Search "best trading bot" and every result looks the same: a ranked list, a handful of screenshots, and confident claims about returns with no way to check any of it. That format isn't an accident — it's optimized to rank in search results and convert clicks, not to give you evidence.

None of that makes automated trading illegitimate. It means the burden is on you to know what real evidence looks like, so you can tell it apart from a well-designed page. Here are the five questions that actually matter.

1. Can you see every trade, or just the highlights?

What to look for
A full, timestamped trade log — every win and every loss — not a curated slideshow of the best weeks. If a site shows you three winning screenshots and a testimonial, that's marketing, not a track record.

2. What's the sample size, and over what time period?

What to look for
Ten trades over two good weeks proves almost nothing — it's well within the range of pure luck. A track record worth trusting needs enough trades, over enough time, to have lived through at least one bad stretch of market conditions.

3. Is the reported return net of fees and slippage?

What to look for
A backtest that ignores spread, commission, and slippage will always look better than live trading. Ask whether the numbers you're looking at are backtested or real, and whether real execution costs are already subtracted.

4. What does the strategy do in a loss, not just a win?

What to look for
Win rate alone is close to meaningless — a strategy that wins 90% of the time can still be a net loser if the 10% of losses are ten times the size of the average win. Look for profit factor (gross profit ÷ gross loss) and maximum drawdown, not just win percentage. Our guide to reading a track record covers exactly what each of these numbers means.

5. Is real money actually at risk, or is this a demo account?

What to look for
Neither answer is automatically disqualifying, but the claim needs to match reality. Demo-account results are real execution against real market prices, which is genuinely useful evidence — but only if the site tells you that's what you're looking at, instead of implying real capital is on the line when it isn't.
A quick sanity check: if a page can't answer these five questions without you having to dig for it, that's itself the answer. Real evidence doesn't need to hide.

What "best" should actually mean

There's no such thing as an objectively "best" trading bot independent of your own risk tolerance, capital, and time horizon — a strategy with a high win rate and small drawdowns suits a different person than one with a lower win rate but a much bigger average win. Anyone claiming a single universal "best" is selling you something, or hasn't looked closely enough at their own numbers.

What you actually want is a bot whose results you can verify, whose failures you can see as clearly as its wins, and whose numbers hold up over a real sample size and a real stretch of time — not the one with the shiniest landing page.

How we hold ourselves to this

We built our own performance page to answer these exact five questions without you having to ask: every closed trade across the fleet, win rate, profit factor, and a live cumulative equity curve, computed automatically from the real trade log — not curated. Every bot runs on an OANDA practice account while it builds a genuine track record, disclosed plainly, not implied otherwise. And our public engineering log documents the real bugs we've found, because a strategy's real weaknesses matter as much as its wins.