What Should a Trading Journal Track? Fields That Matter and Fields That Don't

The test of a trading journal is not how much it contains. It is whether an analyst who has never met you could recompute your results from it. Most journals fail that test, which is why most journals produce feelings instead of findings.

The non-negotiable fields

Six fields make a trade reconstructable: symbol, direction, entry price, exit price, size, and fees, plus timestamps for open and close. With those, everything else is derivable: P&L can be recomputed independently, gross can be separated from net, holding time is known, and any claim in the journal can be checked against arithmetic.

Miss any one of them and whole classes of diagnosis become impossible. Without size, no dollar attribution. Without fees, no cost drag. Without timestamps, no session analysis and no way to join trades to market conditions. Screenshots and feelings do not substitute, because they cannot be summed.

Why recomputability is the whole point

A journal that stores only outcomes, plus commentary, can tell you that Tuesday went badly. It cannot tell you whether Tuesday was bad luck inside a working process or a fair sample of a broken one, because answering that requires slicing trades by objective conditions, and slicing requires raw fields.

It also cannot catch accounting errors, which are more common than anyone admits. If your journal's P&L is typed in rather than derived, it inherits every mistake upstream of it. The ten-minute check that has saved us more than once: sum exit minus entry times size across the journal and compare it against the reported total. Disagreement means stop.

Conditions worth tagging

Beyond the six fields, tag the things you intend to make decisions about. A setup label, because two strategies on one symbol must be separable or neither can be judged. The session, though this is derivable from timestamps and better computed than typed. Anything about market state you believe matters, volatility regime for instance, ideally derived from data rather than mood.

The discipline is that a tag is a hypothesis. Tagging setup means you intend to ask whether that setup earns its place. Untagged books can only ever be judged as a blend, and blends hide exactly the conditions you most need to see.

What not to bother with

Long emotional narratives, untethered to fields, age into noise. Rules you wrote but do not check trades against. Duplicate metrics a spreadsheet can derive. The journal is not a diary competition. It is evidence collection, and evidence has a schema.

Once the schema exists, the questions become answerable: which conditions bleed, what costs consume, whether the edge you believe in survives contact with trades it has not seen. That last one is the highest bar, and we document publicly what it looks like when a book is put through it.