How to Keep a Trading Journal That Improves Your Decisions

Put this into practice

Everything below is the reasoning. To actually keep the journal, try the free Trading Journal tool → — it logs trades, applies the fixed tag vocabulary from this guide, and calculates win rate, R-multiples, and expectancy automatically. No account, nothing sent to a server.

What a Trading Journal Is Actually For

A trading journal creates evidence about behavior that memory does not preserve. Its purpose is diagnostic: to establish whether losses are coming from the strategy, from execution, from risk sizing, or from discipline — four separate problems with four different fixes.

That distinction is the entire value. A trader who changes strategy when the real problem was oversizing, or tightens risk when the real problem was a method without an edge, is applying a correct fix to the wrong fault and will conclude that nothing works. Without a record of what was planned versus what was done, there is no way to tell which of the four is in play.

What Memory Keeps and What It Discards

Recollection of a trading month is not a sample of that month. It is a highlight reel, assembled by salience.

What traders reliably remember: the dramatic win, the painful loss, the trade they almost took that would have worked, and the moments the market appeared to prove them right. These are vivid, emotionally weighted, and easy to retrieve — which is exactly why they dominate any informal review.

What gets discarded: the repeated small rule violations that never produced a memorable outcome, the average quality of the setups actually taken, how often the stop was moved, how frequently a chased entry ended badly, and the cumulative cost of impulsive decisions that were individually too small to notice. None of these are dramatic. All of them are countable, and in aggregate they are usually where the money went.

A journal reverses the weighting. The forgettable, repeated behaviors become rows in a table, and the one memorable trade becomes what it actually was: a single observation.

What to Record

Fields fall into three groups by when they are captured. The before-trade group is the one most often skipped and the one that makes everything else interpretable, because it is the only record of what the plan was.

Before the trade

• Date and time — including time of day, which often turns out to matter.
• Symbol and asset class.
• Setup name — from your written list. A trade with no nameable setup is itself a finding.
• Direction — long or short.
• Market context — the broader trend, volatility conditions, and anything scheduled.
• Entry plan — the intended price or range, and the trigger.
• Stop — the specific price, and the structural reason for it.
• Target — or the trailing method, if there is no fixed target.
• Position size — units or contracts, and notional value.
• Maximum dollar risk — the planned loss if the stop fills as expected. This becomes the R for the trade.
• Emotional state — recorded as a short observable note, not a diagnosis: rushed, calm, distracted, keen to make back yesterday.
• Screenshot — the chart as it looked at the decision, before the outcome existed.

During the trade

• Actual entry — the fill, next to the planned price, so slippage is visible.
• Order type used, and whether it was the type the plan called for.
• Adjustments — anything changed after the position opened.
• Adds or reductions — size changes, with the reason and whether the plan permitted them.
• Changes to stop or target — recorded as events, with the original level preserved. This single field surfaces more behavior than any other.
• Emotional changes — again as observable notes: checking the position every few seconds, switching timeframes repeatedly.
• Unexpected events — news, a halt, a spread widening, a platform problem.

After the trade

• Exit — price, time, and whether it was the planned exit.
• Profit or loss in dollars, net of costs.
• Result in units of initial risk — the R-multiple, explained in the next section.
• Maximum favorable excursion — the best unrealized profit the position reached.
• Maximum adverse excursion — the worst unrealized loss it reached before the exit.
• Rule adherence — scored independently of the dollar result.
• Mistake tags — from the fixed vocabulary below.
• Screenshot — the completed chart, kept alongside the pre-trade one rather than instead of it.
• Lesson — one sentence, and only if there genuinely is one.
• Corrective action — a specific change to a rule or workflow, or nothing. Be more patient is not a corrective action.

Maximum favorable and adverse excursion deserve a note. Together they answer questions that the exit price alone cannot: whether targets are consistently set beyond where price actually goes, and whether stops sit just inside the normal noise of the instrument. Both are sizing and placement problems rather than strategy problems, and neither is visible without the two fields.

Behavioral Tags

Use a fixed tag vocabulary — a closed list, chosen in advance, applied to every trade. A workable starting set:

• FOMO — entered because the move was already underway.
• Revenge — entered to recover a prior loss.
• Boredom — entered without criteria during a quiet stretch.
• Overconfidence — sized up or skipped steps after a good run.
• Fearful exit — closed early, before the plan called for it, with no rule change to justify it.
• Moved stop — the stop was changed after entry.
• Oversized — position larger than the risk rule permits.
• Unplanned add — size increased without a written add condition.
• Valid loss — the plan was followed and the trade lost. Not a mistake.
• Good skip — a setup correctly declined.
• Late entry — the trigger had already passed.
• Poor liquidity — spread or depth outside the stated limit.
• Followed plan — full adherence, whatever the result.

The reason a fixed vocabulary beats free-form prose is arithmetic: it makes counting possible. Eleven moved stop tags in a month is a fact you can act on. Eleven paragraphs each describing, in slightly different words, a moment of hesitation about a stop is a set of stories, and no total can be taken from it. Tags convert behavior into a frequency, and frequency is what reveals a pattern.

Free-form notes still have a place. They carry the context a tag cannot — what the market was doing, what the reasoning was, what made this instance unusual. Keep them as an additional field, not as a replacement for the countable one.

