to the configured top 50 contracts. If fewer markets can be validated, the interface reports the real analyzed count.
3. Lock analysis to completed market data
Each scan uses a stable completed-data cutoff. Market records that are still forming are removed before analysis, so incomplete price action cannot influence the current signal.
4. Analyze multiple market layers
Entry timing
The entry layer examines VWAP interaction, pullbacks, EMA reactions, recent swings, breakouts and volume.
Primary trend
The trend model evaluates alignment, momentum and trend strength and carries the largest influence.
Broader market bias
The context layer identifies strong directional conflicts. A strong bearish structure penalizes a potential LONG; the inverse applies to SHORT setups.
5. Calculate ten indicators plus structure and volume
The engine calculates EMA trend alignment, MACD, RSI, ADX and DMI, ATR, Bollinger Bands, Supertrend, Ichimoku Cloud, OBV and UTC-session VWAP. Market structure and raw volume add context without replacing the ten primary indicators.
6. Compare independent LONG and SHORT scores
Bullish and bearish evidence are accumulated separately across the analysis layers. A range can therefore produce two middling scores rather than forcing one side to equal 100 minus the other. Conflict and entry-quality penalties are then applied.
7. Qualify the setup and validate levels
The winning direction must reach the configured minimum Setup Score and maintain the minimum directional edge. The engine then creates an entry zone from confirmed 1H and 4H supply or demand, repeated swing reactions, displacement, volume and ATR. TP1, TP2 and TP3 target approximately 2%, 3.5% and 5.5% of the total position. A structural stop above the 5% position-loss maximum is rejected.
Support and resistance are checked in the trade direction: a LONG cannot publish when resistance blocks TP1, and a SHORT cannot publish when support blocks TP1. Invalid ordering, excessive extension or weak risk reward also rejects the setup.
8. Publish one stable result
A server-side cache and in-flight promise map prevent duplicate full scans inside one runtime. When a signal cycle changes, the old result stops being current, the scanner state appears and a new result replaces it. If a serverless runtime restarts, deterministic cutoff and ranking rules recreate the calculation without random signals.