Define before you optimise
Write down when to enter, when to exit, how much to risk and when to stand aside. If you cannot state it in a sentence, you cannot test it.
You do not need to write code, connect a broker or risk money to learn the workflow. Start with one rule, watch what it would have done, and only then decide whether an algorithm is worth the next step.
Most people start by choosing a strategy. That is the last step, not the first. These are the things worth understanding before any of it.
Write down when to enter, when to exit, how much to risk and when to stand aside. If you cannot state it in a sentence, you cannot test it.
Brokerage, spread, premium, impact and slippage decide whether an edge exists. A backtest that ignores costs is not a backtest of the real thing.
Win rate means little on its own. Drawdown, trade count and the trade log tell you whether a rule is real or just lucky over a short window.
One index, one pair, one underlying. Breadth makes it harder to tell whether a rule works or whether you got lucky.
The AI copilot turns an idea into an editable draft with entry, exit, invalidation and a starting risk policy. You edit it before it runs.
Not live prices. The point is to see the shape of the result and where it breaks, not to prove anything about the future.
Vary the risk budget, the cost assumptions or the period. If the result only survives one setting, you have not found a rule, you have found a coincidence.
The things worth being straight about before you decide anything.
Not for the guided path. The workspace is built around editable rules and plain-language explanations. If you would rather bring your own source code, the JSON and Python export is there for that.
No. The product preview needs no login, connects to no broker and submits no orders. It runs on generated sandbox data.
Running your rule against simulated fills with no execution venue attached. It is the step between an idea and risking real money, and it is where most surprises show up.
There is no honest number, and anyone giving you one is selling something. The question is whether you can explain why the rule works, what it costs, and what would make you stop.
No login, no broker connection, no orders. Open the workspace and run a rule on generated data.
Risk and regulation. This page describes a research workflow, not investment advice. Nothing here is a recommendation to trade any instrument, and past or simulated results do not indicate future performance.