1. Treating a progression as protection
Increasing the stake after a loss increases exposure; it does not repair the probability of the next contract. A progression can reach account or platform limits quickly during a losing sequence.
2. Using no absolute stop
A bot should have a clear condition that prevents new purchases after a loss limit, profit target, run count, or manual stop. Check that restart logic cannot bypass that condition.
3. Ignoring the disconnected state
A browser, network, or platform connection can close while a contract is open. The interface should not assume that a missing update means the trade did not occur. Check the broker-side report before restarting.
4. Testing too many changes at once
When market, duration, stake, entry rule, and money management all change together, you cannot identify which change caused the outcome. Test one meaningful change at a time.
5. Confusing a recent pattern with certainty
A streak or digit imbalance in a rolling window is an observation, not a guarantee. Building an unlimited recovery sequence around that assumption can magnify losses.
6. Moving to real funds too quickly
A few successful demo runs are not evidence of stable performance. Test operational behaviour across different sessions and keep real-account limits materially lower than an amount that could affect your finances.
7. Not reviewing the journal
Read rejected requests, proposal values, purchases, settlements, and stop events. The journal is where you confirm that the bot followed the rule you intended rather than the rule you accidentally built.