Let me clarify, what I meant to say was that if a truly brilliant (or merely technically competent) person wanted to cheat, all he would need to do is create a mathematical model of the expected observations, including measurement noise, pick a random number seed and let 'er rip.
If he produced "oddly similar" graphs, there are two explanation, the actual underlying physical models are oddly similar or he is cutting corners. Occam's razor points me in a direction that I don't like to go. We need to apply doses of healthy skepticism to any unusual claim in the world of human affairs.
An example of seemingly unrelated physical processes that have similar underlying models are the Fourier Transform of a continuous time process and the far field diffraction pattern of an aperture. (F'rinstance, the graph of the Fourier Transform of a rectangular pulse looks just like the far field diffraction pattern of a uniformly illuminated aperture.) Turns out, on closer examination the equations of a Fourier Transform are similar to Fraunhofer's equations for a diffraction pattern. That would be one example where oddly similar results have a valid explanation.