Tools
How to choose backtesting software for systematic trading
By Koen van Goch · Updated
The best backtesting software is the one that fits what you trade, can test a whole portfolio realistically, works with clean data, and that you'll actually keep using. For portfolio-level stock systems, dedicated tools like QuantumX, RealTest and AmiBroker are popular. NinjaTrader 8 and TradeStation suit futures, Trading Blox is a long-standing choice for multi-market trend following (mainly futures), and MetaTrader 5 is built for forex and CFDs. Python gives full flexibility at the cost of more work.
What to look for in backtesting software
- Portfolio-level testing: can it test many stocks at once with shared capital, position limits and ranking? Many retail tools only test one chart at a time.
- Data quality: for stocks you need survivorship-bias-free data, which includes delisted companies, and correct adjustments for splits and dividends.
- Realistic costs and fills: commissions, slippage, limit orders, next-bar execution.
- Speed: fast tests make it practical to check robustness across many parameter values.
- Path to live trading: can the same rules generate orders for your broker?
- Learning curve: how much code you're willing to write.
Common options at a glance
| Tool | Best suited to | Coding | Notes |
|---|---|---|---|
| QuantumX | Portfolio-level stock and ETF systems | No code | Fast, and the easiest way I know to build and combine many systems at portfolio level |
| RealTest | Portfolio-level stock, ETF and crypto systems | Simple scripting | Very fast, built for testing several strategies together; scripting makes it more work than QuantumX; Windows only |
| AmiBroker | Stocks, futures, broad research | AFL scripting | Mature, flexible, large user base |
| NinjaTrader 8 | Futures, intraday, prop-firm accounts | C# (NinjaScript) | Strong for execution and automation of futures strategies |
| TradeStation | Futures and stocks | EasyLanguage | Broker and platform in one |
| MetaTrader 5 | Forex, CFDs, prop-firm accounts | MQL5 | Widely supported by CFD brokers and prop firms |
| Python (e.g. pandas, vectorbt, backtrader) | Custom research, crypto, anything | Full programming | Maximum flexibility; easiest place to introduce bugs and biases |
For prop-firm trading, any platform that lets you backtest the CFDs or futures your prop firm offers can work.
Data matters as much as software
A good backtester with bad data gives confident wrong answers. For US stocks, check that your data includes delisted stocks and historical index membership, otherwise your tests only include the winners that survived. For US stocks I recommend Norgate Data, which is what I use myself. Budget for a data subscription as well as software.
How to choose
Start from your markets and holding period. End-of-day stock portfolios point toward a dedicated portfolio backtester such as QuantumX or RealTest. Futures point toward NinjaTrader or TradeStation. Forex and CFD accounts point toward MT5. Python makes sense if you already code or need something the others can't do.
Then pick the one you'll stick with. Switching platforms every few months costs more than any feature difference. Once you've chosen, how to build your own trading system walks through the process.