Will a k-period weighted average produce a lower meansquare error (MSE) when making predictions than a weighted average derived from k minus 1 periods? If so, why or why not?
This solution is provided in 92 words in an attached .doc file. It provides the equation for mean square error, including that in relation to a Fourier series. This is explained and it is shown how k is related to the size of mean square error.