Archive / Take apart an impressive trading-bot result without placing a trade

Take apart an impressive trading-bot result without placing a trade

Make an evidence checklist for a performance claim without installing a bot or connecting an account.

Nothing here is financial advice or a recommendation to buy, sell or trade anything. Do your own research.

For: Solo founders, Small business owners

Topics: Investment, Learning

Find out how the number was generated before deciding what it means. A backtest, paper simulation and live-money result answer different questions.

1. Make a claim card

Prompt

Using [public description and supplied reports], identify test type, dates, strategy version, starting capital, cash flows, benchmark and costs. Mark gaps as not stated. Do not install, run, sign into accounts or give trading advice.

2. Check the comparison

Prompt

Review [claim card and benchmark]. Compare period, return basis, fees, dividends, cash flows and risk exposure. Label which trading frictions the simulation excludes if documented. Do not calculate missing values from guesses.

3. Request better evidence

Prompt

Separate documented, claimed but unverified, and missing items. Pick one artifact, such as a timestamped equity series or benchmark calculation, and say what it would verify. Do not suggest trades, connect accounts or deploy anything.

Try it on one example

If a report says one month but does not provide the dates or benchmark method, you cannot fairly compare results yet. Code in a repo does not prove which run produced the reported number.

Inspired by the AlphaLab repository: https://github.com/cve-zh00/alphalab. See https://docs.alpaca.markets/us/docs/paper-trading for simulator limits. No project result is repeated or endorsed.

Use it for: Make an evidence checklist for a performance claim without installing a bot or connecting an account.

Back to the archive