Stefan
Member
Or asking (I just did)But it's php based so there's no way to check whether it works (or how well it works) without hacking.
That's another way to look at it, I guess. I never said x and y are nonzero, so with that I just meant *all* U+D turns, and the first two numbers were those where only one layer got turned. I should've written that in words to make it clearer, sorry.I see, the third number is the first two PLUS the (x,y) with nonzero x and y.
Yeah, some documentation and descriptive variable names would be good. Biases... well, my two definitions (weighing or not weighing the turns) are also biased. And yours might still differ, I'm not quite sure. Question still is whether in the long run, they all end up with the same distribution, and if not, which is "best".Jaap's scrambler [...] has some weird biases and the functioning is pretty much opaque.
But they truly are the most unbiasedI don't think we should ever go by what people who don't "know[...] how to use the puzzle" think.
Ooh, YEAH, my favorite shape. Let's go for this! Not sure whether I'm kidding, to be honest . Though I guess I'll just have to wait until some inventor mixes a 2-layer Square-1 with a 2x2x2... (or does that exist already?).Square-1 in a position [...] halfway through a / move