Hello Klausi,
It is not circular reasoning. I think you have misinterpreted it.
The second "theoretical" part of the NEDU report addresses the issue of why they got the result they did, and whether there are any decompression profiles that could be expected to give a different result (in particular, they asked the question "are there any deep stop / bubble model profiles that would give a better result than the shallow stops / gas content model profile that we tested").
To answer the latter question they started from the premise that tissue supersaturation is the fundamental determinant of decompression stress and bubble formation, and that bubble formation produces symptoms of DCS. There would be few (if any) decompression scientists who do not believe in this paradigm. With that in mind, their analysis calculated supersaturation in a representative fast tissue, a slow tissue, and a combination of the two for 500,000 possible ways of decompressing from the NEDU test dive profile. These possible decompression approaches inevitably included many that would look like a bubble model (with more emphasis on deep stops) and many that would look like a gas content model (with less emphasis on deep stops). No profile that emphasised deep stops more than their tested gas content model decompression (their "shallow stops profile) produced lower combined tissue supersaturations. They therefore concluded that no profile emphasising deeper stops than their gas content model profile would be likely to produce better outcomes.
I believe you are mistakenly equating the basic mathematical tools used to calculate supersaturation in tissues with gas content models (or dissolved gas models as you put it). In fact, all models (bubble models and gas content models) use these same basic tools to calculate dissolved gas tissue pressures and tissue supersaturation. What differs between models is how they then interpret allowable supersaturation in different tissues at different stages of the ascent. But it is perfectly legitimate use these basic tools to compare the supersaturation patterns in tissues produced by different models. This does not amount to evaluating a bubble model using a gas content model as you seem to believe. It amounts to comparing a bubble model and gas content model using mathematical tools common to both of them. It therefore does not create circular argument.
Can I say before this goes any further that I have little appetite for debating this entire issue again. Ross Hemingway and I have finished such a discussion just weeks ago on TDS, and both of us had a fair amount of air time. You can evaluate that for yourself - it is not hard to find. The discussion has also taken place previously on this forum.... on every forum in fact. I really don't see the point in trundling it all out again.
Simon M