Back-testing and Walk-forward testing

Unlike backtesting, which evaluates a system against historical data it has already seen, walk-forward testing requires the system to be tested going forward in time. Results are gathered either from held-out historical data or from new data points as they become available, providing genuine out-of-sample validation.

This is essential for systems that cannot be meaningfully backtested in the traditional sense, including LLM-based trading systems, strategies that rely on natural-language processing, and any approach driven by sentiment analysis. These systems learn and adapt, so the only honest test is one where the data has never influenced the model.

Language-agnostic by design

It does not matter how your trading system is written. The Trading System Test Suite can be used with any programming language. You supply a list of trades and the library does the rest, so you keep full freedom in how you express your strategy.

The workflow is deliberately simple: build a collection of trades, each with a timestamp, a signal type, and a price, pass it to the test engine along with a last price, and read the results object back. From there your application can print a full report, log every trade, and drive further analysis.

Two dozen risk measurements

The Trading System Test Suite turns a trade list into a complete profile of profitability and risk, so you can judge a system on more than its headline P&L. It computes the metrics professional developers rely on to compare strategies and decide what is worth trading.

• Sharpe, Calmar, Sortino, and Sterling ratios
• Maximum drawdown and Monte Carlo drawdown
• Total P&L, largest profit, and largest loss
• Risk-to-reward ratio and percent profitable
• Standard deviation and downside deviation
• Value Added Monthly Index (VAMI)
• Average trades per month and full trade logs
• Profitable versus losing trade counts

Built for strategy optimization

Every good strategy eventually meets a losing stretch as liquidity, volatility, and market dynamics shift, which means systems must be optimized over time or retired. The Trading System Test Suite gives you the objective measurements to know which is which, summarizing the profitability and risk of a system against the exact data it was tested on.

Because results come back as a structured object, optimization loops are straightforward: sweep parameters, run the test for each combination, and compare the metrics that matter to your mandate. Validate before going live, then keep validating as the market moves.

For advanced developers

The Trading System Test Suite was written for engineers who want their testing in code, not buried in a black-box application. It pairs cleanly with the rest of the Modulus toolchain, author strategies in ModScript, compute signals with TA-SDK, and visualize results with StockChartX, and ships with complete source code so you can audit and extend it.

What developers get

A focused, code-first testing engine that quantifies strategy performance the way professional system developers expect.