Return history is your portfolio’s whole record, month by month, on one screen. Open Analytics, and at the foot of the Monthly returns card press Full history →; ← Back to analytics above the title takes you back. It reads the portfolio selected in the header — the combined portfolio included — and every figure on it is a percentage.
Return history is part of Ultimate, like the rest of the analytics board. Every new account gets Ultimate for seven days once the email address is confirmed, with no credit card.
On the free plan the screen is still there, laid out exactly as it is for a subscriber, but filled with a six-position demo portfolio rather than yours. A strip at the top says so — Sample portfolio · illustrative — and carries the Upgrade to Ultimate → link. Nothing on that version is computed from your own positions, so it is a way to see what the screen does before deciding, not a reading of your account.
If the portfolio is younger than one calendar month the screen says No closed months yet: returns are measured between month-end closes, so the first figure appears once a full month has passed since your earliest transaction.
The first card, Returns over time, draws one bar per period, green above zero and red below. Returns are flow-adjusted: money you deposit or withdraw is never counted as a gain or a loss. Two controls set what the bars cover:
Not every pair makes sense. A single year shows months only, 12m shows months or quarters, and Annual needs All; the unavailable buttons are greyed out, and hovering one says Pick “All” to fold the history into longer periods. Your choice is not lost — widen the period again and the granularity you picked comes back. A long history scrolls sideways rather than squeezing the bars.
When there is a benchmark history for your portfolio, the card head carries two more controls. They answer two separate questions, which is why they are two controls and not one switch:
Comparison is a layer over your own series, not a replacement for it, so it is a toggle rather than a third position. It is disabled under Difference: the benchmark is already inside that number, and drawing SPY’s percentages beside percentage points would put two units on one axis. The caption under the chart always says which of the three readings you are looking at.
In Difference, a quarter or a year is not the monthly gaps added up. Each side is compounded into its own quarter or year first, and only then is one subtracted from the other.
Consistency — since inception is eight tiles, always over the whole history whatever the chart above is showing. Each label has an info tooltip with its definition and, where the term has a glossary entry, a link to it.
Win rate is not recalculated on this screen. It is the same figure as the Win rate in the analytics board’s Headline numbers, read from the same source and always over the whole history.
Counting positive months from the heatmap could give a different answer when the history has a hole in it — the two methods treat a gap differently — and the same tile name would then carry two numbers on two screens. One source keeps it one number. Like any win rate, it ignores size: eleven small gains and one large loss still read as a high win rate, which is why it sits beside Worst month and the streaks.
Monthly returns — every month on record is one row per year and one cell per month, shaded green or red by sign and deeper the larger the move; hover a cell for its month and value. The analytics board shows the same table capped at the last eight years; here nothing is cut.
The last column, Yr TWR, is the year’s time-weighted return: its months chained together with deposits and withdrawals left out. It measures the holdings rather than the timing of your money, so in a year with a large deposit it can differ from the “% on capital” in the board’s performance chart. Both are right; they answer different questions, as time-weighted vs money-weighted return explains.
Distribution — how monthly returns cluster sorts every measured month into ranges and draws a column per range: taller means more months landed there. Zero is always a boundary, so no column mixes gains with losses. One extreme month does not stretch the scale: once there is a year of history, the outliers go into the two end columns, labelled with ≤ and ≥. Under the chart one line names the middle of the shape: where half of the measured months landed, and the median month.
Seasonality — the average calendar month is twelve cells, January to December, each the geometric average of that month across every year you have. The number of observations is printed in each cell (3 obs) and is part of the reading: three Aprils are a coincidence, not a pattern.
Sometimes a month cannot be measured — the first month of a history has no previous month-end to measure from, and a break in the history leaves a month without a figure. Such a month is shown as missing everywhere, never as 0%:
Treating a gap as zero would quietly flatter or penalise every figure built on the series, so the screen shows less rather than something invented.