Backtesting

Overview

VECTRA's backtesting engine goes beyond basic historical replay. It provides institutional-grade analysis tools including Monte Carlo simulation for confidence intervals, Walk-Forward Analysis for robustness testing, and detailed trade-by-trade journaling.

Features Beyond Basic Backtesting

Feature	Basic Bots	VECTRA
Candle simulation	Single timeframe	Multi-Timeframe with look-ahead bias prevention
Execution model	Instant fill at close	Realistic slippage model + fee tiers + partial fills
Robustness testing	None	Walk-Forward Analysis (WFA)
Confidence intervals	None	Monte Carlo simulation (1000+ iterations)
Risk analytics	Basic win rate	Calmar, Sortino, Omega, Ulcer Index, Max Drawdown
Trade journal	Basic log	Full entry/exit reasoning, strategy attribution
Equity analysis	Simple curve	Decomposition: alpha, beta, volatility contribution
Breakdown	Global only	Per-strategy, per-symbol, per-session, per-regime
Reporting	Text output	Interactive HTML report with charts
Correlation	None	Inter-strategy correlation analysis

Walk-Forward Analysis (WFA)

1

Optimize (training window)

Perform parameter optimization on Window 1 (the training period). Find candidate parameter sets that maximize chosen objective metrics.

2

Test (out-of-sample window)

Validate the selected parameters on Window 2 (out-of-sample). Record performance and stability metrics.

3

Slide and repeat

Advance the windows forward and repeat the optimize/test cycle across the dataset to evaluate persistence and robustness of the strategy.

This process reveals whether the strategy's edge is persistent or an artifact of curve-fitting.

Monte Carlo Simulation

Randomizes the order and magnitude of historical trades to generate 1000+ possible equity curves. This provides:

• Confidence intervals (5th, 25th, 50th, 75th, 95th percentile outcomes)

• Worst-case drawdown estimates

• Risk of ruin probability

More on what Monte Carlo outputs

Monte Carlo outputs let you quantify distributional outcomes for returns and drawdowns, helping to understand tail risks and expected variability beyond a single historical curve.

Session-Aware Analysis

Performance broken down by trading session (Asian, London, NY) reveals when the strategy performs best and worst.

Regime-Aware Attribution

Performance attributed to market regime shows which conditions drive profitability and which should be avoided.