2x2 random state optimal scrambling

Tim (2008-05-14 23:23:25 +0000)
Hello, It seems logical that a random state scrambler be used for the 2x2, similar to CubeExplorer for the 3x3. After all, the 2x2 is a very simple puzzle (for a computer) and can be solved optimally in 11 moves (according to Jaap's page). Recently, I decided to write a 2x2 solver with this function in mind. I finished the program, and I would like to propose that the WCA switch to this method of scrambling. The main benefit, I believe, is efficiency. The average 2x2 position can be solved in 8.76 face turns. We are currently using 25 moves, this seems excessive. This method of scrambling also guarantees as uniform a distribution as Java's random generator will allow. The URL of the scrambler is [url:pq1nzqod]http://www.rubikscubing.com/2x2Scramble.jar[/url:pq1nzqod]. Feel free to make any comments or criticisms; I will do my best to address them. I have not tested this on anything but Windows; I believe it should work, but let me know if it doesn't. Tim Reynolds
Ron (2008-05-15 04:29:13 +0000)
Hi Tim, Good idea and good work. Would it be possible for you to transfer your program from Java to Javascript? Actually, to embed it in Jaap's script? Thanks, Ron
Tim (2008-05-15 04:33:37 +0000)
Hi Ron, Sure, I'll start working on that as soon as I get a chance (probably not until about Sunday though). Tim
StefanPochmann (2008-05-15 17:33:28 +0000)
Java's Random class is sometimes very bad, producing patterns it really shouldn't. If I remember correctly, the SecureRandom class is much better. Don't know about JavaScript, but I recommend exploring this issue.