DLW wishes to advise that the statement that the treatment of SNIa data, for example, appears to be dismissive would seem to have been made without reading all of the paper, which is understandable given it is 72 pages long. In view of the length of the paper, he has placed a summary of observational results and a faq at
In relation to the supernova data, see Figs 2 and 3, surrounding discussion and section 7.5. Apparent acceleration is now obtained, unlike in the rough approximations of earlier work. Moreover, the goodness of fit is now statistically indistinguishable from the Lambda CDM model. (Further details will be released in a paper with Leith and Ng, ref .) For those who wish to actually pay attention to data, the question of whether or not the "Hubble bubble" should be included (a difference between Riess04 and Riess06 gold data sets), is a serious issue for the Lambda CDM paradigm, as the supernovae data teams have pointed out (in a quiet way) in their recent eprints astro-ph/0612666, astro-ph/0701041. This feature has a natural explanation in the FB model, as discussed in section 7.5, and should ultimately be open to quantitative testing. The difference of the residuals between the Lambda CDM models and the FB model in Fig. 3 also indicate that supernovae data may prove to be a good way to distinguish the models once some thousands or tens of thousands of data points are collected, even though the present data cannot distinguish the two models.