Engine B — Validation
Medium-term USDJPY breakout engine (2018–2025). Five rolling walk-forward slices; B5 selected for production for best IS/OOS alignment and lowest fragility.
Walk-Forward Timeline
Five rolling slices spanning July 2018 to 2025.
| Slice | IS Period | OOS Period | OOS Result |
|---|---|---|---|
| B1 | 07/2018–2020 | 2021 | Profitable |
| B2 | 2019–2021 | 2022 | Profitable |
| B3 | 2020–2022 | 2023 | Profitable |
| B4 | 2021–2023 | 2024 | Profitable |
| B5 | 2022–2023 | 2024–2025 | Profitable |
All five slices profitable OOS. B5 selected for production — best IS/OOS alignment, lowest fragility.
Engine B Walk-Forward Slices
| Sample (IS/OOS) | Win Rate | Profit Factor | Sharpe | DD max | Trades |
|---|---|---|---|---|---|
| In-Sample | 16.4% | 1.37 | 1.99 | 30.3% | 584 |
| Out-of-Sample | 21.3% | 1.60 | 3.28 | 18.3% | 277 |
Behavioral Structure
IS
-3.99
OOS
-2.67
IS
93.7%
OOS
89.6%
Drawdown Stress
IS
10.65
OOS
7.10
Performance Stability
| Metric | Mean | StDev | CV | Min / Max |
|---|---|---|---|---|
| Profit Factor (OOS) | 1.78 | 0.39 | 0.219 | 1.16 / 2.19 |
| Sharpe (OOS) | 2.99 | 1.11 | 0.372 | 0.98 / 4.14 |
| Max DD % (OOS) | 20.24% | 5.59% | 0.276 | 13.64% / 27.05% |
Structural Stability
| Metric | Mean | StDev | CV | Min / Max |
|---|---|---|---|---|
| Runs Z (OOS) | -2.32 | 1.37 | 0.592 | -4.02 / 0.16 |
| Ulcer Index (OOS) | 8.00 | 2.94 | 0.368 | 5.01 / 13.45 |
| Percent Time Below High (OOS) | 90.52% | 2.96% | 0.033 | 87.54% / 96.15% |
Validation Summary
Medium-term USDJPY breakout engine (2018–2025). Adaptive structural logic, deterministic execution.
WFA Slices
5/5 profitable
Selected
B5
Period
2018–2025
Walk-Forward Findings (B1–B5)
- B1 — Moderate stability
- B2 — Strong and balanced
- B3 — Weakest slice but profitable
- B4 — Excellent stability
- B5 — Best alignment with modern volatility
Parameter Behavior
- Controlled drift across slices
- No chaotic jumps
- Stable PF/DD characteristics
Independent Review Checks
- Slicing integrity was verified
- No overfitting indicators were detected
- Behavior matches expectations for a medium-horizon structural engine
Robustness Conclusion: Engine B passes WFA with stable OOS results. B5 is selected for live deployment due to regime alignment and low fragility.
Monte Carlo Stress Test
Simulations were generated via bootstrap resampling of historical trade returns (with replacement), producing 10,000 equity paths to estimate variability in cumulative returns and drawdown distribution.
Results reflect resampled historical trade returns under the same position sizing model and are not forecasts of future performance.
Equity Path Simulation
10,000 paths · 1,088 trades resampledMedian Max Drawdown
26.1%
95th pctl: 38.0%
Original Max DD
28.5%
Backtest (single path)
Median Final Multiple
×606
5th–95th: ×85.2 – ×4,788
Original Final Multiple
×614
Backtest (single path)
Drawdown Probability
>20%
90.3%
>30%
26.5%
>40%
3.5%
>50%
0.2%
1,088 trades · 10,000 paths · bootstrap with replacement · seeded PRNG for reproducibility