Home › Learn › Lesson 49

The trading journal — the six columns worth keeping

By TradingPrimer · Published 11 Sep 2026

Everyone is told to keep a journal. Almost nobody is told what makes one entry worth more than another, so people log seventeen fields for three weeks, learn nothing, and stop. This lesson takes the awkward objection seriously — a journal is written by the same mind that made the trade, so why believe it? — and follows it to the only honest answer: trust the lines that were written before you knew how it ended, and the lines the exchange wrote for you. That leaves six columns, and it is worth knowing exactly which one of them is soft.

The padlock is the whole idea: four of the six columns are sealed before you know how the trade ends, which is why they are the only ones your hindsight cannot rewrite.

What is a trading journal actually for?

It is a record of decisions, not of money. Your exchange already keeps a perfect record of the money — every fill, every fee, every timestamp, and it cannot be talked into rounding anything in your favour. What no exchange holds is the thing you believed when you clicked, and what you had already agreed to do if you turned out to be wrong. A journal exists to hold those two things, and it is only worth keeping to the extent that it does.

The course this site is built on places it as the last of six steps: analyse, plan, size and set the stop, manage, close — and then write the journal. The note attached to that final step is blunt: this is the step that makes us better every day and cuts down the mistakes we have already made. Not the step that tells you how much you made. Notice that the journal is listed as part of the trade, not as homework that happens afterwards.

There is a mechanical reason writing beats thinking it through. When you flip between timeframes weighing a trade in your head, whatever you are already hoping for quietly recruits evidence. The bullish reading gets stated; the bearish one gets felt for half a second and never finishes forming. On a chart that is genuinely balanced, that is enough to tip you every single time, and you will not notice because the missing half was never a sentence. Writing the case down forces both sides into words where they can be counted. This is also why the useful entry is the one written before you act: afterwards, the recruiting has already happened.

Which columns survive the fact that you wrote them yourself?

Only two kinds of line: the ones written before the outcome existed, and the ones the exchange wrote for you. Everything else in a journal is a memory, and memory has already been told how the story ends.

This is the awkward question the previous lesson left on the table. If a journal is filled in by the same mind that made the trade, why would it be any more honest than the mind was? The answer is not that you become more honest. It is that four of the six standard columns are sealed by the clock. You wrote them when you did not yet know whether you were about to be right, so there was nothing to flatter.

Here are the six columns from the handwritten template our course uses, sorted by the only two properties that decide whether a line can be trusted:

Five of six are hard. The sixth is the one the course calls the column that makes you better — and it is also the one your own hindsight can rewrite. Section six is about how to stop that, and the trick is not honesty. It is turning column six into a comparison with column four instead of a story.

How many logged trades before the journal can tell you anything?

Around fifty, if what you are trying to read is your own behaviour. Below that the honest answer is that your journal does not yet know anything, and reading conclusions out of it is just a slower way of guessing.

The number worth measuring first is not your win rate but your plan adherence — the share of trades where the exit matched what column four said before you entered. Lesson 3 makes the case for promoting yourself on adherence rather than on profit; this section asks the follow-up question nobody asks, which is how much evidence an adherence figure needs before it means anything.

Say you look back at your journal and count that you followed the plan on 80% of trades. Here is the 95% confidence interval around that figure, by sample size — the range of true adherence rates that could plausibly have produced what you counted:

Read the second row again, because it is the one that costs people a year. A twenty-line journal cannot distinguish a trader who follows the plan six times in ten from one who follows it nine times in ten. Those are completely different traders with completely different futures, and twenty trades put them in the same bucket. Most people quit journalling somewhere around trade fifteen, having decided it was not telling them anything — and they were right, and they stopped exactly one row before it started to.

Why does every extra column make you know less?

Because a column is not just something to fill in — it is a way of cutting the data, and every cut divides your year into smaller piles. The cost of the seventeenth field is not the eight seconds of typing. It is that each pile gets too small to read.

Suppose you take 200 trades in a year, which is an active year of swing trading, and you want to know which conditions your method works in. Each dimension you record multiplies the number of buckets you will eventually sort into:

By the third cut, a bucket showing a 40% win rate and a bucket showing a 60% win rate are statistically the same bucket. You will still read a difference into them, because the numbers are sitting there looking different, and you will change your trading on the strength of it.

