Let w ^ be an unbiased estimate of an unknown w ∈ R . Given a function t ( w ) , we show how to choose a function f n ( w ) such that for w * = w ^ + f n ( w ) , E t w * = t ( w ) . We illustrate this with t ( w ) = w a for a given constant a . For a = 2 and w ^ normal, this leads to the convolution equation c r = c r ⊗ c r .