[ ] Can we predict how many levels of sub-pixel sampling that we need as a function of r, b, and t_{exp}?
[ ] Revisit Fisher information analysis ala Carter et al. (2008), without time-integration, and Price & Rogers (2014) with time-integrated light curve. In addition, it would be interesting to use celerite in this analysis to investigate the impact of granulation noise on transit light curve parameter determination, and thus the limits of transit transmission spectroscopy and TTVs in the presence of correlated astrophysical noise.
r,b, andt_{exp}?