Optimization Analysis

Period Recommendation Report
June 02, 2026
OPT-20260602-191938
Portfolio Composition — 5 Strategies
#STRATEGYSYMBOLTFTRADES
1The Nomad Trader EA 71.0AUDUSDH1394
2The Nomad Trader EA 71.0GBPUSDH1521
3The Nomad Trader EA 71.0NZDUSDH1637
4The Nomad Trader EA 71.0USDCADH1406
5The Nomad Trader EA 71.0USDCHFH1370
Portfolio (5 Tests)
MULTI • MIXED
2,324 trades
2,320 days (6.4 years) analyzed
Portfolio
01 Equity Projection 1,000 simulations · 1,099 trades forward · ~36.6 months
MONTE CARLO EQUITY PROJECTION — Portfolio (5 Tests) Starting: $126,198.35
Historical equity
Simulation paths
Median (+174.6%)
Best case ($493,429)
Worst path (−9.9% DD)
5TH PCT
$276,681
+119.2%
25TH PCT
$318,148
+152.1%
MEDIAN
$346,568
+174.6%
75TH PCT
$376,540
+198.4%
95TH PCT
$430,747
+241.3%
95% DD CONFIDENCE
−6.7%
Worst DD in 95% of simulated paths
AVG RECOVERY
6 trades
Mean trades to recover from a DD episode
EST. DURATION
~36.6 mo
At current trade frequency (1.0 days/trade)
SIMULATIONS
1,000
Percentage-based resampling (exact method)
STATISTICAL OBSERVATIONS
DISTRIBUTION SHAPE
The outcome distribution is right-skewed — the upside range ($276,681–$430,747) spans $154,066. Adverse scenarios are bounded closer to the starting balance than favourable ones.
DRAWDOWN PROFILE
The 95% confidence drawdown of −6.7% is the stress-test reference — the worst expected drawdown in 19 out of 20 simulated scenarios. The median path experiences materially lower drawdowns.
TRADE FREQUENCY CONTEXT
At 1.0 days per trade, the 1,099-trade projection spans approximately 36.6 months. Actual duration varies with market conditions — lower-frequency periods extend the timeline without changing the statistical distribution.
RESAMPLING METHOD
Simulations use percentage-based resampling of historical trade returns. This preserves the return distribution but assumes trade independence. Structural regime changes may produce outcomes outside the simulated range.
Statistical context: Monte Carlo projections are probabilistic estimates based on historical trade return distributions. They do not predict future outcomes. Results should be interpreted as a statistical characterisation of the strategy's historical behaviour under resampling, not a forecast. Past simulation results are not indicative of future performance.
BACKTEST WINDOW ANALYSIS
Recommended Period
1447 days
4.0 years
Statistical + Academic + Bootstrap

What This Means

Based on analysis of your strategy's characteristics on MULTI MIXED trading systems, we recommend using 1,447 days (4.0 years) of historical data for optimization and backtesting.

This period balances having enough data for statistical significance while avoiding outdated market conditions. The recommendation is synthesised from three independent methods: rolling-variance statistical analysis (767 days (2.1 years)), literature-informed period benchmarks (2,737 days (7.5 years)), and Politis-Romano block bootstrap spectral analysis (837 days (2.3 years)).

02Analysis Metrics
Statistical Period
767 days (2.1 years)
Academic Period
2,737 days (7.5 years)
Bootstrap Period
837 days (2.3 years)
Sharpe Ratio
3.34
Max Drawdown
-7.2%
Recovery Factor
50.8
Volatility Ratio
0.32x
Confidence Level
MEDIUM
Re-opt Frequency
Semi-annual ↓ (low vol)
Regime Changes
18
Profit Factor
1.97
Win Rate
62.7%
03Three-Method Synthesis

📈 Statistical Analysis

Rolling-variance stability test finds the lookback window where your strategy's return distribution is most stable and least regime-dependent.

Optimal period found767 days (2.1 years)
Regime changes detected18
Volatility adjustment0.32× current/historical

📚 Literature Benchmarks

Period benchmarks for MULTI MIXED strategies, informed by asset class volatility profiles and FX market microstructure research. These are calibrated heuristics, not direct citations.

Base period (MULTI MIXED)2,737 days (7.5 years)
Volatility adjustment↓ low vol
Adjusted period2,737 days (7.5 years)

🔬 Block Bootstrap — Politis-Romano (1994)

Spectral autocorrelation analysis estimates the memory structure in your returns using block resampling. Accounts for serial dependence that rolling-variance methods miss.

