Purpose
Combines three classic Western technical analysis approaches into one composite signal engine:
| Dimension | Indicators | Purpose |
|---|---|---|
| Trend | EMA(12/26) + ADX(14) | Determine direction and trend strength |
| Mean reversion | Bollinger Bands(20,2) + RSI(14) | Detect overbought and oversold conditions |
| Volume-price | OBV + volume ratio | Confirm volume participation |
Signal Logic
Three-dimensional voting mechanism:
- Long: trend is bullish + RSI is not overbought + OBV is rising
- Short: trend is bearish + RSI is not oversold + OBV is falling
- Stand aside: mixed signals
Key Implementation Details
- RSI and ADX use Wilder EWM (
ewm(alpha=1/period)), not a rolling mean - Full ADX chain: +DM/-DM → TR → +DI/-DI → DX → ADX
- OBV =
(volume * sign(close.diff())).cumsum()
Parameters
All parameters have default values and can be overridden at instantiation time:
| Parameter | Default | Description |
|---|---|---|
| ema_fast | 12 | Fast EMA period |
| ema_slow | 26 | Slow EMA period |
| adx_period | 14 | ADX calculation period |
| adx_threshold | 25.0 | ADX trend-strength threshold |
| bb_window | 20 | Bollinger Band window |
| bb_std | 2.0 | Bollinger Band standard deviation multiplier |
| rsi_period | 14 | RSI period |
| rsi_oversold | 30 | RSI oversold threshold |
| rsi_overbought | 70 | RSI overbought threshold |
| vol_ma_period | 20 | Volume moving-average period |
| obv_ma_period | 20 | OBV moving-average period |
Dependencies
pip install pandas numpy requests
Signal Convention
1= long,-1= short,0= stand aside