Time series data and metric types

All metrics used at pfolio are calculated from an asset's daily prices time series data or its derivative, the daily returns time series data.

While many other types of data can be utilised to analyse an asset (e.g. a stock's fundamental data, a currency pair's carry etc.), we limit ourselves to price action data as it is easily available for any asset and carries a lot of signal.

Daily prices and returns time series

To standardise across all assets, the daily prices time series data is exclusive of weekends but inclusive of stock exchange holidays.

To learn more about how we process data to handle missing data and stock exchange holidays, read this article.

Daily returns distribution metrics

Daily returns distribution metrics take all daily returns in the considered time frame into account.

For example, the mean return for a 10-year time frame is calculated as the arithmetic mean of all daily returns in the time frame. The result can then be scaled to a given period, e.g. multiply the daily value by 252 (average trading days per year) to get the annualised value.

Examples: Mean return, volatility, Sharpe ratio, value at risk

This metric type is useful as a statistical measure, given the robustness against outliers and the independence from the order in which the values appear.

Price points metrics

Price points metrics compare two price points in the considered time frame.

For example, the cumulative return for a 10-year time frame is simply the return between the last day and the first day.

Examples: Cumulative return, maximum drawdown, Calmar ratio

This metric type is useful for reporting purposes, e.g. return per calendar year.

Rolling metrics

A rolling metric calculates a given metric every day in the selected time frame.

Example of rolling metric chart

For example, the rolling 12-month annual Sharpe ratio calculates the Sharpe ratio on every day in the time frame over the preceding 12 months of that day and therefore creates a time series of Sharpe ratios.

Rolling metrics are useful for exploring trends in the data.

‍

‍

‍