Kitson Strategy Translator + Backtest Lab: Plain English to a Tested Strategy (Opus 5.5)

Describe a strategy in plain English. Get rules you can read, a hypothetical backtest, and a critique of how far to trust it.

Strategy Translator + Backtest Lab works in four steps:

- Translate: Claude Opus 5.5 turns your description into strict rules from a small, safe grammar (moving averages, RSI, ATR, highs and lows, stops, take-profits, time limits), shown back as plain-English rule cards. Nothing is tested until you tick every card.

- Test: our own backtest engine runs the rules on your own historical price file (CSV) or the made-up sample, on your computer. No look-ahead (signals on a bar's close, fills at the next bar's open); commission, slippage and stops applied to every fill; a trade list, an equity chart, drawdown, win rate, profit factor and a buy-and-hold comparison on the same data.

- Stress-test: the same rules are run on each half of the file and with every lookback nudged 20% either way, to show how fragile the result is.

- Critique: Opus critiques the test, not the market: too few trades, one lucky trade, sensitivity, costs, one market period, and what to test next.

What makes it different

- Nothing the AI writes is ever run as code. The rules are read as data from a small grammar, and anything outside it is refused.

- You confirm every rule before anything is tested. The rule cards are written by the app, not the model.

- Every number in the critique is checked in code against the app's own figures, and the result of that check is shown to you.

- Your price file stays on your computer. Only a summary of it (dates, bar count and column names) and the app's figures are sent to Anthropic, and you can see exactly what will be sent before anything goes.

- It never tells you whether to trade a strategy. If the critique drifts into advice, forecasts or promises, the app catches it and asks for a rewrite.

- The cost is shown before every run, and you set a limit. In our own test on made-up data, about 19 cents on our own Anthropic account for the whole thing: the translation about 6 cents, the critique about 8 cents and one change about 5 cents. The backtest itself is free.

- A Word report and an Excel file of the rules, the results, the robustness checks, every trade and the equity at every bar.

What you need

- Free Node.js 18+ (Windows, Mac or Linux).

- Your own Anthropic account and API key, with a little credit. A Claude chat subscription does not include API access.

- Your own historical price file (date, open, high, low, close; volume optional), or use the made-up sample in the zip.

You pay Anthropic directly for what you use, on top of the price of the app.

Educational only. Hypothetical backtest on historical or made-up data. Past or simulated results do not predict future results. Not financial advice. It tests the rules you describe and does not recommend any trade or strategy. The app never places trades. Trading involves risk of loss. If we can't get the app running for you, email kit376351@gmail.com within 14 days for a refund of the app price. Kitson Workshop is not affiliated with or endorsed by Anthropic.