Random Forest Research Engine
ApexTrend’s AI-driven scoring system draws on every recorded trade and all historical research to evaluate the setup you’re looking at. In milliseconds it produces a simple 1 to 10 score. A higher number reflects how often similar setups have outperformed in the past. This synthesized “AI score” helps you make faster, more confident decisions by summarizing thousands of prior trades and conditions into a single, easy-to-read number.
How It Learns
The model is trained on a large database of historical bullish setups, capturing conditions at the moment of entry and logging what happened next. Each sample includes:
- Signal trigger time and entry price
- Performance after trade execution (profit/loss, price change, time held)
- Maximum drawdown before peak gain
- Complete indicator and pattern context at the entry point
By training on thousands of these real-market setups, the Random Forest engine learns what combinations of factors most often lead to strong short-term results.
Feature Engineering
The RF model is powered by dozens of indicators and flags computed on a minute-by-minute basis:
- Trend strength (EMA slope, VWAP alignment, MACD state)
- Momentum factors (RSI, ADX, volume spikes)
- Volatility zones (Bollinger Band position, high/low wicks)
- Candle signals (red-to-green, flat top, inside bar breaks)
Backtesting & Continuous Learning
Every time the model would have triggered a trade, the system logs the outcome and updates the signal library. This allows for constant improvement:
- Logs actual PnL at the time of exit
- Captures peak price and lowest pullback after trigger
- Measures which indicator combinations have the best historical returns
- Supports risk filtering and future condition weighting
This feedback loop forms the foundation of ApexTrend's short-term research strategy engine constantly refining its accuracy and adaptability as the market evolves.