Forex backtesting software
The spread is the strategy. Most forex backtests get it wrong by assuming it away.
The finding
14 of the 20 platforms assessed here list forex. 11 support it natively. The other 3 qualify it: you bring the data, or it depends on the broker you connect, or the support stops short of a full backtest.
That distinction is the page. A list of asset classes counts all 14 the same way, which is how people end up choosing a tool and discovering the gap afterwards.
Who ships forex, and how
| Platform | Support | Rigour |
|---|---|---|
| QuantConnect | Native | 6.8 |
| QuantRocket | Native | 6.8 |
| NautilusTrader | Native | 5.8 |
| StrategyQuant X | Native | 5.8 |
| Wealth-Lab | Native | 5.8 |
| MetaTrader 5 | Native | 5.3 |
| AmiBroker | Native | 4.5 |
| Forex Tester | Native | 4.3 |
| TradingView | Native | 4.3 |
| Backtrader | Nativeforex (spot) | 3.8 |
| TrendSpider | Native | 3.8 |
| NinjaTrader | Broker-dependentforex (data/backtest via third-party feeds) | 5.5 |
| VectorBT | Bring your own data | 5.3 |
| Backtesting.py | Bring your own data | 3.5 |
Levels come from a structured field on every platform, normalised from each vendor's own wording, with the original qualifier kept and shown above.
Why the gap is so wide here
Forex has no central exchange, so there is no single price. Your broker quotes its own, the spread widens around news and at rollover, and a backtest run at the mid on a fixed spread will show an edge that does not survive contact with a real quote.
A forex backtest is only as honest as its spread model. A fixed spread over years of history flatters every high-frequency strategy and most low-frequency ones.
Support is not trustworthiness
Reading the two columns together is the point. Whether a platform can load forex is a data question. Whether the number it hands back means anything is a different one, and it is the one almost nobody checks: only 2 of 20 platforms correct a result for how many variants you tried. A tool can have perfect forex data and still hand you a number that is mostly search luck, which is what overfitting is.
Frequently asked questions
Which backtesting platforms support forex natively?
11 of the 20 assessed here: QuantConnect, QuantRocket, NautilusTrader, StrategyQuant X, Wealth-Lab, MetaTrader 5, AmiBroker, Forex Tester, TradingView, Backtrader, TrendSpider. A further 3 list forex but qualify it, either because you supply the data yourself or because it depends on the broker you connect.
Why do so many tools list forex without really supporting it?
Forex has no central exchange, so there is no single price. Your broker quotes its own, the spread widens around news and at rollover, and a backtest run at the mid on a fixed spread will show an edge that does not survive contact with a real quote.
What goes wrong when I supply my own forex data?
A forex backtest is only as honest as its spread model. A fixed spread over years of history flatters every high-frequency strategy and most low-frequency ones.
Does native support mean the backtest is trustworthy?
No. Support is about data and plumbing. Whether the number is trustworthy depends on whether the engine controls look-ahead, models fills honestly, and corrects for how many variants you tried, which is what the rigour column measures and where most platforms score badly. Check both.
Related
Educational content only, not investment advice.