Gurzadyan dismisses the critical analyses as "absolutely trivial", arguing that there is bound to be agreement between the standard cosmological model and the WMAP data "at some confidence level" but that a different model, such as Penrose's, might fit the data "even better" " — a point he makes in a response to the three critical papers also posted on arXiv5. However, he is not prepared to state that the circles constitute evidence of Penrose's model. "We have found some signatures that carry properties predicted by the model," he says.
I only read the BC paper (Moss et al, [3]), because I knew one of the authors (and he's a smart guy). Anyway, they show that you can find triangles just as easily, and that you can find similar circles in computer-generated random data with the same properties as the background.
At the same time, Penrose likes to be controversial.
So I guess I'm saying that I'm willing to take your money...
From an Bayesian/Occam's Razor standpoint, Penrose's data could fit the data "even better", but as it's (arguably) a more complicated model, still have a lower posterior probability.
Would a universe with infinite time extent be easier to believe than one with a definite start point? I have enough trouble merely comprehending the possibilities, much less assigning prior probabilities to them.
Gurzadyan dismisses the critical analyses as "absolutely trivial", arguing that there is bound to be agreement between the standard cosmological model and the WMAP data "at some confidence level" but that a different model, such as Penrose's, might fit the data "even better" " — a point he makes in a response to the three critical papers also posted on arXiv5. However, he is not prepared to state that the circles constitute evidence of Penrose's model. "We have found some signatures that carry properties predicted by the model," he says.
My money's on Penrose.