Backtesting
Backtesting runs the PG scoring algorithm against historical data to evaluate how well the scores would have predicted actual price movements.
How It Works
1. Select an instrument and a historical date range
2. PulseGrid™ replays all events that occurred during that period
3. At each point in time, the algorithm computes what the PG Score would have been
4. The actual subsequent price movement is compared against the score's prediction
Backtest Output Metrics
| Metric | Description | Good Value |
|---|---|---|
| Total Return | Cumulative return if you had followed the PG score signals | Positive and higher than buy-and-hold |
| Max Drawdown | Largest peak-to-trough decline during the backtest period | Lower is better; below -20% indicates significant risk |
| Win Rate | Percentage of signals that correctly predicted direction | Above 55% is meaningful; above 65% is strong |
| Profit Factor | Gross profits / Gross losses | Above 1.0 means profitable; above 1.5 is good; above 2.0 is excellent |
| Sharpe Ratio | Risk-adjusted return of the strategy | Above 1.0 is good; above 2.0 is excellent |
Backtest Equity Curve
The equity curve chart shows how a hypothetical $10,000 investment would have grown (or declined) following the PG score signals. Compare it against the "Buy & Hold" line to see if the algorithm added value.
Limitations
- Survivorship bias: Backtests only include instruments that still exist today
- Look-ahead bias: The algorithm is designed to avoid this, but event timing precision may vary
- Transaction costs: Not included in the backtest results. Real-world returns would be lower.
- Past performance: Historical accuracy does not guarantee future results