So the working rule is uncomfortable and specific: one year of trading buys you one slicing dimension, at most two. Choose the dimension before the year starts — usually setup type, since it is the thing you can act on — and let the other twelve fields go. This is the real argument for six columns rather than sixteen, and it has nothing to do with laziness.

The time cost points the same way. Six columns is about ninety seconds a trade, which is five hours a year; a seventeen-field form is closer to four minutes, which is thirteen. Those eight extra hours are the ones people quietly stop paying, and what they buy instead is the problem in the next section.

What does a journal with gaps actually report?

A better year than the one you had. Not slightly better — in the worked example below, a journal missing 20% of the trades reports an edge 4.4× larger than the account delivered.

The reason is that the trades that go unlogged are not a random sample. Nobody forgets to record the trade they are proud of. The ones that slip are the impulsive ones, the oversized ones, the ones taken outside the written setup — which is to say, the exact trades the journal exists to catch.

Here is that priced out. The assumptions are stated so you can disagree with them: 200 trades in a year, 160 taken inside the plan (45% win rate, +2.0R on winners, −1.0R on losers) and 40 taken outside it (25% win rate, +2.0R on winners, but −2.0R on losers because the size was doubled or the stop was widened). Everything after that is arithmetic.

Two things fall out of that. First, 20% of the trades ate 71% of the gross profit — the year went from +56.0R to +16.0R because of forty entries. Second, and this one is derived rather than assumed: if the average loss on that rule-breaking group is worse than 2.53R, the whole year is zero or negative no matter how well the other 160 were executed. Solve 56R + 20R − 30x = 0 and that is where it lands. Doubling your size on a bad day does not need to happen often to be the whole story.

Which gives the one rule in this lesson that has no exceptions: log the trades you do not want to log, or do not bother. A complete journal of a bad year is useful. A flattering journal of the same year is worse than no journal, because it hands you confidence you have not earned.

How do you write the last column so it is worth keeping?

Answer a yes-or-no question first, then write the sentence. The question is always the same: did the exit match what column four said — yes or no? And if no, what rule replaced it?

That ordering is the whole fix. Left to itself, column six produces narrative — the market was strange, the news was unexpected, the setup was not quite clean — and narrative written after the close is exactly the kind of evidence the previous lesson warned you about. But a binary written against a line that was sealed before you entered cannot be edited, because column four is sitting right there and you cannot claim you planned the thing you did not plan. Once the answer is on the page, the sentence you add afterwards has to live with it.

This also gives you the adherence rate from section three for free: it is just the share of yes answers. If you use our free journal tool, the field labelled “Did you follow the plan?” is that binary, and “What did this trade teach you?” is the sentence that comes after it. Fill those two and the free-text box is doing the job the course intended. Skip the binary and it becomes a diary.

For what a good column six looks like when it has teeth, here is one from the course, written on 16 July 2023 after a futures account was liquidated. It does not describe the market at all. It lists four failures — overtrading, position size far too large, stop loss set far too wide, and wanting to win back what the stop had just taken — and then names the single cause underneath all four: wanting to get rich fast. Four behaviours, one root, no mention of price. That page is worth more than the twelve months of ledger around it.

Notice what it is not, too. It is not a record of feelings, and it is not a plan for the next trade. It is an audit of four decisions against what had been agreed before each one.

Practice corner

Everything above collapses into one habit you can start on your next trade: fill columns one to four before you click, and do not click until they are filled. If you cannot write the reason and the stop, you do not have a trade — you have an urge with a chart attached. Do it for fifty trades before you read anything into the results, and count the yes answers in column six as you go.

PRACTICE CORNER

The four columns written before entry are only real if the stop and target actually exist as orders. If you want to practise that on a small spot account — plan written first, stop placed at entry, every fill logged including the boring ones — these are the exchanges we use:

Binance→ OKX→ Bybit→

Affiliate links — we may earn a commission at no cost to you. Disclosure · Education only, not financial advice.

