I'm saying you're in denial of the basic rules of PROBABILITY you are so fond of. Your knowledge in the area is null, otherwise you would maybe understand the first law of independent events (e.g. hashes in your case): they don't have memory, and their history (sequences) do not matter. You're also totally clueless on the MATHEMATICS level as well, otherwise you would quickly realize (it takes 30 seconds) that using the prefix as some sort of base for averaging and scan-jumping is flawed, in the same way that using the suffix, or any middle part, is flawed. What you should be looking for are the RIPEMD-160 bits matching 1:1, not some base-whatever representation, which actually diverges from the raw information you're seeking. And the only way to do that matching is brute-force, no BS crackpot probability statistical theory. What you are doing is exactly the Gambler's Fallacy put in practice, and unfortunately nothing anyone tries to prove it to you in any way will ever change your mind. You will continue to believe that because you found X prefixes there are only some semi-exact number of prefixes, and that they are somehow spread apart, which is in complete opposition with their intended design: nothing stops the range from having much more or much less than N amount of prefixes, and nothing stops two consecutive/close keys to hash to the same value.

Also you seem to use some very slow software, so basically you're just wasting electricity and are proud of that. All I can say is keep it up. NB: I said many times I don't care about any puzzle below 100 ~ 110 bits, so my advices to you are totally free, take them or leave them.