Three ways that need no code: run a ready-made strategy with its free backtest, read the results of the textbook strategies, or describe your own idea and let the AI write and test it. Every run on real data, with costs charged.
The concept library explains each idea, from RSI and SuperTrend to VWAP and breakouts, with its Pine script written out. A strategy's page has a Free backtest button that runs it on real data with costs charged.
Open the concept library →Seventeen textbook strategies, run with their usual settings on SPY, QQQ, NVDA and others, on daily bars since 2016 and intraday bars for the day-trading ones, costs charged and the last third held out, with every trade listed. No account needed.
See the strategy results →Tell the Pine Copilot the entries, exits and stops in plain words. It writes the strategy in Pine v6, checks it and runs it here, so you see whether the idea did anything without writing a line.
How the AI writes it →Yes. Start from a ready-made strategy in the concept library and run it with its free backtest button, read the published results of the textbook strategies, or describe your own idea to the Pine Copilot, which writes the Pine and runs it.
The ready-made strategies and the published results are free: a free account backtests without limit on the Free plan, with no card, and the results pages need no account at all. The Pine Copilot, which writes a script from your own words, has 50 AI credits a day on the Free plan (a new script uses 15) and more on the paid plans, each with 30 days money back.
US stocks and ETFs, with daily bars from 2016 and intraday bars from one minute, and NQ, MNQ, ES and MES futures on CME's 1-minute bars for signed-in accounts.
That costs were charged, that there are enough trades to mean something, how deep the worst drawdown went, and how the strategy did on the part of the data it was not tuned on. A result without costs, or with a handful of trades, says little.
Alpha Charts is software for writing and testing trading scripts and gives no investment advice. A backtest shows what rules would have done on past data, not what they will do.