Rolling return and observed drawdown calculator

A rolling window is not one start-to-finish result

A single holding-period return can conceal how outcomes differed for investors entering at other dates. Rolling windows repeatedly apply the same observation span to neighboring points. This calculator reports the lowest and highest annualized result across those windows.

The window input counts observation intervals. It is not a promise of a fixed number of calendar months. Dates determine each actual duration, so irregular observations can produce windows of different day lengths even when they contain the same number of intervals.

The XIRR and XNPV dated cash flow calculator incorporates dated deposits and receipts. A personal money-weighted return is a different question from rolling performance of an adjusted NAV series.

Drawdown remembers the preceding peak

At each supplied date, the tool compares the NAV with the highest NAV already observed. A recovery can shrink the current drawdown without erasing a deeper historical drawdown. The largest recorded percentage is retained as maximum observed drawdown.

Only the observations you provide are visible to the model. A monthly series can miss a fall and recovery inside the month. The output therefore says observed drawdown rather than claiming to reconstruct every intraday or daily trough.

Use comparable index values

Supply adjusted NAVs or index levels on a consistent basis. A payout that reduces an unadjusted price can look like investment loss if distributions are omitted. Personal contributions and withdrawals belong in a dated cash-flow return calculation instead.

Short windows can annualize to extreme percentages. That is the consequence of extending a short observed ratio over a year, not a forecast. Compare the raw dates and day counts in the table before interpreting a very large annualized number.

Work through the example

For NAV observations 100, 120, 90 and 110, maximum observed drawdown is 25 percent: the decline from the observed peak of 120 to 90.

Use the worked table to trace where each input enters. Then change one assumption and calculate again, leaving the others fixed. A comparison is easier to interpret when the change has one identifiable cause rather than a different set of measurements, dates and rates all at once.

What this model does not establish

Positive strictly date-ordered observations, at most 500 rows. Observation-count windows; actual/365 annualization. No personal cash-flow adjustment, live NAV, distribution repair or future forecast.

Keep the input assumptions with any exported result. Illustrative values do not become observations merely because the output is precise. For a real purchase, financial reconciliation or technical decision, check the underlying records and applicable requirements before treating this scenario as a final answer.