Conjecture:
If
or
then the m by n grid is unsolvable.
[Also, side note: m refers to column, n refers to row. This is because of the proper alphabetical order and the usual way of referring to things as row then column.]
Examples that I have (some) solutions for.
Numbers: [{1 2 2 ... 2 1}]
XO grids: [{XOXX...OXXO}] and [{OXXO...XXOX}]
2] m x 2:
Numbers: [{2 3 3 ... 3 2}{2 3 3 ... 3 2}]
XO grids: [{XOXO...}{OXOX...}], [{OXOX...}{XOXO...}], [{XXX...}{OOO...}], [{OOO...}{XXX...}]
3] 5 x 3:
Numbers: [{2 3 4 3 2}{3 5 6 5 3}{2 3 4 3 2}]
XO grids: [{XOXXO}{OXOXO}{OXXOX}], [{OXXOX}{OXOXO}{XOXXO}] + possibly others
4] 8x8:
Grids shown below:
A few of these I have after doing the 5 x 3 case, which is why a few are specific to odd ms and ns.
- In the number grid, the numbers increase towards the center, both horizontally and vertically.
- The two solutions are reflections of each other over either the vertical or horizontal axis (so far, I think that both transformations yield the same result, so either gets from one solution to the other)
- The number of solutions pairs (those that are reflections of each other) is related to n, the number of columns. At the moment, I believe it is the number of columns.
- In the case where n and m are odd, the middle row and column will be the same in both solutions. (I'm not sure how this works yet if they're even. Maybe it doesn't. Yeah, I don't think it does, because of the m x 2s.)
- If m and/or n are odd, the pattern of Xs and Os will be alternating in the middle row/column. (I am unsure about this.)
- In any m x n and in any vertical or horizontal group of three, there will be at least one O and one X. (Also unsure about this one.)
- The patterns for the rows will be either O+XXOs+X, X+OXXs+O, XOXO... or OXOX. (Again, unsure...)
The 5 x 3:
I started with the middle column (number 3) with XOX going down. This is because if the conjecture about flipping was correct, than this middle column would remain the same in both solutions. And XOX seemed like a likely candidate...
From there, I started filling in Xs. I tried to make it symmetrical over a kind of diagonal line (I'm actually not sure why...it just sort of happened.) I think I might have misinterpreted my own conjecture. Anyway, it worked. If this is confusing, look at the picture below. I thought about the grid squares as corresponding to others and those pairs had to have the same thing in them. The colors that are the same are the ones that correspond.
I filled in the Xs. Here they are below with same colors as the grid squares above.
And with this done, I have the solutions below.
They are reflections over horizontal or vertical axes.





Those are lovely pictures and great examples of why those sizes aren't solvable. It looks like from your mx2 cases that you have puzzles that have x's and o's, which are not simple transformations of each other. (Really I think you can put a single x in every column, in any row and you have the [{2 3 3 ... 3 2}{2 3 3 ... 3 2}] pattern. Right?) So I think you've disproved the hypothesis that they had to be transformations.
ReplyDeleteSide note, has your book arrived yet?
Our book has not arrived yet, but hopefully it will soon.
Delete