Satellite-2: the third system, with its numbers
I have successfully automated the "YosugaDow" manual trading method, resulting in a system that proves discretionary logic can be quantified into a…
I have successfully automated the “YosugaDow” manual trading method, resulting in a system that proves discretionary logic can be quantified into a reliable edge. I developed the Satellite-2 system by focusing on pullbacks following a Dow Theory trend reversal, combining this with multi-timeframe analysis and proximity to statistically significant horizontal price levels.
Performance Overview (2015 to 2026)
This system is designed to operate on five specific pairs (XAUUSD, USDJPY, GBPJPY, EURJPY, and CHFJPY) using the H1 timeframe at a 0.3% risk per trade. I tested a broader basket of non-JPY pairs, but that approach diluted the performance and led to a drawdown of 19%. This resulted in disqualification. Restricting the system to these five robust pairs yielded the following results:
| Metric | Result |
|---|---|
| Total Return | +66.3% |
| Average Monthly Return | 0.38% |
| Max Drawdown | -7.84% |
| Profit Factor (PF) | 1.62 |
| Sharpe Ratio | 1.16 |
| In other words, the profit factor of 1.62 means the system generates 1.62 units of profit for every 1 unit of loss, while the Sharpe ratio of 1.16 indicates a smooth, consistent growth curve. |
Stress Testing and Reliability
The system passed my internal “STEP 1” verification, showing profitability in 9 out of 11 years. Notably, it performed well in both the OOS (out-of-sample) period with +5.0% and the IS (in-sample) period with +60.8%. I also subjected the system to rigorous stress tests:
- Monte Carlo (MC) Pass Rate: 84% for STEP 1 and 75% overall.
- M1 Intraday Worst: The worst single-day loss found by rebuilding equity from 1-minute bars was 1.52%, with zero days breaching the critical threshold.
- Correlation: The system maintains a 0.52 correlation with my core system.
Strategic Positioning
Because of its high Sharpe ratio, this system is well-suited for regular profit withdrawals. However, my analysis suggests that simply adding this to my existing core system does not significantly improve the overall portfolio Sharpe ratio or Monte Carlo results. Instead of being treated as a simple “add-on” to increase volume, it serves better as a high-quality alternative or a parallel system to run alongside the core. My core system (v1.1.0) remains unchanged. The primary significance of this project is the validation of a long-standing hypothesis: that discretionary logic can be successfully translated into mechanical, numerical rules. While common wisdom suggests that pullback-based discretionary strategies disappear when subjected to forward testing, I have effectively overturned that expectation by correctly quantifying horizontal price levels.
How this connects
This verification builds on earlier ones (what failed before and what I tried this time, comparisons between approaches).