Wednesday, August 20, 2014

August 20


I looked over the nx1s again. It appears that if n is 2 more than (or one less than) a multiple of 3, it is not necessarily solvable. The pattern of numbers that makes this so is 1s with 2s in the middle; in particular, the number of 2s is appears to be a multiple of three. From there, looking at the patterns of Xs and OS that can be generated, they're in the form of repeating XXOs with an O and an X at the end. I'm assuming that these qualify as unsolvable even though the two solutions are reflections. If transformations like this are counted as the same, then these would be "solvable."

I also looked over the counts for "without transformations" for nx1s and revised the formula. I think it works now, but I'm not completely sure.

1 comment:

  1. 1xn's are perfect, and you've got the pattern. This could be described mod 3 to make it shorter.

    I'm now trying to think about your without transforms and needed a clarification. Earlier you said (as in, XO=OX and [XO,XO]=[OX,OX]=[XO,OX]=[OX,XO]) , I understand the first = the second and the third = the fourth, but would not have said they all were the same even without transformation. What sort of transformations are you eliminating? Are we talking being able to flip it over in one direction or another and having it match? Are we including rotations as a valid transformation?.... I may just be confused.

    Any conjectures as to which nxm's are going to not be solvable now that you know which 1xn's are solvable?

    ReplyDelete