Dragan Huterer, from the University of Michigan, gave us a very entertaining talk that stimulated much discussion about how to falsify entire paradigms of dark energy rather than just specific models of dark energy. In their recent paper Mortonson, Hu and Huterer allowed for a general form of w(z), the dark energy equation of state, by describing it with principal components. They then constrained the parameters describing these principal components, using MCMC (Markov chain Monte Carlo) methods both in the case of simulated future data from the SNAP and Planck satellites, and in the case of current supernovae and CMB data. They then used the resulting constraints to constrain the distance, Hubble and growth (all functions of redshift) and showed that even in the most general case there exist tight correlations between the growth at different redshifts which provide a very general consistency test of the dark energy paradigm.
Dragan also spoke about “pink elephant clusters”: recently detected high redshift clusters which are claimed to be very problematic for the standard model. He pointed out that, in addition to Poisson errors, one must also allow for the fact that one does not know the underlying cosmological parameters perfectly. Another important point to consider is the Eddington bias, which is the bias introduced when any given data point is more likely to have been scattered up from the theoretical curve then scattered down from an outlier because there are many more smaller mass clusters than larger mass clusters (see the Mortonson, Hu and Huterer paper for details). Taking into account these two effects, the group has shown that ΛCDM (and other models) has not been ruled out by these high redshift clusters.