Publications and Research

Document Type


Publication Date



A simple method for the construction of empirical confidence intervals for time series forecasts is described. The procedure is to go through the series making a forecast from each point in time. The comparison of these forecasts with the known actual observations will yield an empirical distribution of forecasting errors. This distribution can then be used to set confidence intervals for subsequent forecasts. The technique appears to be particularly useful when the mechanism generating the series cannot be fully identified from the available data or when limits based on more standard considerations are difficult to obtain.


This article originally appeared in the Journal of the American Statistical Association.



To view the content in your browser, please download Adobe Reader or, alternately,
you may Download the file to your hard drive.

NOTE: The latest versions of Adobe Reader do not support viewing PDF files within Firefox on Mac OS and if you are using a modern (Intel) Mac, there is no official plugin for viewing PDF files within the browser window.