Reading a Trading Bot's Record

The Trading bots page runs paper strategies side by side. Each starts with the same simulated $100,000, sees the same live signals and trades by its own written rules. You can watch them, copy one into your paper account or build your own. None of it is real money, which is exactly why it is a good place to learn how records mislead.

A bot's return is never shown alone: beside it are its number of closed trades and its worst drop (max drawdown). The trade count says how much luck could be in the return; the drawdown says what you would have lived through to get it. A 30% return with a 40% drop is a much rougher ride than 15% with a 5% drop.

With many bots running for a few weeks, the one at the top is mostly the one that met the friendliest stretch of market. It is the many-tries trap from the previous lesson, played out live. That is why a bot isn't ranked until it has closed 30 trades, and why the default order is return per unit of drawdown: sorting by raw return puts first whoever took the most risk while risk happened to pay.

Check how many trades it has closed (and how many it still needs before its record means much), its worst drop, whether open positions could swing the figure, and its rules. Could you explain in two sentences why its edge should exist? A bot whose rules you can't state is one you wouldn't hold through a losing streak.

Copying a bot mirrors what it does NEXT, scaled to the share of your paper account you choose; you never get the record already on its page. And a record, however long, describes the market it ran in. Use it to decide what to study, not what to expect.