How We Test Our Bots
What we test, how we test it, and the honest limits of what any test can tell you about future results.
Last updated October 28, 2026
What we test
Each Tradion bot runs a published, versioned strategy: a set of explicit rules (which markets, what qualifies a signal, how the entry is confirmed, when a setup is cancelled). Before a version is published, we check three things:
- Correctness: the bot does exactly what the rules say. Automated tests replay recorded Deriv ticks and compare every signal and trade decision with the expected result.
- Safety: stakes, session limits, authorization windows and account checks work, including when the connection drops or the market data is late.
- Behaviour: how often it signals, how long its losing streaks are, and how its results are distributed, measured on historical replays and on demo accounts.
How we test strategies
- Recorded ticks: we download real tick history from Deriv's public market data for each market the bot trades.
- Out-of-sample checks: rules are designed on one period of data and judged on different periods the rules never saw. A strategy that only looks good on the data it was designed on is overfitted.
- Comparison with chance: results are compared with the break-even rate implied by Deriv's payout and with simple baselines, using confidence intervals rather than single numbers.
- Demo forward tests: a new version runs on demo accounts in live conditions before it is offered for real trading.
What we have found
Deriv's synthetic indices are generated by a random number process. In our replays on recorded ticks, we have not found a strategy that consistently beats the break-even rate set by Deriv's payouts. That is why we do not publish win rates or profit claims for any bot.
What the bots do offer is consistent, rule-based execution: the same rules every time, with your stake and loss limits enforced and every trade recorded. Whether that suits you depends on why you trade and how much risk you accept.
Limits of any test
- Past results, including backtests, do not predict future results.
- Small samples are noisy. A few days of good results can be pure chance.
- Live trading adds delays, rejected contracts and price changes that a replay cannot fully reproduce.
- Deriv can change payouts and contract terms, which changes break-even rates.
Read Trading Bot Risk Management before trading with real money.