Engine A — Validation
Published historical tests for the long-horizon USDJPY strategy: what the results support, where they are weaker, and how they relate to the deployed engine.
Past performance, including backtested and simulated results, does not guarantee future returns. Trading leveraged instruments involves risk of loss.
Historical evidence · live validation ongoing
Profitable OOS slices
4/4
Overlapping periods; not independent trials
OOS profit factor range
1.15–2.18
Across the published slice reports
Largest OOS equity DD
37.81%
At the sizing used in these backtests
All four published out-of-sample (OOS) runs closed profitably. Results vary: A1 is the weakest slice, while later slices show stronger profit factors. This supports continued evaluation of the strategy across different periods.
The live engine uses the full-window 2014–2025 parameter set. The rolling slices evaluate the strategy family; they do not make that final configuration out of sample throughout 2014–2025. Subsequent observed live results provide additional evidence.
Walk-Forward Timeline
Each slice pairs an in-sample (IS) development window with a later OOS test window. The reports cover research from 2014 and OOS tests from 2019.
| Slice | IS Period | OOS Period | OOS Result |
|---|---|---|---|
| A1 | 2014–2018 | 2019–2020 | Profitable · lowest PF |
| A2 | 2016–2020 | 2021–2022 | Profitable |
| A3 | 2018–2022 | 2023–2024 | Profitable |
| A4 | 2019–2023 | 2024–17 Nov 2025 | Profitable |
A3 and A4 both include 2024. The four profitable slices are overlapping historical results, not four independent trials or a single non-overlapping OOS track record.
A1: the weaker period and financing contribution
The 2019–2020 OOS report has a profit factor of 1.15, with 32.97% balance drawdown and 37.81% equity drawdown.
Its closed-trade net profit is €715.19, including €1,037.79 in positive swaps. Subtracting those swaps from the recorded trades leaves -€322.60. This is an accounting breakdown, not a zero-swap rerun: changing financing can also change subsequent position sizes.
Carry can contribute to an FX strategy, but the result depends on its financing assumptions. These exports alone do not establish that historical swap rates were reproduced throughout the test.
Engine A Walk-Forward Slices
| Sample (IS/OOS) | Win Rate | Profit Factor | Balance Sharpe* | Balance DD | Equity DD | Mean trade return | Trades |
|---|---|---|---|---|---|---|---|
| In-Sample | 16.6% | 1.44 | 1.90 | 24.3% | 34.1% | 0.43% | 1419 |
| Out-of-Sample | 14.4% | 1.15 | 0.79 | 33.0% | 37.8% | 0.13% | 541 |
*Balance Sharpe is an annualized return-to-volatility estimate based on closing balances, carrying quiet weekdays forward and assuming a zero risk-free rate. It excludes daily unrealized P/L. Mean trade return is net P/L divided by each trade’s pre-trade balance, averaged across recorded trades; it is not a risk-normalized R multiple.
Behavioral Structure
IS
-6.66
OOS
-3.76
IS
92.2%
OOS
96.3%
Drawdown Stress
IS
9.45
OOS
12.32
Across-slice statistics and their limits
Descriptive variation across the published slices, including overlapping periods. These are not independent samples or confidence intervals; low variation alone does not establish robustness.
Performance metrics
| Metric | Mean | StDev | CV | Min / Max |
|---|---|---|---|---|
| Profit Factor (OOS) | 1.75 | 0.38 | 0.215 | 1.15 / 2.18 |
| Balance Sharpe (OOS) | 2.00 | 0.74 | 0.368 | 0.79 / 2.73 |
| Balance DD (OOS) | 27.35% | 4.51% | 0.165 | 20.38% / 32.97% |
Structural metrics
| Metric | Mean | StDev | CV | Min / Max |
|---|---|---|---|---|
| Runs Z (OOS) | -3.93 | 0.63 | 0.161 | -4.72 / -3.00 |
| Ulcer Index (OOS) | 10.02 | 2.04 | 0.204 | 6.95 / 12.32 |
| Observations Below High (OOS) | 91.83% | 3.15% | 0.034 | 88.27% / 96.32% |
Reading the evidence
The published exports document simulated outcomes under their recorded settings. They do not independently prove the absence of data leakage or selection bias, establish parameter-neighborhood stability, or guarantee that live execution will reproduce the results.
Drawdowns describe the risk experienced at the sizing used in each test. Balance drawdown excludes open-position losses; equity drawdown includes them. Neither is a limit on future loss.
Watch the original WFA review →Monte Carlo · Historical Resampling
Source: Engine A demo-condition research export · 11y 11m 13d (19/01/2014 - 01/01/2026) · 3,288 balance-return samples · 10,000 simulated paths.
This is a separate research export from the full-window report used on the engine overview. The simulation describes variability conditional on this input, not the likelihood that the strategy will succeed in live trading.
Each simulated path independently resamples historical balance returns with replacement. This removes their original time ordering and does not preserve regime or loss clustering, model new execution costs, or establish an edge independently of the source backtest. Drawdowns use closing balances and exclude intraday equity losses.
The blue FPT line is a continuous live-condition backtest from account inception, normalized to the simulation’s starting capital. It shows a simulated closing-balance path, not observed live returns or a comparison at matched deployment risk.
Equity Path Simulation
10,000 paths · 3,288 trades resampledMedian Max Drawdown
34.1%
95th pctl: 47.5%
Original Max DD
34.9%
Backtest (single path)
Median Final Multiple
×59,977
5th–95th: ×4,189 – ×967,804
Original Final Multiple
×62,390
Backtest (single path)
Resampled Paths Above Each Drawdown Threshold
>20%
100.0%
>30%
77.1%
>40%
20.5%
>50%
2.8%
3,288 trades · 10,000 paths · bootstrap with replacement · seeded PRNG for reproducibility. These frequencies describe the resampled trades, not calibrated probabilities of future drawdowns.