Result in Units of Risk (R)

R is the initial planned risk on the trade — the dollar amount you accepted losing if the stop filled as expected. The result expressed in units of R is the R-multiple. If the planned risk was $250, then a $500 gain is +2R and a $250 loss is −1R.

Recording results this way solves a comparison problem. Dollar results mix together position size, share price and instrument, so a $600 gain on a large position and a $150 gain on a small one cannot be ranked by quality. R-multiples strip that out and put every trade — different sizes, different prices, different markets — on a single scale that measures how much was earned per unit of risk taken.

Hypothetical example — for education only.

Five trades, each with its own planned risk:

Two things are visible here that the dollar column alone hides. Trade 4 is the largest dollar loss at $400, but in risk terms it is identical to Trade 2 at −1.0R — both stopped out exactly where planned, and neither is a behavioral problem. And Trade 5 made real money in dollars while returning only half the risk taken, which is a sizing and target question rather than a win to celebrate. The account gained $220; the process returned +2.0R across five trades.

One caveat: R-multiples are only meaningful if the planned risk was recorded before the trade and the stop was actually honored. A trade where the stop was widened has no clean R, and forcing one onto it hides precisely the behavior worth measuring.

Logging Good Skips

A trade correctly declined is a decision, and decisions are what a journal records. Most journals contain only executed trades, which quietly biases the whole record: every correct pass disappears from the log and survives only in memory, where it gets stored not as a rule working but as money left on the table.

Log the skip with the setup name, the specific rule that disqualified it, and what happened afterward. Two useful things come out of that. First, a correct pass becomes a countable event that the review can credit, rather than a source of regret — this is directly relevant to FOMO trading, since the feeling of having missed something is fed almost entirely by unrecorded passes. Second, if a large share of skipped setups would have reached their targets, that is evidence the filter is too tight — a finding that is impossible to reach if skips are never written down.

Recording every setup you glanced at is not the goal. The ones worth logging are those that met your criteria, or came close enough that a decision had to be made.

Reviewing Without Rationalizing

The review question is: what decision was justified by the information available at the time? It is not: what would have made the most money. Those two questions produce different answers on most trades, and only the first one generates a lesson that transfers to the next trade.

Reviewing against the outcome instead of against the information is how a journal becomes an engine for bad conclusions. A rule-following loss gets recorded as a mistake and the rule gets weakened. A rule-breaking win gets recorded as a good read and the violation gets repeated at larger size. Do this for a few months and the journal will have taught the opposite of what it was kept for.

Hindsight makes past outcomes look considerably more predictable than they were — the completed chart contains the answer, the live chart did not. The practical defenses are procedural: read the pre-trade note and pre-trade screenshot before looking at the result, record the adherence score before the dollar column is visible, and leave the original note unedited so it cannot be quietly reinterpreted. Cognitive biases in trading covers hindsight and outcome bias in more depth, including why the two need different corrections.

Journal Formats

Three formats cover almost everyone. A plain spreadsheet, with one row per trade and one column per field, which sorts and counts without any setup work. A written template — the same fields as headings in a document, one entry per trade — which suits longer-horizon trading and richer notes but makes counting harder. Or a dedicated tool, which can import fills and compute statistics automatically, at the cost of accepting someone else's field structure.

The format matters far less than two things: consistent fields, and actually reviewing it. A spreadsheet with twelve reliable columns will outperform an elaborate tool that gets filled in for half the trades. And a journal nobody reviews is a diary — a complete record that changes no future decision.

Swoopr's authenticated dashboard includes manual trade tracking for signed-in users, which can serve as the record for trades placed outside an automated strategy; you can sign in to see what it holds. Whether that or a spreadsheet is the right home for your journal depends entirely on which one you will keep current.

Common Mistakes

• Recording only losses. A log of bad outcomes cannot show what a well-executed trade looks like, and it makes the win rate and the average result impossible to compute.
• Recording outcomes but not the plan. Without the pre-trade entry, stop, target and size, there is nothing to compare the execution against — so no trade can ever be scored for adherence.
• Free-form notes with no countable fields. Pages of prose describe individual trades well and support no totals at all. Without a fixed tag list, a repeated error stays invisible.
• Never reviewing. Data collection is the cheap half. The review on a fixed schedule is where the finding comes from, and it is the step that gets dropped first.
• Changing the strategy after every entry. A single trade is one observation. Adjusting rules trade by trade guarantees that no version of the method ever accumulates a sample worth judging.

Limitations

A journal does not guarantee profitability. It is a measurement instrument, not an edge, and it cannot fix a strategy with negative expectancy — a well-documented losing method is still a losing method, and the documentation may even make it easier to keep operating. It is also only as honest as the person filling it in: an adherence score set after the result is known, or a mistake tag omitted because the trade happened to profit, corrupts the record in the exact place its value sits.

Small samples reveal very little. Ten trades cannot distinguish a working method from a lucky one, and patterns that look clear across a handful of entries frequently vanish across a hundred. Journaling reduces avoidable, repeated errors in a process; it does not remove market risk, and no field or review schedule prevents losses.

Trading Journal FAQs

Related Guides

• Trading performance metrics — the statistics a consistently filled journal makes computable.
• Trading discipline and the pre-trade checklist — the rules an adherence score is scored against.
• Cognitive biases in trading — why reviews drift toward rationalizing, and what stops it.
• Trading psychology guide — the full framework this page is part of.