A Trading Journal That Survives a Bad Month

Almost everyone who trades has started a journal. Far fewer have one that is still being written six months later, and almost nobody has one that was still being written through their worst stretch. That is the interesting part. Journals do not usually die in the quiet months when nothing is happening — they die in the middle of a losing run, which is precisely the period they exist to explain.

The usual diagnosis is a discipline failure: you should have kept writing, you did not, try harder next time. That diagnosis is comfortable and it is mostly wrong. A journal that stops during a drawdown is usually a journal that was built in a way that made it unbearable to write during a drawdown. The fix is in the design, not in the resolve.

What a journal is actually for

Start with what it is not for. It is not a record of your results — your broker already keeps that one, it is more accurate than yours, and it takes no effort to produce. Anything you write down that duplicates the account statement is work that buys nothing.

The thing the statement cannot tell you is why you were in the trade. Two trades with identical outcomes can be a rule followed and a rule broken; two losses can be a good decision that lost and a bad decision that lost. Nothing downstream of the entry can separate those, because by then the outcome has arrived and it colours everything. A journal is the only place that separation can be recorded, and it can only be recorded at the time.

So the whole purpose narrows to one sentence: a journal exists to preserve what you thought before you knew how it turned out. Every design decision below follows from that, including the ones that look like laziness.

Why the bad month is the one that breaks it

Three things happen at once during a losing run, and each of them attacks a different part of the habit.

The first is that writing becomes unpleasant. A journal built around outcomes turns into a list of losses you are asking yourself to transcribe by hand, one at a time, at the end of a day that already went badly. Nobody sustains that, and calling it weakness does not make it sustainable.

The second is that the entries quietly change shape. They get longer. They acquire context and explanation — the news was unusual, the spread was wide, the session was thin. Some of that will be true. But the shift from short factual entries to long explanatory ones is itself the most reliable signal in the whole document, and it is invisible unless the earlier entries were short enough for the contrast to show.

The third is that the rules start moving. Somewhere in the middle of a drawdown, the size comes down, or the stop gets wider, or a setup that was never on the list starts appearing. Sometimes that is a sensible adaptation. The problem is that without a written record of what the rules were last month, an adaptation and a drift look exactly the same from the inside — and by the following month, you will not be able to tell which one you did.

Design rule one: the entry is written before the outcome

This is the rule that does most of the work, and it is the one most journals get backwards. If the entry is written after the trade closes, everything in it has been contaminated by knowing the answer. A losing trade gets recorded with the reservations you now have; a winner gets recorded with a confidence you did not feel at the time. Both are sincere and both are useless.

Write the line at entry, and write it before the position is open if you can. It needs almost nothing in it: the instrument, the setup name, where the trade is wrong, and what you expect to happen. Four short pieces. It should take less time than placing the order.

There is a useful side effect. A setup you cannot name in one line is usually not a setup — it is a feeling with a chart attached. Being unable to write the entry is information, and it arrives while you can still act on it.

Design rule two: short enough to survive a bad day

The length of an entry is not a matter of thoroughness. It is the single variable that determines whether the journal is still being written in month four. A format that takes two minutes on a good day takes twenty on a bad one, because on a bad day you will want to explain yourself.

So cap it deliberately. One line before the trade, one line after it closes. The after-line is not an essay: it says whether the rule was followed, and nothing else. Followed and lost is a perfectly good outcome to record. Broken and won is the entry that will be most valuable to you later, and it is the one an outcome-focused journal will never contain, because nothing about a winning day prompts you to write anything down.

If you want to write more, write it somewhere else. Keeping a separate place for the long thinking is not a compromise; it is what protects the short record from becoming an argument.

Design rule three: record the decision, not the feeling

Journals that ask you to log your emotional state sound rigorous and tend to collapse into noise. The problem is that the vocabulary is thin — after a few weeks everything is 'impatient' or 'confident' — and that the entries are unfalsifiable. There is no later test that can tell you whether 'anxious' was the right word.

The behavioural version of the same question is answerable. Did you take the trade before the level was reached? Did you move the stop after entry? Did you take this one because the last one lost? Those are facts about what you did, they can be checked against the chart afterwards, and they capture the state you were in far more reliably than a mood label does. Emotions matter — they just leave better fingerprints in actions than in adjectives.

Design rule four: nothing is deleted

The temptation to tidy up a journal is strongest in exactly the month you should not. An entry that reads badly in hindsight, a trade you would rather not have on the record, a week where the rule column says 'broken' four times — these are the entries with the highest information content in the entire document, and they are the first ones to be quietly dropped.

We take the same view about our own public record, for the same reason: a channel that deletes its losing calls is not producing evidence, it is producing an advertisement. A journal you have curated is doing the same thing to an audience of one. If you find yourself wanting to remove an entry, that impulse is worth a line of its own — note it and keep the entry.

