It's paradoxical, because while I went straight to the point with what I wanted to show, you don’t accept it because the math turned it into an ego race. Your proposal is total distance / simulations.
It's like saying Usain Bolt runs 10 races of 100 meters against Pepe, wins 7, and then concluding that both are equal by adding up 100m x 10 = 1000m. Since both have covered the same distance, both runners must be equal.
That’s magnificently absurd.
Except you forgot that your code counts EXECUTED checks, not FULL RANGE checks. So they are not running the same distance at all, they do not cover the entire track, only up to finding the solution.
So what is absurd, is your method of comparation by dividing actual work by number of wins, without caring about relative effort, and even ties.
By your judgement, if Usain wins by 1 millisecond, than Pepe didn't even compete.
And also by your judgement, the time (ops) quantity does not even matter, all wins are equal.
And I think you forgot that somehow, magically, prefix method "wins" even when not even looking for a prefix - simply by breaking itself up once every whatever chances.
Really, lol... I'll leave you so you can keep deceiving yourselves among yourselves.
What you get by setting SHARED_BLOCK_ORDER = true is exactly what you get with your original code. Setting it to false changes the result, precisely because we're introducing true randomness and making a genuinely fair comparison between the two methods.
Fatal mathematical error-comparisons must always be made under the same conditions. That’s the only way to truly determine which is better. Comparing under different conditions is unrealistic and inaccurate, but I know it's done to slip in bias, which I deliberately introduced in my first script to catch your attention.
You're obviously failing, but you can still deceive the uninformed with a beautifully crafted fallac-one I already anticipated from my first script.
Now, to play around with your theory: if both follow the same order, and that order changes in each simulation, why on earth does it favor prefixes? Where’s the magic?
Does it prove that prefix-based search is better? Or is Python rigged?