Monthly P&L

The Monthly view aggregates P/L into a year × month grid with annual totals, so seasonal patterns and clusters of bad months jump out at a glance.

The grid

• Rows are years; columns are the twelve months plus a yearly Total.
• Cells are the sum of daily P/L for that month, in account currency. Months with no trades read as zero.
• The grid is not cumulative — each cell stands on its own.

An insight line summarises the share of positive months and names the best and worst months.

Following the framed period

By default the grid covers the entire history of the loaded portfolio. Narrowing the workspace range alone does not change that: on its own the range selector is a chart zoom.

Turn on align starting capital in the workspace and the range becomes the portfolio's reading calendar: the grid then lists only the framed months, and a band above it names the active period with a full history button to widen back. Two details follow from that window:

• Edge months are partial. A window that starts on the 17th gives you a first cell covering the 17th onwards, not the whole month.
• The first framed day is the starting balance, not a move. It is the day the rebased curve starts flat on, so its P/L is absorbed into the base and left out of the cell — exactly as in the Metrics ledger. That is what makes the cells add up to the net profit the ledger reports for the same period, down to the cent.

The cells carry the same rebase as the curve and the Metrics ledger: the window is recomposed strategy by strategy, so a fixed-contract leg keeps its real P/L while a dynamically-sized one is scaled to the initial capital. A leg whose size is pinned by a max/min contracts limit counts as fixed — its P/L no longer follows the account. See align starting capital.

The variant tabs are clipped to the same window, so switching scenarios compares month for month. Opening a month keeps the window too: an edge month shows only the framed days, so the card never contradicts the cell that opened it. Exports are unaffected — the Monthly P&L export always ships the full history.

Month detail

Open any month to see it in isolation:

• Entry equity — the equity carried in from the end of the previous month (or the initial capital for the first month).
• Intra-month equity and drawdown — the daily curve within the month, with drawdown reseeded from the entry equity.
• Stats — month P/L, return %, win rate, trading days, best and worst day, and the month's max drawdown.

Economic events

more ▾, next to Table and Heatmap, opens Economic events: the portfolio's P/L on the sessions of each scheduled release (FOMC, CPI, the jobs report…) and of the market events (monthly opex, VIX expiration, month end…), against the other sessions. It answers the question to ask before stepping aside on those days: are they really different, or is it noise? It is the same analysis as the backtest results, on the scenario shown: the current portfolio or the variant tab you picked, with the framed period when align starting capital is on.

• Event day, Day before or Day after picks which session of each event is compared.
• For each event: the sessions, the average P/L per traded session and the win rate against the other sessions, the difference with its 90% band and a verdict (worse, better, no clear difference, or too few sessions to tell: below 10 on either side), and if skipped: the change in P/L and in Sharpe with those sessions at zero. A verdict needs the band to exclude zero and a placebo test (the event's dates moved to other sessions) to agree, as in the backtest results. A line on top says whether the most different event is beyond what chance produces when this many types are compared.
• The analysis runs on trading sessions: weekend and holiday days of the portfolio's daily P/L are left out, and the sessions between its first and last day where nothing traded count as zero, as in a backtest. The figures are in the currency of the scenario shown.

The dates come from the economic event calendar, which lives with the Research data: without Research the menu doesn't appear. Skipping is a what-if here; a Veeman backtest can skip those sessions for real with an Economic events rule.