The easiest way to understand noise is by thinking about measurements. The variability of the error is noise – and that’s important. In the mathematics of accuracy, the expression for total error is very simple and quite compelling. It is bias-squared plus noise-squared.