Oh well, OK. It's definitely not MY job to convince YOU of anything, it is YOUR job to convince US that the established and well proven facts we know of are the wrong ones.

So far, you failed to prove in any way that your theory or scripts are in any way better than sequential search. To do that, it is your responsibility (not ours) to come up with actual benchmarks and results, that compare directly, one to one, success rates of your method, applied to the same set of data as the sequential method.

Obviously you need to do that over a lot (and I mean a lot) of actual random data sets, not the ones you do with Mersenne-based PRNG so far.

If you consistently get even a 0.01% faster solve times over sequential, then you can claim that your method has the slightest seed of truth in it. If not, please stop bragging about whatever you believe to be "AI" posts, when they are actually simple scipy one-liners. Or is SciPy written by AI, not by actual real statisticians?

And I definitely don't need to bother running your script above, simply because I can follow what it's doing simply by reading it. Also, just as a side-note, that entire script can be written in half the number of lines.

Good luck anyway.