Backtest trading strategies honestly
One engine for crypto, stocks and forex. Costs come out of every result, buying and holding sits beside it, and the luck checks tell you when one trade is doing all the work. No coding required.
Three Markets, One Engine
The market decides where the candles come from, never how the trades are simulated.
Crypto
Every Binance USDT spot pair, on 5-minute to daily candles. Perpetual-futures plans pay the funding charged at each 8-hour settlement.
Stocks & ETFs
Daily and weekly bars for listed tickers, international listings included, and US intraday bars going back to 2016.
Forex
28 major and cross pairs, with intraday history back to 2004. Same engine, same metrics, so results compare across markets.
Know Before You Trade
Most backtests flatter the strategy. These are built to find out whether it would really have made money.
Costs in every result
Fees are charged on both sides of every trade. Stops that gap fill at the open, not at your stop price, and when a bar hits both your stop and your target, the stop counts first.
Compared with doing nothing
Each run sits next to buying and holding the same asset over the same window, with both drawdowns shown. A strategy that trails holding says so.
Luck checks
See the result without its single best trade, and reshuffle the trade order to find the drawdown you should expect, not just the one history happened to give you.
The numbers that matter
Win rate, profit factor, Sharpe ratio, max drawdown and an equity curve, plus an account simulation with fixed, compounding or risk-based position sizing.
Trade-by-trade report
Every trade the strategy would have taken, with entry, exit and result, and a CSV export when you want the raw data.
No code
Set entry and exit rules, take-profit and stop-loss in a form, or describe the idea in plain words and let Sensei, the AI coach, write the rules.
Start From Strategies That Already Passed
Every entry in the strategy library went through the same six checks: profitable without its best trade, positive in two independent multi-year windows, working on more than one instrument, net of fees, a survivable drawdown, and better than 400 runs of random entry timing. Where an entry misses a check, its page says so.
The ideas that failed are published too, with their numbers, in the research log. Adopt a library strategy on the free plan, run the backtest yourself, and get an email when it signals.
What's Included
Backtesting Questions
Is the backtesting free?
Partly. A free account, with no card required, gets 10 backtests a month on strategies you adopt from the strategy library. The Analyst plan runs unlimited backtests on every market and adds Sensei for building your own strategies. See pricing.
Which markets and timeframes can I test?
Crypto on 5-minute to daily candles, stocks and ETFs on 5-minute to weekly bars (US intraday history starts in 2016), and 28 forex pairs on 5-minute to weekly bars, with intraday history back to 2004.
Do the results include trading costs?
Yes. Fees come out of every run, on both sides of each trade. Plans built with Sensei can also model slippage and fill at the next bar's open instead of the signal bar's close, and plans on perpetual futures pay funding at every settlement. The next-open fill matters: our own RSI(2) dip-buy gives up a third to half of its edge that way, and its write-up shows it.
Do I need to know how to code?
No. Rules go into a form of entry and exit conditions, or you describe the strategy to Sensei in plain words and it builds the rules, runs the backtest and suggests what to change.
How do I know a good backtest isn't just luck?
Start with the robustness checks on every result: if the profit disappears without the best trade, there was no strategy. For the library we go further and test every candidate against 400 runs of random entry timing. We wrote up what that test catches, and our guide covers the mistakes that make most backtests worthless.
Test Your Strategy
Stop guessing whether your strategy works. Find out what it would have done, after costs.