Write a strategy in plain English. Ethy tests it on real market history, shows every trade, then runs the same rules on paper or live.
AI backtesting is testing a trading strategy on historical data by describing it in plain words: the AI turns your sentence into precise rules, replays them on past prices and reports every trade. Ethy’s backtesting software does this for crypto, stocks and gold, and the rules it tests are the same ones that trade when you switch it on.
A backtest is a simulation of how a rule would have done, not a forecast of how it will do. Its real job is to rule out weak ideas before real money meets them, so a good result earns a check on unseen data and a run on paper.
Type “long HYPE when the 1h RSI falls under 25, exit above 55, stop at 4%” and Ethy shows the rules it understood before it tests a single candle. No code, no scripts.
Test any perp on Hyperliquid, from BTC to the stocks, indices, gold and currencies on trade.xyz, or spot tokens. Coins Ethy can’t trade still test as research.
Each entry and exit is marked, and the report sets the return against buy and hold, with max drawdown, win rate, profit factor, Sharpe and a month-by-month view.
It tunes thresholds, indicator lengths, stops and targets on one stretch of history, then tests each candidate on a held-out stretch it never saw. Your position size stays yours.
Three steps from an idea in plain English to a strategy you can run.
Explain the rule as you would to another trader: the market, the timeframe, when to get in and out, and where the stop goes. Ethy turns it into precise rules and shows them back to you.
The backtesting tool replays the rule on up to a year of real market history and marks each entry and exit. Check the return against buy and hold, the worst drawdown and how often it traded.
Run the same rules in paper mode on live prices with no money moving, then switch them to live when you are ready. Nothing is rewritten between the backtest and the trade.
Tap one to open Ethy with it typed in, then swap the market or the numbers.
Not sure how to word a rule? Read how to backtest a trading strategy, with a real example.
Each of these can backtest trading strategies. They differ in what you have to bring.
| Question | Ethy | Charting platforms | Your own script | ChatGPT or Claude on their own |
|---|---|---|---|---|
| Writing the rule | Plain English, read back to you as rules | Built-in strategies, or code in the platform’s own language | Code you write and debug | A prompt, and code the model writes |
| Price history | Built in for Hyperliquid perps and spot tokens | Built in for the markets they list | You find, clean and store it | Files you upload or paste |
| Seeing each trade | Every entry and exit marked on the chart | A trade list and marks on the chart | Whatever you plot | A trade list, if you ask for one |
| Out-of-sample check | The optimizer scores a held-out window | You hold dates back by hand | You code the split | Only if you ask, and then you check it |
| After the test | Paper mode, then live, on the same rules | Depends on the platform | Execution code you host | You place the trades yourself |
Thinking of asking a chatbot instead? Read what ChatGPT and Claude get right about backtests, and where they invent results.
What traders ask about testing a strategy with AI.
Yes, AI can backtest a trading strategy when it runs real rules on real price data, and that condition matters, because a chatbot can also describe a backtest it never ran. Ethy turns your plain-English rule into precise rules, shows them to you, tests them on market data it supplies and marks every trade. Our guide to AI backtesting with ChatGPT and Claude covers the difference.
The best backtesting software is the one that tests the rule you will actually trade, on data from the market you will trade it on, and shows you every trade. Charting platforms suit chart traders, Python libraries suit coders who want full control, and Ethy suits traders who want to describe a rule in plain English and run that same rule afterward.
Yes, free backtesting is possible. A spreadsheet or a free Python library costs nothing but your time, and Ethy is free to start with no card needed, backtests included. Paid plans add more; see the pricing page for what each one includes.
Backtesting is accurate about the past only when its data and assumptions are realistic, and even a careful backtest is a simulation, not a forecast. Treat it as a filter: it is good at showing that a rule loses money, rarely trades or falls further than you could stand. A strong result is a reason to test on unseen data and on paper, not proof.
Yes. Crypto backtesting in Ethy covers every perp on Hyperliquid and spot tokens, with up to a year of history depending on the timeframe. The same tool tests the stocks, indices, gold and currencies on trade.xyz, and coins Ethy can’t trade can still be tested as research.
No. To backtest a trading strategy in Ethy, you explain the idea as you would to another trader, Ethy writes the rules and shows them back, and you can change any of them before the test runs. Ready-made templates from the strategy library are there if you would rather start from one.
Test it on data it was not tuned on. Ethy’s optimizer ranks settings on a training window, then checks each candidate on a held-out window it never saw while tuning, and a result that holds up there is stronger evidence than one that only fits the training stretch. Paper trading on live prices is the next check.
Yes. Templates from Ethy’s strategy library go through the same backtester as your own rules, and a basket of markets can be tested in one run, with a breakdown per market. Alerts, scanners, research reports and rules that follow smart money or an AI chart read are the exception: they can’t be backtested, so try them in paper mode.
Trading involves risk, including the loss of your money. Ethy AI is a research and trading tool and does not give investment advice or recommendations. Backtests are simulations on historical data and do not predict future results.
Start free with no card, and test your first idea today.
