LSTM-Optimized VWAP Adaptive Trend

Family: hybrid · Regime: trending · Complexity: high · Asset classes: Equities, FX, Crypto · Timeframes: 15m, 1h, 4h

Thesis

The hypothesis is that institutional volume clusters around VWAP, and by filtering these entries through an AI-optimized volatility trend model (ML SuperTrend) and a high-frequency momentum delta (UDT), we can isolate high-probability trend continuations while maintaining strict risk control via accelerating stops.

Components

• Imported UDT typed UDF params Test (regime) — Acts as a high-frequency momentum filter, ensuring the immediate 'point-to-point' delta aligns with the broader trend.
• ML SuperTrend (Ultimate) - Auto-Optimized AI with LSTM (direction) — Provides the primary directional bias using an adaptive, volatility-adjusted trend line optimized via DQN/LSTM.
• VWAP (entry) — Identifies institutional 'fair value' entry points to avoid overextending entries into the trend.
• Donchian Channels (exit) — Used for structural exits, capturing the end of a trend when price breaks out of the recent range against the trade.
• Parabolic SAR (risk) — Calculates position sizing based on volatility-adjusted distance and provides an accelerating trailing stop.

Known failure conditions

• ML SuperTrend 'Factor' remains static for extended periods, indicating model stall.
• Market enters a tight bracket where VWAP stays centered in Donchian Channels for >50 bars.
• UDT delta oscillates around zero with no directional follow-through.

Explore the full interactive blueprint, parameter ranges and evidence on WOBR StrategyVerse, or generate this strategy as an MT4/MT5 Expert Advisor with QuantMogul AI Engine.