(0.1) check calibration as a function of priors and n_gals

[ismael-mendoza]

Good to keep in mind: preliminary tests suggest that the calibration is also sensitive to bias in each of the posteriors. Both the bias and the width of the contours plays a role. Experiment with Gaussian example would elucidiate this.

• attempt to run with 10k
• make plot of calibration 10, 100, 1000, 10k
• intuition for calibration based on Gaussian toy example
• attempt to quantify sigma_true disagreement based assuming posterior is Gaussian
• check with sigma_e_int=0.5, sigma_e_int = 3.0
• try chaing prior on shear to be N(0.0, 0.3)