The sample autocorrelation function and raw periodogram both show the same data, but the autocorrelation function can be estimated more accurately using a parametric time series method. This method, implemented in a MATLAB program, selects the best time series model for unknown data characteristics. The parametric estimates of autocorrelation are more accurate than mean-lagged-product estimates. Eventually, parametric estimates will become more reliable and provide a clear answer to how long the autocorrelation function truly is.