Quote from: Bram24732 on April 17, 2025, 07:09:23 AM

Got bored this morning so I wrote a little snippet which looks for a value in a uniform space.
It does it using both methods :

Sequential random (i.e. the one I use)
Prefix based, I believe this is the WP version, but the code can easily be adjusted to your flavor of prefix optimization.

By defaults it runs 1000 simulations of a search in a 100M elements space with a range of 100k.
This is exactly equivalent to the 69 problem in terms of statistics. Just simpler because it uses integers instead of hashes, so we can run many more simulations for cheap.

How to run :

Code:

Have nodeJs installed on your machine.
Download script
Run "node prefixVsSequential.js"

Results, as you can see there is very little deviation. The more simulations/bigger ranges you run, the closer both values will be to 50% :

Code:

Average match percentage for random sequential method : 49.63%
Average match percentage for prefix sequential method : 52.35%

Link to script :

Code:

https://gist.github.com/Kowala24731/e9a52fcf08b14934a5e8e96ae967c729

You can change those parameters to update number of simulations / size of problem / ranges / percentage of deviation before skipping a range

Code:

const PREFIX_DEVIATION = 1.35;
const TOTAL_SIZE = 100_000_000;
const RANGE_SIZE = 100_000;
const NUMBER_SIMULATIONS = 1000;

so you were going to prove that I was wrong, someday?

Quote from: Bram24732 on April 16, 2025, 06:19:21 PM

Quote from: mcdouglasx on April 16, 2025, 05:50:26 PM

Quote from: Bram24732 on April 16, 2025, 05:36:58 PM

We already explained a million times why, formally, the prefix theory would break 3 different mathematical constructs if true.

What I proposed doesn’t do it, you assume too much, you say, 'the theory doesn’t work'. Which one? Which of the 20? Just me alone, I have 2 approaches, WP has another, Bibilgin has another, NoMachine... and I could go on. That’s why your words don’t hold value.

Regarding those spreading FUD, read about Curt Richter, optimism, and hope as a psychological construct.

You’re right. Let’s have fun.
Please write down how your approach works here. I’ll replicate it with a python script searching a key in a 100M keyspace. Both fully random and using your approach. Well be able to check if it does provide an edge.