1: \begin{abstract}

2: We study the Blackwell-Rao (BR) estimator of the probability

3: distribution of the angular power spectrum, $P(C_{\ell}|\mathbf{d})$,

4: by applying it to samples of full-sky no-noise CMB maps generated via

5: Gibbs sampling.  We find the estimator, given a set of samples, to be

6: very fast and also highly accurate, as determined by tests with

7: simulated data.  We also find that the number of samples required for

8: convergence of the BR estimate rises rapidly with increasing $\ell$,

9: at least at low $\ell$.  Our existing sample chains are only long

10: enough to achieve convergence at $\ell \la 40$.  In comparison with

11: $P(C_{\ell}|\mathbf{d})$ as reported by the \emph{WMAP} team we find

12: significant differences at these low $\ell$ values.  These differences lead

13: to up to $\sim 0.5 \ \sigma$ shifts in the estimates of parameters in

14: a 7-parameter $\Lambda$CDM model with non-zero

15: $\textrm{d}n_{\textrm{s}}/\textrm{d}\ln k$.  Fixing

16: $\textrm{d}n_{\textrm{s}}/\textrm{d}\ln k = 0$ makes these shifts much

17: less significant.

18: 

19: Unlike existing analytic approximations, the BR estimator can be

20: straightforwardly extended for the case of power spectra from

21: correlated fields, such as temperature and polarization.  We discuss

22: challenges to extending the procedure to higher $\ell$ and provide

23: some solutions.

24: \end{abstract}

25: