Pyramiding into winners is a hazard in a prop account
Adding a pyramiding strategy to my existing trading engine does not help when aiming for consistent capital growth.
Adding a pyramiding strategy to my existing trading engine does not help when aiming for consistent capital growth. While the logic behind adding to winning positions (often called “riding the trend”) is sound in theory, it proves fundamentally incompatible with the strict risk constraints required by professional funding programs. I spent the last few weeks updating my core engine to support multi-position logic, ensuring that the original single-position system remained identical in performance to verify that my code changes didn’t introduce errors. Once the engine was stable, I built a breakout strategy that adds units as the trend moves in my favor, using ATR (Average True Range, a measure of market volatility) to determine when to add and where to place stops. The results show a clear trade-off between growth and stability:
| Metric | Baseline System | Pyramiding Strategy |
|---|---|---|
| Return / DD Ratio | 11.7 | 14.5 to 20.2 |
| Sharpe Ratio | 1.02 | 0.84 to 0.96 |
| Max Drawdown | -7.1% | -18% to -28% |
| M1 Worst Day | -4.06% | -11.6% to -19.9% |
| MC Pass Rate | 78% | 55% to 70% |
| In other words, while the pyramiding approach improved the return-to-drawdown ratio, it caused the actual drawdown to explode by 31% to 53%. The Sharpe ratio, which measures risk-adjusted return, also declined. Most importantly, the pyramiding strategy struggled with the daily loss limits imposed by most prop firms. Because the system concentrates exposure when a trend is strong, it inevitably gives back a large portion of those gains in a single day when the market reverses. This creates a structural conflict with the “no loss over 5%” rule. | ||
| The data confirms that the M1 intraday loss is the true constraint for any system. Even when a strategy looks profitable on paper, the risk of a massive single-day wipeout makes it unsuitable for professional accounts. My tests show that the pyramiding version faces disqualification between 17% and 25% of the time due to maximum loss rules, whereas the baseline system remains well within safe boundaries. | ||
| This experiment has a silver lining. The multi-position engine I built is a permanent upgrade to my research infrastructure. It will be incredibly useful for future work, such as partial profit-taking or non-prop trading accounts where aggressive growth is prioritized over strict drawdown limits. However, for my core system, I am keeping the no-pyramid logic. Trying to squeeze out 2% monthly returns while keeping drawdowns low remains a difficult challenge, and this research confirms that pyramiding is not the path to achieving it. |
How this connects
This verification builds on earlier ones (what failed before and what I tried this time, comparisons between approaches).