How to Review a Trading Month

The point of reviewing a month on a checklist is comparability. A review improvised from whatever feels salient will be thorough after a losing month and cursory after a winning one — which is backwards, because winning months hide problems just as well. Running the same six steps in the same orderproduces a written summary that lines up against last month’s, and the month after that.

The example running through the steps below is one month on a €20,000 account — May, with 38 closed trades — and the full summary is assembled at the end.

May closed at +€680 (+3.4%). That sounds like a result, but it is half of one. Mid-month, equity peaked at €20,710 and then fell to €19,620 before recovering — a €1,090 drawdown, 5.3% from the peak and 1.6 times the net gain. The honest description of May is “earned 3.4% while being down 5.3% at the worst point”, not “made 3.4%”.

Writing the two numbers side by side every month builds the comparison that matters: is the return being bought with more risk, less, or about the same? A month that doubles its P/L while tripling its drawdown has not improved. The Drawdown Calculator converts peak and trough into the percentage figure if your journal does not already show it.

The second step is activity, and the reference point is your own recent history, not a textbook figure. Over the previous six months this account averaged about 45 closed trades per month; May produced 38. The review question is simply why: a holiday-shortened week, an EA that found fewer setups, or sessions that were skipped? Each explanation is checkable against the trade dates, and each implies something different about whether June should look similar.

Win rate came in at 55% against a 51% baseline. At 38 trades, a four-point swing is unremarkable — two trades landing the other way would erase it. It goes in the summary as “normal”, not as evidence of improvement.

Step three compares the edge metrics with previous months. May’s gross profit was €2,520 against €1,680 of gross loss — a profit factor of 1.50, up from 1.22 in April. Average net result per trade (the journal’s expectancy estimate) rose from +€3.8 to +€17.9. Both moved the right way; neither is conclusive.

What each of these metrics measures — and what each hides on its own — is covered in the journal metrics guide; the monthly review only needs their direction of travel and the caveat above.

An account-level +€680 says nothing about where the money came from. Splitting the month answers that in two cuts:

• By symbol:EUR/USD +€612 over 17 trades, USD/JPY +€293 over 12, GBP/USD −€225 over 9. Two symbols carried the month; one drained it.
• By strategy:the trend EA (magic number 2103) made +€540 over 22 trades; manual trades added +€140 over 16.

The strategy cut only works if EA trades carry their magic numbers and manual trades are tagged consistently — the practical methods are covered in the guide to separating manual and EA trades. With the split in hand, GBP/USD stops being a vague unease and becomes a line item with a number attached.

Commission and swap are written into the account history: €114 of commission in May (38 trades averaging 0.5 lots at €6 per round-turn lot) and €46 of swap, most of it from one USD/JPY position held across a triple-swap Wednesday. Spread is the cost the history does not itemise, because it is embedded in the fill prices — so it gets estimated:

For May: 38 × 0.5 × 0.7 pips × €9 ≈ €120. Total costs come to roughly €280 — 11% of the €2,520 gross profit— a share worth tracking month over month, because it creeps up with trade frequency and holding time long before it shows up as a losing month.

Execution gets a quick scan rather than a metric: any fills far from their requested prices, anything filled inside a news window, gaps over weekends, oddly large swap lines. May’s scan found two fills slipped by 1.8 and 2.4 pips where the rest of the month stayed under 0.5 — both within minutes of the same scheduled release. That is an observation worth a line, even though it cost little this time.

The lessons entry is two or three sentences, each anchored to something in the data. “I was undisciplined this month” is a mood, unfalsifiable and useless next month. Compare May’s entries:

• GBP/USD finished negative for the third consecutive month (−€95, −€180, −€225) while carrying the widest spread of the symbols traded.
• The entire €1,090 drawdown formed in one week, when the EA and two manual trades were positioned the same way on correlated pairs at the same time.
• The only two fills that slipped more than 2 pips both sat inside the same news window; every other fill slipped under 0.5.

Each observation suggests a change — which is exactly why the review ends with a constraint. Change at most one thing per month. If June begins with GBP/USD paused, the correlation cap tightened and entries banned around releases, and June turns out better, nothing was learned: any of the three could be responsible, and one might have hurt while another helped. In this example the trader pauses GBP/USD and leaves everything else alone; the other two observations stay on the list as candidates for later months.

Assembled, the whole review fits in eight lines — short enough to write every month and to re-read a year later:

Every number above comes straight out of an account history plus five minutes of arithmetic — nothing requires forecasting or judgement until the final two lines. That is the property worth protecting: the review measures first and decides last, one change at a time.