Trading journal

A log kept alongside trading activity. The mechanical half is already produced by any platform: instrument, direction, size, entry and exit prices, times and costs, all recoverable from the account history. The half a journal adds is the part no platform records: the reason a position was entered, the condition that would have shown the reasoning to be wrong, the state the writer was in, and what was expected to happen. The convention that distinguishes a journal from a report is timing, since an entry written before the outcome is known is a record and an entry written after it is a reconstruction.

What it is conventionally read for is patterns across many entries rather than the story of any one. Repeated grouping of losses in one instrument, one session or one condition; a difference between the plan as written and the position as taken; a tendency for results to change after a run of either kind. None of that is visible in a single record, and all of it is arithmetic over a sufficient number of them, which is why the sample size is the thing that makes the exercise worth anything.

Whether keeping one improves results is genuinely unsettled. The psychological literature finds that being aware of a bias does little to remove it, and the stronger evidence sits with structural measures such as conditions fixed in advance and exits set before entry rather than with reflection after the fact. What a journal does reliably is create a record that cannot be quietly rewritten, which is a narrower claim than the one usually made for it, and the honest reason confirmation bias is discussed alongside it.