Hindsight Bias in Trading Journals: The Same Trade, Written Twice

By Sofia Harchich | Trading Psychologist & Behavioural Finance Writer

Hindsight bias in trading journals is the reason a diligent record can run for eight months and change nothing.

The entries are honest. The trader is not lying to the page, and would pass a polygraph on every line. What has happened is subtler and much harder to catch: the entry was composed by someone who already knew how the trade ended, and knowing the ending reorganised everything that came before it.

What follows is a reconstruction — a composite, assembled from patterns that recur across behavioural records rather than from any single person’s file. The trade is representative rather than real. The two versions of it are the point.

10:41 — what was true at the time

A long position, opened forty minutes earlier on a setup that met three written criteria. Size as planned. Stop under the swing.

At 10:41 the price is nine points below entry. The trader is looking at a five-minute chart, and the following things are true simultaneously: the higher timeframe structure is intact, the volume on the pullback is unremarkable, a scheduled release is due in nineteen minutes, and a similar setup two weeks ago went badly.

None of these is decisive. The trader is holding four partial considerations and no conclusion, which is the ordinary condition of being in a trade. There is mild discomfort, an awareness of the clock, and a half-formed intention to reduce before the release that does not get acted on.

At 10:47 the price recovers. By 11:15 the position is up twenty-two points and closed at target.

17:20 — what the journal says

That evening the entry gets written. Here it is, in the form it usually takes:

Held through the pullback rather than panicking. Higher timeframe structure was clearly intact and the volume confirmed it was just a shakeout. Stayed patient and let the setup work. Good example of not letting short-term noise shake me out of a valid idea.

Every clause in that paragraph is defensible. Not one of them is what happened.

The four competing considerations have become one clear reading. The half-formed intention to reduce has vanished entirely. “Stayed patient” has been applied retrospectively to a period the trader spent uncertain and slightly uncomfortable. And a decision that was, in truth, not taken has been recorded as a decision that was taken, deliberately, on the basis of evidence.

Read back in a month, that entry teaches a lesson: hold through pullbacks when structure is intact. The trader has no way of knowing the lesson was manufactured after the fact.

Distortion one: the reason arrives after the outcome

The first and largest distortion is causal. Once an outcome is known, the mind supplies a reason for it, and the reason feels like recall rather than construction.

Hindsight bias is the tendency to look back at an event that was genuinely unpredictable and experience it as having been obvious. It is one of the more robust findings in the literature, and it does not weaken with expertise or with warning — knowing about it does not stop it happening, which is what makes hindsight bias in trading journals so difficult to write around.

In the reconstruction, the volume observation was one of four things noticed and was not weighted heavily at the time. In the record it has become the reason for the decision. That promotion happens silently.

Distortion two: the alternative disappears

The second distortion is the one that costs the most, and almost nobody notices it, because you cannot see an absence.

At 10:41 the trader nearly reduced before the release. That intention was real, it was reasonable, and on a different day it would have been the better call. It appears nowhere in the entry — because it did not happen, and because the outcome made it look unnecessary.

A record that only contains what was done cannot support a review. Decision quality lives in the comparison between what you did and what you nearly did, and the near-miss is exactly what the evening write-up deletes.

Distortion three: the timing collapses

The third is quieter and it destroys the most useful data in the record. It is also the distortion that makes hindsight bias in trading journals invisible even to a careful reader of their own file.

In the reconstruction, discomfort peaked at around 10:44 and the position was closed at 11:15. Thirty-one minutes separate them. The journal entry compresses that into a single narrative beat — held through the pullback and let it work — with no interval in it at all.

Timing is precisely what memory does not keep. The gap between a signal appearing and an action being taken is the single most informative number a discretionary trader can hold, and end-of-day reconstruction cannot produce it. It has to be captured live or not at all.

The evening entry is not neutral in the instability loop. It rewrites the outcome into a story, and the story becomes the reinforced belief at step five — the part of the loop that decides how the next trigger will be seen.

Why hindsight bias in trading journals is worse than keeping no record

Worth stating plainly, because it sounds like an overstatement and it is not.

A trader with no journal knows they do not know. They hold their conclusions loosely and remain open to being wrong about themselves.

A trader with eight months of distorted entries has something that looks like evidence. Every entry is sincere, they accumulate into a coherent account, and the account is wrong in a consistent direction — because each entry was written by someone who already knew the answer. Confidence rises with the page count while accuracy does not, and decisions get made on the basis of a file that was never measuring what it appeared to measure.

