On this page
Count what one row represents
If each observation is a full year, use one period per year. If each is an equal month, use twelve. Two monthly observations then cover two twelfths of a year. Two observations alone do not establish that two years elapsed.
Compare frequency settings
For +20% and −20%, cumulative return remains −4% under either annual or monthly labeling. Annual labels give geometric rate √0.96−1; monthly labels give 0.96⁶−1 because the two-month horizon is one sixth of a year.
| Periods per year | Covered years | Annual exponent |
|---|---|---|
| 1 | 2 | 1/2 |
| 12 | 1/6 | 6 |
Diagnose an implausible annual result
A large annual equivalent can arise from a short observed horizon. First inspect the row duration and frequency before assuming the cumulative calculation failed. The annual number is not a prediction that the short-period outcome repeats.
Keep uneven dates out of this control
An irregular-date series cannot be correctly annualized merely by choosing a convenient frequency. Supply the actual elapsed time in a suitable method. The tool deliberately asks for equal periods so the conversion can be reproduced.