It starts with the assumption that our observations of cosmic expansion are false and due to time dilation from a clumpy universe, then works out how clumpy the universe must be to account for our observations.

For timescapes to "work" time must run 38% faster in cosmic voids. This works out to a density difference of about 100,000 times what we observe with many existing methods.

> It starts with the assumption that our observations of cosmic expansion are false

It starts with the assumption that our measurements (observations!) of attenuation and red-shift of standard candles are correct. Your characterization is wildly wrong, unless you meant to use a different word than 'observations'.

> and due to time dilation from a clumpy universe, then works out how clumpy the universe must be to account for our observations.

Yes, it does work backwards. That's insufficient by itself to say that timescapes is incorrect. It's concerning though because they work out what densities must be in order to explain away all of the attenuation/red-shift disparities. However the reasoning time dilation is not wrong -- in fact, it's blindingly obviously correct if GR is correct, it's just that it might also be wildly insufficient -- it might just be in the noise.

> For timescapes to "work" time must run 38% faster in cosmic voids. This works out to a density difference of about 100,000 times what we observe with many existing methods.

Thanks. This is responsive to my question, but it's not yet dispositive. We could find that:

  - some of our observations are incorrect
  - some of our interpretations of those
    observations are incorrect
  - timescapes explains some but not all
    of the apparent acceleration of the
    expansion of the universe

Also, density differences need to take into account density differences at the time that a ray of starlight we observe today entered each void along the way to us, not current apparent density disparities. It has to be that 8 billion years ago the difference in density between clusters and voids was much starker, though the voids would have been smaller, and perhaps the difference was not enough orders of magnitude.

There's a ton of theoretical work out there that is about building a foundation for future work. This is one of those works. But people tend to overlook the part where the work is just a piece, and many other things most first fall into place and over hype it. This is course isn't the fault of the authors.