Given the generic definition of statistics, it is clear that astronomers have engaged in statistical analysis of some variety since astronomy first emerged as a science. However, from the early nineteenth century until the beginning of the twenty-first the two disciplines have been somewhat estranged – there was no formal relationship between the two. This has now changed, as is evidenced by the recent creation of the International Astrostatistics Association (IAA), the ISI astrostatistics committee, astrostatistics working groups authorized by the IAU and AAS, and this Symposium. The challenge for us to come is in establishing how statisticians and astronomers relate in developing the discipline of astrostatistics. I shall propose a direction for how the discipline can progress in both the short term and well as for future generations of astrostatisticians.

I conclude the SCCC21 conference highlighting some of the contrasts we heard about, some specific topics (the statistics of random fields), and discuss how some of these play out in the analysis of cosmic microwave background data. I conclude with a hopeful look at the efficacy of blind analyses in the CMB and elsewhere in cosmology.

The proof corrections supplied by the author were not made to the final version of the article (Guarnizo, Amendola, Kunz and Vollmer, 2015).

We apologise to the authors and to the readers for this omission.

After (2.6) on page 348, the statement “being S = AR.” should read

“being S = AR. and P1 = R/A.”

Improved versions of the figures are also presented below.