By trying to be right, you're being silly. Your craps simulation is correct for that specific game, but it doesn't model prefix pruning. With prefixes, we abort at the first false positive without betting on future events. The probability of success is 63.2%, as my simulations demonstrate. You still fail to distinguish between sequential processes with aborts (prefixes) and conditional bets (your dice).
You inadvertently demonstrate that when you change the rules, the probability changes.
This validates my claim that "Every stopping criterion changes the probability of success."
Prefix pruning: 63.2%
Random stopping: 50%
Your craps bet: 48.2%
Dude, my craps bet (which is actually yours) is exactly what you described as "natural to bet that 6 doesn't appear again in 4 rolls". So if there's anything crappy about it, it was your original statement, which was proven to be false. If you happen to find something's wrong with it, please let me know and I'm happy to refactor it according to your conditionals or whatever. But the results will be identical, since it's simply empirical proof that in whatever 4 rolls (sequential, skipped, timed out and resumed, changing the dice, etc etc etc), any value has a 51.8% chances of showing up at least once. Conditionals are irrelevant, just like skipping is irrelevant. I think you should be the one to go back to the drawing board here.
If you have some other statement about something, or you discover what's wrong with the simulations (not flakes of snow) let me know. Congrats on discovering and demonstrating that there's a 63% chance of success to hit X at least once, when p=1/N, in N tries, btw. No one really bothered to do the math on that one. You're definitely the first.
Forget it, this guy’s ego is so inflated he never even considered that if prefix search actually worked or had any real merit, academics would’ve published on it by now. But there’s nothing because it’s just hot air. You can use any kind of filtering you want, but without the public key, you’ll still end up brute-forcing every private key anyway. In fact, its so-called "prefix search" just makes things less efficient. Good luck to him on his impossible quest.