Logging trades is the input. The weekly review is where a trading journal produces output, and it is the step most traders skip because nobody tells them what to do in it. "Look at your trades" is not a routine. This is one: twenty minutes, seven steps, the same order every week, with the specific questions that turn a pile of trades into one change for next week.
Why weekly, and not daily or monthly
A daily review reacts to noise; five trades cannot tell you anything, and a review after a losing day becomes self-criticism. A monthly review is too late; by the time you notice that a mistake cost $1,800, it has cost $1,800. A week is long enough to hold ten to thirty trades and short enough that the fix lands while the memory is fresh.
Pick a fixed slot: Saturday morning or Sunday evening, never Friday after the close. Put it in the calendar like a trade.
The 20 minute routine
| Minutes | Step | The question |
|---|---|---|
| 2 | Calendar | Which days made and lost money, and how many trades did each day have? |
| 3 | The numbers | Net result, R, win rate, profit factor, expectancy this week, next to the last four weeks. |
| 5 | Best and worst trade | Was the best trade skill or luck? Was the worst a bad setup or a bad decision? |
| 3 | Mistake tags | Which tag cost the most this week, in money? Is it the same tag as last week? |
| 2 | Discipline | What share of trades had a plan and no mistake tag? How long is the clean streak? |
| 3 | One change | One sentence, specific, for next week. |
| 2 | Write it | The review as five lines in the journal, so next week can compare. |
Twenty minutes is a ceiling, not a target. Weeks with six trades take ten.
Questions that produce answers
Most review advice asks "how do you feel about the week?" Feelings are what the journal is supposed to replace. Ask questions that have a number or a yes/no as an answer:
- Did I trade my setups? Share of trades with a strategy tag from the playbook, against trades tagged "no plan".
- Did the setups work? Expectancy per strategy tag this week, with the sample size next to it. Under five trades the answer is "no data", not "the setup is broken".
- Was the size right? Risk per trade in percent, every trade. Any trade over the limit gets the "oversized" tag now if it did not already have it.
- Where did the losses cluster? Day of week, hour, session. One slice usually carries most of them.
- Did I honour the stops? Count of trades where the loss exceeded 1R by more than slippage explains.
- Did I leave money on the table? Average exit efficiency, if the journal computes it from MAE and MFE. See MAE and MFE.
- Did the open positions get managed? Anything still open from more than a week ago needs a decision today.
The one change
The output of the review is one sentence that describes a behaviour, a condition and a limit. "Be more disciplined" fails all three tests. "No entries in the first ten minutes after the New York open, for the whole of next week" passes them. Write it at the top of next week's plan and put it in the pre-trade checklist if the journal has one.
Change one thing. Two changes cannot be evaluated; three will not be kept.
What not to do
- Do not change the strategy on one week of data. A week is for behaviour; a quarter is for strategy. Twenty trades is the minimum before a setup's numbers mean anything.
- Do not review after a bad Friday. You will not see patterns, you will see guilt. Sleep, then review.
- Do not skip weeks with few or no trades. A week with two trades still has a calendar, a discipline rate and a plan for next week. Skipping once is how the habit ends.
- Do not rewrite history. If a result looks wrong, fix the fill that produced it; never edit the number.
Turn the review into a report
If you have a mentor, a trading group or a prop firm that wants to see progress, the weekly review is the thing to send, not a screenshot of the account balance. A one-page report with the week's KPIs, the daily bars, the mistakes that cost money and your five written lines is credible precisely because it shows the losses and what you concluded from them.
Doing it in SetupBook
The dashboard's "This week" card shows the week's numbers and reminds you when last week's review or an evening note is missing. The Journal page holds day notes and the weekly review as a text you write; the calendar and the analytics breakdowns answer the questions above for any period, including a custom range. The weekly report dialog renders net, R, win rate, best and worst day, the mistakes that cost you and your review as a PNG you can download and send. The journal section of the user guide shows the flow with the demo account's weeks, and a free account is enough to start the routine this Sunday.