You closed eleven trades last month and you can describe three of them. The other eight are rows in a table you have not read since you logged them. A generated review starts from all eleven, because it reads the table rather than your memory, and that is the whole of its advantage.
What the review reads and what it prints
The Reviews tab inside AI Insights holds this. Pick a period from the row of buttons, Weekly, Monthly, Quarterly, 6 Months or Yearly, and press Generate Review. Nothing runs on a schedule. There is one review per completed period, and once it exists the button changes to show when the next one becomes available.
What goes in is every trade you closed inside that period. For each one: the ticker as TICKER.EXCHANGE, whether it was long or short, the realised profit or loss, the R-multiple, and the strategy attached to it. Alongside those, four period totals: trades closed, win rate, total P&L and average R-multiple. The itemised list is capped at 20 trades. The four totals are computed over every trade in the period, so a busy month still gets accurate figures on the card.
What comes back is a card. Collapsed, it shows the date range, the number of trades closed, and those three figures: P&L, Win Rate, Avg R. Expanded, it adds:
- Summary. A short read of the period.
- Best Trade and Worst Trade. Each shown as
TICKER.EXCHANGEwith its realised P&L, and each one clicks through to that trade. - Strengths and Areas to Improve. Bulleted, drawn from the period's numbers.
- Focus for Next Week (or next month, quarter, half or year, matching the period you chose).
- Action Items. Short chips, one line each.
You can delete a review. The figures on the card come from your trade records, so they are checkable against the ledger.
What it can see, and what it cannot
Be precise about this, because it decides whether a finding is worth acting on. The review reads the fields listed above and nothing else. Outside its input: the time of day you entered, your notes, your emotional state, the sector of the stock, the overnight gaps you sat through, and how far a trade ran before you exited it.
So a review can find that your short trades carry the losses, that one strategy holds every negative R-multiple, or that the worst trade was several times the size of the average loss. It cannot find that you trade badly on Thursday afternoons, because nothing in its input says what time it was.
That matters because a confident sentence about a pattern the data cannot contain is worse than no sentence. When you read a finding, check it against the fields the review was given.
Where the other patterns live
The patterns the review cannot see are not missing from the app. They are on the Analytics page, in the Trading Patterns section, which needs at least 5 closed trades before it renders, and in the Risk & Exit Analysis section above it, which needs 3.
Win Rate by Day of Week breaks your results down by the day. That is the card that answers a Thursday question.
Sector P&L lists each sector with its trade count, win rate and average R, so a sector you keep losing in shows up as a row rather than a feeling.
P&L by Setup Type compares your strategies. Market Regime compares your results across the four market regimes the app stamps on each trade at entry: Uptrend, Downtrend, Ranging and Volatile. Performance by Setup × Regime crosses the two, which is where you find out whether a strategy that works overall is one strategy working in a single regime.
Exit Capture puts one bar per trade showing how much of the peak favourable move you captured before exiting, or how much of the drawdown you absorbed. That is the card that answers whether your exits are early relative to how far the trade ran.
Overnight Gaps, on the Market tab of any trade, counts the gaps you held through and splits them into favorable, adverse and flat for the direction you were positioned in.
Read the review for the summary of a period. Read these cards for the mechanism.
A 15-minute review routine
Weekly, after the market closes for the week.
- Generate the review for the period. Read the Summary, then Areas to Improve. Take one item, not all of them.
- Open the Worst Trade from the card. It is one click. Check the entry against the strategy rules you attached to it, and write what you would need to see to skip that trade next time.
- Open one Analytics card that bears on the item you took. If the review flagged a strategy, that is P&L by Setup Type and the Setup × Regime crosstab. If it flagged exits, that is Exit Capture.
- Write the change as a rule, or write nothing. A change you cannot state as a rule is a mood, and it will not survive Monday.
The step that gets skipped is the fourth one, and it is the only step that changes a future trade.
Where these reviews go wrong
Small samples. Four trades can only produce a win rate of 0%, 25%, 50%, 75% or 100%, and nothing in between. The app takes this seriously enough to dim any cell in the Setup × Regime crosstab holding fewer than 5 trades, so a two-trade square cannot be mistaken for an edge. Apply the same discipline to a weekly review: over one week, the review is a record, and the pattern claims need a quarter behind them.
Reading a finding past its evidence. The review writes prose from numbers. Prose can imply causes that the numbers do not carry. If a sentence says you exit winners early, check it against Exit Capture, where the arithmetic is visible.
Fitting the last month. A rule written to remove last month's losses will fit last month exactly. Before you adopt one, count how many trades in your history it would have changed, and whether the change survives across regimes rather than in the one you have just been trading.
Set the review against your own records
The review is generated when you press the button, from the trades you closed in the period. Its value depends on those records being complete, so the strategy attached to each trade and the stop that sets its R-multiple do more for the output than anything else you can control.
Close out this week, open the Reviews tab in Swingfolio, and generate the weekly review. Then open Areas to Improve, take one line, and check it against the Analytics card that holds the underlying numbers. Start the 30-day trial if you have not set up an account yet.