Divergence Backtest Strategy
The Divergence Backtest Strategy is a specialized backtesting mode that simulates trading based on disagreements between PulseGrid's parametric signal and market consensus, rather than PG score thresholds alone.
How It Works
1. The engine scans historical data for periods where PG Signal and Market Signal diverge
2. When PG says Buy but Market says Sell (or vice versa), a trade is initiated following PG's direction
3. The trade is held until the divergence resolves (signals converge) or a stop-loss/take-profit is triggered
4. Performance metrics are computed across all divergence-based trades
Strategy Configuration
| Parameter | Description | Default |
|---|---|---|
| Strategy Mode | Select "Divergence" to enable this mode | PG Score |
| Min Severity | Minimum divergence severity to trigger trades (1=minor, 2=moderate, 3=significant) | 1 (all) |
| Stop Loss | Maximum loss per trade before forced exit | 5% |
| Take Profit | Target profit per trade for automatic exit | 15% |
| Position Sizing | How capital is allocated across positions | Score-weighted |
Divergence Analytics Tab
When viewing results from a divergence backtest, a dedicated "Divergence" tab appears with:
- Total Divergences Detected: How many divergence events occurred in the backtest period
- Divergences Traded On: How many met the minimum severity threshold
- Divergence Win Rate: Percentage of divergence-based trades that were profitable
- Avg Divergence Return: Mean return per divergence trade
- Breakdown by Severity: Win rate and average return for minor, moderate, and significant divergences
- Top Divergence Trades: The most notable individual trades with entry details and outcomes
Interpreting Results
A positive divergence win rate above 55% suggests that PG's contrarian signals have genuine predictive value. Compare the divergence strategy's Sharpe ratio against the standard PG Score strategy to determine whether divergence-based trading adds risk-adjusted value.
Validation with Accuracy Tracker
The Accuracy Tracker page provides a complementary view by aggregating divergence outcomes across all instruments without simulating actual trades. Use both tools together for a comprehensive assessment of PG's predictive accuracy.