Rolling returns vs point-to-point

How you measure a past return changes the story it tells. A point-to-point return uses a single start date and a single end date. Rolling returns measure the same length of holding, but starting on many different dates in turn. The difference between them is mostly about how much the choice of dates is doing the talking.

What a point-to-point return is

A point-to-point return is the plainest kind: value at the end compared with value at the start, over one chosen stretch. It is easy to quote and easy to understand. Its weakness is that it rests entirely on those two dates — change either one and the figure can change a great deal.

Why the dates can mislead

Because markets rise and fall, a point-to-point figure can be flattered or damned by where its endpoints happen to land. Measure from a low point to a high one and the return looks wonderful; measure from a high to a low and the same holding looks poor — even though nothing about the holding changed, only the dates. This is how a carefully chosen start-and-end window can make almost any case.

What rolling returns do

Rolling returns tackle that by not relying on one window. You fix a holding length — say a period of some years — and then calculate that return starting from every date across a long history: the return for the period beginning in January, then February, then March, and so on. That gives you many overlapping results instead of one, which can be summarised to show how the holding behaved on average and how widely it varied, rather than on a single lucky or unlucky run.

An illustration

Illustrative, not any real fund: suppose a holding was measured just from a market bottom to a market peak — the point-to-point return would look striking. Roll the same holding length across hundreds of start dates, including the bad entry points, and you get a spread of outcomes that is far more representative of what an ordinary investor, who did not happen to buy at the bottom, would have experienced. The rolling view is harder to cherry-pick.

What it still cannot do

Rolling returns are a fairer way to read the past, not a way to see the future. They still describe history, and past performance is not a promise. They are also best read as an annualised CAGR over each window. Use them to judge how consistent a holding has been rather than to be dazzled by one number, and always alongside the scheme document, never as a forecast.