Suppose we have a set of measurements of n quantities, xi
(i = 1 to n) each with a standard deviation si,
and true mean mi. We can calculate a quantity called chi-squared, defined as
will look at a simple example. Here the measurements are of the same quantity, i.e. a single
parameter, so mi = m
independent of xi . Therefore
varies with assumed value for m: Vary the
value on the xi plot (up or down).
for the minimum C2, it should equal the weighted mean.