Journaling that actually changes behavior

You kept a journal for three weeks. The entries said things like "got chopped up, need more patience" and "should have waited for confirmation". Then you stopped, and nothing about your trading was different.

That is the normal outcome, and it is not a discipline problem. It is a design problem. Free-text feelings cannot be counted, and a record you cannot count cannot tell you anything you did not already suspect.

The difference between a diary and an instrument

A diary records what happened. An instrument produces a number you can compare across time.

"Felt impatient today" is a diary entry. "Trade opened 4 minutes after a losing exit, risk 1.4%, setup score 2 of 5" is an instrument reading. The second one joins fifty others and becomes a distribution. The first one joins fifty others and becomes a mood.

The test for a field is simple: could you sort by it. If you cannot sort or count it, it belongs in a notes box at the end, not in the structure.

The fields that earn their place

Keep it short enough that filling it in takes under a minute, or you will stop. Ten fields is plenty.

Setup name. One of a fixed list you defined. Not free text. This is the field that eventually tells you which of your setups actually pays.

Score at entry, 0 to 5. Written before the outcome exists, against your written conditions.

Planned risk and actual risk, in percent of equity. Two fields, not one. The gap between them is where overconfidence and revenge live.

Planned stop and actual exit price. Again two fields. Exits taken away from the plan are the single most informative thing in most journals.

Result in R, not currency. Currency amounts change meaning as the account grows. R does not, and it is the unit every useful comparison is made in.

Minutes since previous exit. One number, and it exposes revenge trading and overtrading at once.

Time of day opened. Reveals fatigue clustering.

State tag, from a fixed list. Rational, FOMO, revenge, fatigue, overconfidence, overtrading. Fixed list, chosen at entry, one value. This is the same taxonomy Indikora's Coach applies automatically to trades on a read-only connected account, and the reason to use a closed list is that closed lists can be counted while free text cannot.

One sentence of notes. Last, optional, and never the main event.

The review is the product

Writing entries does nothing. The review is where behavior changes, and it needs a fixed shape or it becomes rereading.

Weekly, 20 minutes, same time. Not after a bad day, when the conclusion is predetermined.

Ask four counting questions. How many trades had actual risk above planned risk. How many exits happened away from the written plan. Which state tag has the worst average R. What is the average R of trades opened within 15 minutes of a loss, against everything else.

Then change exactly one rule. One. A review that produces five changes produces none, because you cannot attribute the next month's results to any of them. Write the change as a sentence with a number in it: "no new order within 45 minutes of a losing exit", not "be more patient".

And write the date next to it. Six weeks later you can compare the metric before and after. That is the whole loop, and it is the only part that makes journaling worth the time.

Why it fails, specifically

Too many fields. Twenty-five columns is a form you will abandon by week three. Ten you can live with.

Filled in after the outcome. Once you know the result, your recorded reasoning is reconstructed. The score and the state tag have to be entered at entry or they are worthless.

Only losses get written up. Winners contain your worst habits, because a bad decision that paid is a habit you are not examining. Log everything.

No review. This is the most common failure by a distance. People collect months of data and never sort it. An unsorted journal has produced zero information.

Indikora's journal attaches notes and tags to closed trades so the fields stay joined to the fill, and the Coach's classification gives you one of the countable columns without you having to be honest about yourself in real time. Neither replaces the review. The review is the part that costs 20 minutes a week and is the reason any of it works.

Three months, one metric

If you want the smallest version that still works: log setup, score, planned risk, actual risk, result in R, and minutes since last exit. Review Sunday. Change one rule. Repeat.

After three months you will have something most traders never get: a specific, counted description of how you actually behave, rather than a memory of how you meant to. That is what makes a rule change land, because you are no longer arguing with an impression.