Large increase in precision for the CaM frame order pseudo-ellipse model test data set.
The higher precision is because the number structures in the distribution is now 20 million rather
than 1 million and the numpy.float128 data averaging has been used.