Generally they are all constructed as the summation of the squared auto correlation coefficients up to mth order. ( in your case is 6).
The basic idea is that if Q(m) is zero, every auto correlation coefficients up to mth order should be zero. (e.g. if a^2+b^2=0, a=0 and b=0)

The degree of freedom is adjusted so that to let the statistic follow chi-square distribution.

The original version is very simple:



The modified one is:


The tao here are auto correlation coefficients. T is the number of observations.

I think we use modified one more often.

best,

谢版主！

谢版主！