A record assembled after the outcome is a well-organised memory. Memory is the thing the record was supposed to replace.

The weekend review compounds it. Reading a run of entries produces a narrative, and narratives resolve — a pattern that would be obvious in a column of numbers becomes, in prose, a series of individually reasonable decisions. Confirmation bias then operates on the reading as freely as it did on the trade, and the lesson extracted from a month of entries is usually the one that fits a view already held.

What survives the distortion

Not everything in a written entry is compromised, and it is worth knowing what to keep.

Anything recorded before the outcome was known survives intact. Anything countable survives — number of entries outside written criteria, planned size against actual size, clock times. Anything checkable against the platform record survives.

What does not survive is every sentence containing a reason, an interpretation, or the word because. Those are written by the wrong person.

This is the practical resolution to hindsight bias in trading journals, and it is a change of format rather than a change of effort: shift the weight of the record from prose to fields, and from the evening to the session.

The two-entry method

The method takes about forty seconds a trade and it is designed around the distortions above.

Entry one, before the position opens. Three items: the written criteria that were met, the single observable event that would invalidate the idea, and one line naming the strongest alternative you considered and rejected. That last item is the antidote to distortion two, and it is the one traders skip.

Entry two, during the session, at the moment something happens. Two clock times: when the invalidation condition occurred, and when you acted on it. Four words each, no narrative.

Nothing is written after the close. If you want to write reflectively about your trading — and it is a reasonable thing to want — keep it in a separate document that is never used as the record. Separating the two is what stops hindsight bias in trading journals from contaminating the part of the file you intend to act on.

Ten trades of that produces something the evening write-up cannot: a comparison between the alternative you named beforehand and the decision you made, plus a distribution of latencies. Both are immune to knowing the outcome, because both were fixed before the outcome existed.

What the weekend review was actually for

Fairness to the practice, because the advice to journal is not wrong and the people giving it are not careless.

Written review does something real. It builds a habit of attention, it slows the week down, and it creates a relationship with your own behaviour that most traders never develop. The trader keeping eight months of entries has done something difficult and has built genuine detection capacity — they notice more, earlier, than they did in month one. That is step one of the sequence and everything else installs on top of it.

The failure is one of aim rather than effort. Journalling in its standard form is a reflect-stage practice, sitting at step six, and it is being asked to diagnose failures that occur at steps one and two. A tool built for consolidation is being used as an instrument of measurement, and it is a poor instrument because it runs after the fact.

The conclusion the record was quietly building

One thing worth naming, because it is usually what makes this hard to hear.

A trader who has kept a journal faithfully for months and seen no behavioural change generally reaches one of two verdicts. Either the journalling was pointless, which leads to abandoning the record entirely — the worst available outcome. Or the fault is theirs: insufficiently honest, insufficiently rigorous, insufficiently disciplined in the writing.

The second verdict is the more damaging, because it prescribes more of the thing that is not working. The response is longer entries, more searching self-examination, more careful prose — and entries that get longer and better while the behaviour holds constant are a specific failure in themselves, where the description discharges the discomfort that would otherwise force a change.

Neither verdict is supported by what actually happened. What happened is that a sincere record was compiled by a person who knew the ending, three times over, in three specific ways that can each be designed around.

Where to start

Ten trades, two entries each, and no evening write-up at all.

Before the position: criteria met, invalidation condition, the alternative you rejected. During the session: two clock times. That is the whole method, and it costs less time than the entry you were writing at 17:20.

At the end of ten, compare the alternatives you named beforehand against what you actually did. Most traders find something in that column they have been unable to see for months, because it is the one part of the record that hindsight bias in trading journals cannot reach.

For the sorting framework this bias belongs to, read behavioural finance for traders; for what to put in the fields, trading journal psychology; for the bias that operates on the same record while the position is still open, confirmation bias in technical analysis.

The pattern quiz identifies which behaviour is running your trading under pressure, and the free sample it sends maps that result onto the four surface patterns — reactive, controlled, avoidant and impulsive. Either is a more reliable starting point than a reading of your own past entries.

---

About the Author

Sofia Harchich is a Trading Psychologist and Behavioural Finance Writer with a Master’s in Psychology. She works at the intersection of Jungian shadow work, neuroscience, and market behaviour — helping traders understand the psychology driving their decisions, not just the strategy.

Read more at thewealthmirror.com/about