Strategy Mechanics
- Core Logic: Perceptrader AI uses a neural network to identify trend reversals and momentum shifts. Trades are executed with a fixed stop-loss and take-profit, but position sizing is dynamic, increasing after wins and decreasing after losses.
- Asset Focus: Primarily trades major forex pairs (EUR/USD, GBP/USD) and indices (S&P 500, DAX).
- Trade Frequency: Approximately 15-20 trades per month, with an average holding period of 2-4 days.
Drawdown Analysis
- Maximum Drawdown: -15.4% occurred during a period of low volatility in August 2023. The drawdown lasted 45 days.
- Recovery Time: Historical drawdowns have recovered within 2-3 months, but the current drawdown duration is increasing.
- Equity Curve: The equity curve shows a steady upward trend with periodic sharp declines, indicative of a high-risk strategy.
Risk Warnings
- Potential Martingale: The dynamic position sizing algorithm increases lot size after consecutive wins, which can amplify losses during drawdowns. This resembles a Martingale system.
- Survivorship Bias: The signal has only been active for 18 months. Backtested results may not reflect live performance due to overfitting.
- Lack of Transparency: The exact neural network architecture and training data are not disclosed, making it impossible to verify robustness.
- Correlation Risk: High win rate (78%) but low risk-reward ratio (1:1.2) suggests the strategy relies on high frequency of small wins, which can be vulnerable to black swan events.
Performance Metrics
| Metric | Value |
|---|---|
| Monthly Return | +4.2% |
| Max Drawdown | -15.4% |
| Win Rate | 78% |
| Profit Factor | 1.8 |
| Sharpe Ratio | 1.2 |
| Average Trade Duration | 3.2 days |
Conclusion
Perceptrader AI exhibits strong returns but carries significant risk due to its dynamic position sizing and lack of transparency. The high win rate and moderate Sharpe ratio are attractive, but the drawdown profile and potential Martingale behavior warrant caution. Suitable only for aggressive risk-tolerant investors.