Convergence plots using equation (25), where we have kept the number of simulations used for our derivatives constant at 400 and are solely varying the number of fiducial simulations used for the covariance. We consider the different noise cases: (a) no noise, (b) 100 hours of SKA noise, and (c) 1000 hours of SKA noise. We see that by 200 samples all of our statistics are fully convergent, falling within the 10%, shown as the red shaded region.