Bootstrap spectral period837 days (2.3 years)
MethodPolitis-Romano stationary bootstrap
Combined (all three)1,447 days (4.0 years)
04Why This Period Works

Interpreting the Optimal Range

Too short (under 940 days (2.6 years)): Not enough data to capture full market cycles. Results are statistically unreliable and prone to overfitting to a single regime.

Above range (over 2,098 days (5.7 years)): The window extends beyond the calculated optimal range. Older data points contribute proportionally less signal in recent-regime models.

The sweet spot (940 days (2.6 years) – 2,098 days (5.7 years) for MULTI MIXED): Provides enough trades for statistical confidence while focusing on recent, relevant market conditions.

The three-method synthesis gives 1,447 days (4.0 years) — averaged from rolling-variance statistical analysis (767 days (2.1 years)), literature-informed period benchmarks (2,737 days (7.5 years)), and Politis-Romano block bootstrap spectral analysis (837 days (2.3 years)).
05Your Backtest Period Diagnostic
◉
Period Assessment
ABOVE RANGE
The backtest window is moderately longer than the calculated optimal range.
Period Coverage
160%
2,320 days (6.4 years) of 1,447 days (4.0 years) optimal
Trade Coverage
1162%
2,324 of 200 estimated minimum
Detailed Findings
Backtest window above calculated range — 2,320 days (6.4 years) vs 1,447 days (4.0 years) optimal
The backtest window differs from the calculated optimal by 873 days (2.4 years).
⚠️ Sharpe Ratio 3.34 — Exceptionally High
López de Prado (2014) shows Sharpe ratios above 3.0 are extremely rare in professional trading. Top-tier hedge funds rarely sustain above 2.5. With 2,324 trades, the standard error of your Sharpe estimate is ±0.02, meaning the true Sharpe could realistically be 3.3–3.4 at 95% confidence. Verify backtest conditions match live: spreads, slippage, rollover costs.
⚠️ 18 Regime Changes Detected
Multiple structural shifts detected in the analysis window. Consider walk-forward analysis to confirm parameter robustness across different regimes.
References

Bailey, D. et al. (2014) "The Deflated Sharpe Ratio" — Minimum track record length, multiple testing correction
López de Prado, M. (2014) "The Deflated Sharpe Ratio" — MinTRL formula, Sharpe standard error at finite samples
Politis, D. & Romano, J. (1994) "The Stationary Bootstrap" — Block resampling for serially dependent data
Pardo, R. (2008) "The Evaluation and Optimization of Trading Strategies" — Walk-forward, regime-based optimization

06Acceptable Period Ranges
Range TypeDaysUse Case
Minimum Acceptable940 days (2.6 years)Quick tests, high-frequency strategies
Recommended1,447 days (4.0 years)Calculated optimal — three-method average
Maximum Useful2,098 days (5.7 years)Conservative analysis, lower-frequency strategies
07 Re-optimisation Frequency
Recommended
Semi-annual ↓
(low vol avg)
Derived from optimal period
The re-optimisation frequency is derived from the recommended period and your timeframe's typical regime change rate. Re-running optimisation semi-annual ↓ using a rolling 1,447 days (4.0 years) window keeps strategy parameters aligned with current market conditions.
REGIME CHANGES DETECTED
18
Multiple shifts — walk-forward testing advised
VOLATILITY REGIME
0.32×
Low vol — extend lookback
08How to Apply These Results

💡 Practical Recommendations

1
Set Your Backtest Period
When running backtests, use approximately 1,447 days (4.0 years) of historical data. This gives enough trades for statistical validity while focusing on current market conditions.
2
Re-optimise Semi-annual ↓ (low vol)
Markets change over time. Re-run optimisation semi-annual ↓ (low vol) using the most recent 1,447 days (4.0 years) window to ensure parameters stay relevant to current market conditions.
3
Use Walk-Forward Testing
For robust validation: optimise on 1,447 days (4.0 years), test on the next 477 days (1.3 years), then roll forward and repeat. This prevents overfitting and validates parameter stability across time.
4
Use 95% DD (6.7%) as Your Risk Budget
Monte Carlo simulation projects that in the worst 5% of scenarios, maximum drawdown reaches 6.7%. Size your position so this drawdown is within your risk tolerance before allocating capital.