Turing Tumble Community

Puzzle 48 - 26 parts, only 11 balls

This solution only was possible because I accidentally placed register A one column further to the right than the diagram indicated. This allowed me to place a ramp that bypassed the 2 bit when the 1 bit was switched from on to off. From there I just traced the paths of which bits would need to be flipped to get to the next multiple of three. I probably should have counted it as 27 parts and put an interceptor at the bottom of register A so that an overflow would be caught.

IMG_4322

Clever, but if you look carefully at the problem, Register B(the four bit register on the left) is supposed to start at zero.