Reading it: not daily, and not forwards

Writing a journal and reading one are separate habits, and most people only ever do the first. An unread journal is a diary. What turns it into a tool is a fixed review, at a cadence far slower than the writing.

Once a month is enough, and the review has a specific shape. Do not read it as a story from the start of the month to the end — read it looking for disagreement between two columns. The trades where the rule was followed and the result was bad. The trades where the rule was broken and the result was good. Those two groups are where everything you can actually learn lives, and both of them are invisible if you sort by profit and loss.

Then one further pass, and it is the one that catches the drift: compare the entries at the start of the month with the entries at the end. Are they the same length? Are they naming the same setups? Did the stop distances change? A rule that moved without a decision will show up as a difference in the writing long before it shows up in the results.

A format that fits on one line

The specific columns matter less than the fact that they are few and fixed. This is a version that has survived contact with bad months; a spreadsheet holds it perfectly well and so does a paper notebook.

• Date and instrument. Written at entry, not filled in later from the statement.
• Setup name, from a short list you wrote in advance. If the trade does not fit one of the names, that is the finding — record it as 'unnamed' rather than inventing a category to make it fit.
• Where it is wrong. The price or condition that would tell you the idea has failed. This is written before the trade and never edited afterwards.
• What you expect. One clause. Not a target price you will feel obliged to defend — what you think the instrument is about to do.
• Rule followed: yes or no. Filled in when the trade closes, and nothing else goes in this column. No explanations, no 'yes, but'.
• Outcome: win, loss, breakeven, or did not fill. Two seconds from the platform, and it is here only so you can sort against the rule column.

That is the whole thing. Notice what is missing: no screenshot, no profit figure, no lesson-learned field. Screenshots record what the market did, which you can always reconstruct; the profit figure is in your statement; and a lesson-learned field written the same day is nearly always the outcome talking. The lessons come from the monthly read, not from the entry.

What it will actually tell you

Expect the first useful finding to be smaller and more boring than you hoped. It is usually something like: one of your named setups accounts for most of the broken-rule entries, or the trades taken in the hour after a loss look different from the rest, or half of your entries are 'unnamed'. None of that is dramatic. All of it is actionable, and none of it was visible from the account statement.

And expect the bad month to be the one that pays for the whole exercise. A drawdown with a journal behind it is a set of questions with answers — the rules were followed and the market was unkind, or they were not followed and you can see where it started. A drawdown without one is just a number going down, and the story you tell yourself about it will be written afterwards by the number.

Common questions

How many trades before a journal tells me anything?

Refusing to give a number, because any number would be invented and the honest answer is not a count. What you are waiting for is not a sample size, it is a disagreement — the first cluster of trades where the rule column and the outcome column point in opposite directions. That can appear in a fortnight if the trades are frequent, or take several months if they are not. The one thing that guarantees it never appears is a journal that only records outcomes, which will simply agree with your statement forever.

Should I journal every trade, or only the interesting ones?

Every one, and this is the rule least worth compromising on. The moment you are choosing which trades are interesting enough to record, you are selecting the sample after the fact, and you will select on outcome without meaning to. It is also self-defeating in a practical way: the entries that look boring at the time — the ordinary trades taken by the rules — are the baseline that makes the unusual ones legible. If the format is short enough, logging all of them costs almost nothing, which is the other reason the length rule matters.

Spreadsheet, dedicated app, or notebook?

Whichever you will still open on a bad day. The dedicated tools do one genuinely useful thing — they import the fills, so the mechanical columns fill themselves — and one unhelpful thing, which is that they present a wall of performance analytics that pulls attention straight back to outcomes. A spreadsheet does nothing for you and asks nothing of you. Paper is slower but has the highest survival rate for a lot of people, probably because it makes the shortness feel normal. This is genuinely a matter of what you will keep doing; the format above works in all three.

My journal shows I break my rules constantly. Should I change the rules?

Maybe, but not during the month in which you noticed. A rule that is broken constantly is either wrong or unenforceable, and both are worth fixing — the trap is that a drawdown is the worst possible moment to tell the difference, because the pressure to change something is at its highest and the evidence is at its thinnest. Write down the proposed change and the date, keep trading the existing rule until the next monthly review, then decide with the entries in front of you. A change made on a specific date and recorded is a decision you can evaluate later. A change that happened gradually is the drift this whole piece is about.

Is there any point journalling if I am on a demo account?

Yes, with one caveat worth naming. The mechanics all transfer — the setup names, the rule column, the monthly read — and building the habit while nothing is at stake is easier than building it later. The caveat is that the part a journal is best at capturing, what you did under pressure, is the part a demo account does not produce. Expect the rule-followed column to look implausibly clean, and expect that to change when the money is real. That change, when it arrives, is itself the most useful thing your first live month will show you.