The current scrambler for Pyraminx gives a solution to a random position, but the move sequences tend to be pretty long (sometimes longer than the solution I use to solve it!). Since modern computers are more than capable of solving a pyraminx optimally, I thought I'd make an optimal, random-state pyraminx scrambler. It can be found [url=http://mzrg.com/miniSites/scramblers/scramble_pyraminx2009.htm:243mdoy3]here[/url:243mdoy3]. I got the code for an optimal solution from Jaap's website - his Pyraminx applet has had that feature for a long time. It took some work to remove all the unimportant stuff and to figure out how it worked well enough to be able to generate a real random position, but as far as I can tell it works great now. (The rest of the code, such as all of the image processing stuff, comes from the current official Pyraminx scrambler. I suggest that this scrambler be used as the official one, unless people have objections to this. Note that it is possible to make a small change in the end of the dosolve() function to guarantee that every given solution will be exactly 11 moves (excluding tips) - this may be a good idea if you want to make sure scramblers don't know when a given scramble will be particularly easy.

Looks great, but for reasons stated previously in reference to 2x2 I'd suggest the change that you mention.

I personally prefer the concept of simply discarding all scrambles that are too close to solved. 11 moves is actually a lot for Pyraminx - very few positions have an optimal solution greater than 9 moves - so making sure everything is 11 moves would significantly increase the number of moves scramblers have to perform. And of course it is still possible to have a lucky scramble... I just don't want to have to compete with someone who got a 3-move solution.