Evening review
Neutral
Weekly review: the register is up, and three of the four winners broke the rules
This is a review week: no new ideas, only an audit of what the register already contains.
The numbers first. Five ideas published, four closed and scored: two reached the first target, one ran to the second, one was stopped out. Total +6.6R, about +7.7% on the simulated 1,000 of paper capital at 1% risk per idea, 75% win rate, maximum drawdown 1R. A fifth idea on BTC/USD was voided and left out of the statistics; that instrument has since been removed from the universe, because on most European retail CFD accounts the short side is neither realistically executable nor closable at the weekend.
Those numbers look good, and they should not be trusted. Of the four scored trades, three were marked bad_process in their post-mortems: the checklist was broken and the market paid anyway. The single trade that followed the process cleanly is the one that lost. A register that rewards its own errors is the most dangerous kind, because it teaches the wrong habit fast and the right one slowly.
Four closed trades is not a sample. No pattern here, by instrument, by session slot or by conviction score, is statistically meaningful, and none will be for several weeks.
What changes next week: nothing in the rules, because no error has three recurrences behind it yet. The two existing lessons stay in force and will be verified before publishing rather than after. Informational material, not investment advice.
The numbers first. Five ideas published, four closed and scored: two reached the first target, one ran to the second, one was stopped out. Total +6.6R, about +7.7% on the simulated 1,000 of paper capital at 1% risk per idea, 75% win rate, maximum drawdown 1R. A fifth idea on BTC/USD was voided and left out of the statistics; that instrument has since been removed from the universe, because on most European retail CFD accounts the short side is neither realistically executable nor closable at the weekend.
Those numbers look good, and they should not be trusted. Of the four scored trades, three were marked bad_process in their post-mortems: the checklist was broken and the market paid anyway. The single trade that followed the process cleanly is the one that lost. A register that rewards its own errors is the most dangerous kind, because it teaches the wrong habit fast and the right one slowly.
Four closed trades is not a sample. No pattern here, by instrument, by session slot or by conviction score, is statistically meaningful, and none will be for several weeks.
What changes next week: nothing in the rules, because no error has three recurrences behind it yet. The two existing lessons stay in force and will be verified before publishing rather than after. Informational material, not investment advice.
Generated with AI assistance (claude-opus-5) and reviewed against a fixed checklist.