Hi @Talha_Muhammad!
The value is still quite low. The values get higher because you are increasing the magnitude of your computation. With 20 sides, the sums go from 0 to 40, so it is natural that both standard deviation and covariance get higher.
I think the idea you have in mind is about the correlation coeficient, this is adimensional and between -1 and 1, and it should be very low for every experiment you make in this fashion.
Thanks,
Lucas