Criticism of the thermal model of the Pioneer anomaly

I have just read through the papers by Turyshev et al. (arxiv: 1204.2507, 1107.2886) in which they argue that the Pioneer anomaly is due to thermal radiation from the RTGs bouncing anisotropically off the spacecraft antenna. First of all, well done to them and the Planetary society for the Pioneer data rescue: the data is the important thing, and I appreciate a lot of work has gone into the finite element modelling. However, I have some criticisms of their thermal explanation of the anomaly:

1) An identical anomaly to the Pioneers' was also found by Anderson et al. in data from the Solar Ulysses probe, and, less conclusively, in Galileo probe data. See: http://arxiv.org/abs/gr-qc/9808081. These spacecraft were very different, and it's unlikely that thermal effects would cause the same acceleration.

2) As they themselves say in their 2011 paper (page 4) the Pioneer data is still too noisy to prove whether there is a decay with time in the anomaly or not, and a thermal explanation can't be supported without a proven decay.

3) The half life of the decay with time that best fits their thermal model is 28.8 or 36.9 years whereas the half life of the Plutonium on board is 87 years.

4) Their predicted anomaly is at its largest in the inner Solar system where there was no Pioneer anomaly. They have got around this by proposing there was an exactly cancelling push because the Sunward side of the craft was warmer due to sunlight (Turyshev kindly emailed me to point this out), but looking at their Fig. 2 from their latter paper this does not exactly cancel the onboard thermal effects, so I guess they had to adjust the momentum flux from photons close to the Sun that was originally assumed by Anderson et al.?

5) Anderson et al. (2002) (arXiv:gr-qc/0104064, p32-33) said that, since most of the heat from the RTGs was radiated from fins whose flat surfaces were not pointing at the antenna, only their narrow edges, only 4W of power, could have hit the antenna, leading to a maximum acceleration of only 0.55*10^-10 m/s^2.

6) Since the power radiation Q is proportional to the fourth power of temperature, I'd like to see the temperature errors they get, since any errors from this source would be hugely magnified.

More generally, I always find it difficult to accept a paper when a very complex and unrevealed process (over 3000 finite elements, not fully detailed in the paper) and with fitting parameters, is used to get to a previously known answer, and no experiment is suggested that might unambiguously test it against rivals. I dislike dark matter for similar reasons. I'd like to see them present a simplified order-of-magnitude calculation so others can reproduce what they have done on paper.

Here's my explanation for the Pioneer anomaly using MiHsC, published in MNRAS: preprint.