A date period always starts on the first of the month and ends on the last day of the month and is inclusive, i.e., if the date period is [Jan 1 1980, Jan 31 1980] then calculations will include both Jan 1 & Jan 31 and all days between. Since the app only supports monthly returns right now any calculations using that date range will include only the January 1980 return.
Shortcuts are pre-defined periods. The end date for a period is always the reference date. The reference date is updated when we refresh the data each month and should typically be the latest month end, i.e., if the current date is July 15, the reference date should be June 30. If the reference date is June 30, 2023, choosing the 1 year shortcut will set the period to [July 1 2022, June 30, 2023].
If a series does not have data for all the dates in the date period then the app will subtly indicate this by changing the date range and by changing the background color on the date field to light red to warn you that it did so. For example, if you set the date period to [Jul 1 2022, Jun 30 2023] and then add a series that only has returns from Feb to Mar, 2023 then the date selection will be changed to [Feb 1 2023, Mar 31 2023] and background on both the start and end date fields will be red. The date selector does remember the originally selected period and removing the series will cause the date selection to revert back to [Jul 1 2022, Jun 30 2023].
The All shortcut is useful as it will select largest common period available for all the series that are currently loaded in the view. As series are added and removed it will adjust to reflect the largest common period.
The Previous shortcut remembers the last selected date period and is useful for switching between two date periods.
TIP Some charts support changing the date period by dragging. If you left click in some charts, drag the mouse and release the date range will be updated based on your selection. Use the Previous shortcut to undo the selection.
The app only supports monthly returns for now. Performance returns are reported as annualized geometric (compound) returns, unless otherwise stated.
Represents return per unit of risk for an investment and is useful for comparing investments with different risk levels.
Sharpe Ratio = Rp / σp
where
Represents active return per unit of active risk for an investment and is useful for comparing relative performance of investments with different active risk levels.
Information Ratio = Rp / σp
where
Difference in means test
Many times in finance we observe differences in returns between securities or differences in return of the same security under varying economic conditions.
In such cases it is very useful to understand whether the differences are meaningful, i.e., beyond mere random chance. The most common test to answer this is a difference in means test.
In our app, we report the Welch t-test which tests if two populations have equal means where the samples may have unequal variances. Wikipedia
When the absolute t-test values exceed ~2, then the test is statistically significant at the 5% level (i.e., would occur only 5% of the time by mere chance).