What people get wrong about trade journals

• “I’ll write it up at the end of the week.” By then every outcome is known, so all six columns become memory and the four that were supposed to be sealed are not sealed at all. A weekly write-up is a report, not a journal.
• Logging only the interesting trades. Section five prices this: the missing 20% took 71% of the year’s gross profit. The boring wins are the evidence that the process works when you leave it alone, and the ugly losses are the entire diagnosis.
• Recording the reason after the fact. “I entered because of the breakout” written on a trade that worked is not a reason, it is a caption. Only the dated version is evidence.
• Judging the journal by whether it is green. The P&L column is the one thing your exchange already tracks perfectly. The journal exists for the other five.
• Adding fields to make it more scientific. Seventeen fields feels rigorous and produces buckets of eight trades, which is where you start confidently acting on noise.
• Quitting at trade fifteen because it “isn’t telling me anything”. Correct diagnosis, wrong conclusion — at fifteen trades nothing could be telling you anything. The first readable number arrives around fifty.

When this lesson is wrong

Three conditions, and they are worth knowing before you take any of the above as a rule.

If you take very few trades — a position trader with twenty entries a year — the sample-size argument in section three never resolves, and no amount of discipline will fix that. Your journal is then a record for reasoning about individual trades, not a dataset, and you should read it that way rather than computing rates from it.

If the trades you fail to log are missing for ordinary reasons — you were travelling, you were busy — then the gap is noise rather than bias, and the 4.4× distortion in section five does not apply. The distortion depends entirely on the missing trades being the rule-breaking ones. That happens to be the usual case, but it is an assumption and not a law.

And if you have no edge at all, a perfect journal will document the loss without preventing it. The journal is a measuring instrument. It tells you whether you did what you said you would; it cannot make what you said you would do profitable. That part belongs to the risk arithmetic and to the strategy work in the stage that follows this one.

Frequently asked questions

Do I have to write the journal by hand, or is a spreadsheet fine?

Either works, because the property that matters is when the line is written, not what it is written on. A spreadsheet row filled in before you click buy does the job; a beautiful handwritten page filled in at the end of the week does not, because by then you know how every trade ended. If you are choosing, the only practical argument for writing by hand is speed of thought - a pen keeps up with a half-formed objection, and a form with dropdowns tends to flatten it into whichever option is closest. Pick whichever one you will still be doing in three months.

Lesson 48 said a feelings journal is the weaker tool. Isn't this lesson contradicting it?

No, it is the same argument carried one step further. The objection in that lesson is that a journal entry is written by the mind that made the decision, so it can be flattered after the fact. Four of the six columns here escape that objection by construction: they are dated before the outcome exists, so there is nothing to flatter yet. The two columns written after the close are exactly the ones the objection applies to, which is why this lesson turns the last one into a yes-or-no comparison against column four rather than a paragraph about how it felt.

How long should I keep a trade journal before I decide my strategy doesn't work?

Longer than feels reasonable, and the arithmetic is unkind about it. Around an observed rate of 80%, a 20-trade sample gives a 95% confidence interval of 58.4% to 91.9% - wide enough to contain both a trader who follows the plan six times in ten and one who follows it nine times in ten. At 50 trades the interval is 67.0% to 88.8%. So 50 logged trades is roughly the point where the journal starts to be evidence rather than mood, and judging a strategy needs more than judging adherence, because outcomes are noisier than behaviour.

What about screenshots of the chart - should I attach one to every trade?

Attach one if it costs you seconds, skip it if it costs you minutes. A screenshot taken at entry is genuinely useful because it is another thing recorded before the outcome exists. But the failure mode of journalling is abandonment, not insufficient detail: six columns is about five hours a year at ninety seconds a trade, and a seventeen-field form is about thirteen. The eight-hour difference is what people quietly stop paying, and a journal that stops has the gap problem in section five.

Sources and method. The six-column template, the six-step trade process and the journal page dated 16 July 2023 all come from our own ten-part course (Part 1, risk management and the trading journal; Part 4, the six-step strategy). Confidence intervals are Wilson 95% intervals, computed exactly. The gap scenario in section five states its assumptions inline — 200 trades, a 160/40 split, 45% and 25% win rates, and a doubled loss on the unlogged group; the 2.53R break-even threshold is derived from those, not assumed. Full working: de-cuong-bai/do-so-bai-49.py. Published 11 Sep 2026.

Next: The trading plan — the five steps in order, and why the journal you have just built is the thing that tells you whether you followed it.