# Talk:74: Su Doku

If that puzzle is 4 (i.e. 2x2) domains of 1x1 cells or 1 domain of 4x4 cells then it's actually an impossible puzzle. Sudoku grids for 'n' symbols (ignoring some very interesting variants) need to be of n² cells in total with n cells in each direction, composed of n 'domains', each of n cells so as to contain *one and only one* of each symbol in use. That's 81 cells in a traditional 1-9 digit 9x9 format, being 3x3 array of 3x3 individual cells in typical ~~Sudokus~~ ~~Sudokii~~ ~~Sudoka~~ puzzles, but can be irregularly domained instead as long as the domains still have nine cells. In a 12-digit that's often 3x4 cells in each domain, arrayed 4x3 (or 4x3 arrayed 3x4) to make a 12x12 full grid but can be 2x6 6x2s (or <=>) or of irregular, but still equally-sized, subdivisions. ("Killer" variations typically augment the row, column and domain parities with a 'fourth dimension' of *unequally*-sized irregular domains (no larger than any other domain, containing a *maximum* of one of each digit, but possibly zero of some) labelled as having a stated sum total within (more than or equal to m*(m-1)/2 for m cells in that given sum-zone, assuming the lowest digit is 1, and less than or equal to (n*(n-1)/2)-((n-m)*(n-m-1)/2), if n digits are being used as unique symbols throughout the whole grid), but that's generally in leiu of *all* pre-existing clue digits, using Kakuro-like calculations to break ground on the puzzle's answer.)

Realistically, therefore, the comic must be 1x2 domains of 2x1 cells. Or the other way round. Although it's not obvious from the line-weighting which it might be. As each subdivision is the same as the row-grouping or column-grouping it could effectively be just a 'simpler' puzzle that abandons or considers redundant domains *other* than the basic rows and columns, given that each possible domain-type would be congruent with one or other of the two implicit groupings. However, it definitely could *not* be an "X" variant of the puzzle type (repetition dissallowed across the major diagonals, as well as across rows and columns), otherwise it reverts to being impossible again...

However, none of what I've just said is particularly entertaining, so please feel free to ignore it and instead try the following Unary Sudoku.... (Hint: its major diagonals are also valid domains to solve!)

□

178.98.31.27 15:07, 24 June 2013 (UTC)

## Major update[edit]

I think we even can discus PIxPI grids here at discussion page, but the explain should be simple as possible. Please help on that bad remaining language. AND: Since Randall is from the US we have AE (American English) here.--Dgbrt (talk) 21:19, 1 July 2013 (UTC)

I think this description of the alt text is a little inadequate. The reason the puzzle is 'medium difficulty' is because any given puzzle in a binary sudoku is going to be essentially the same... you'll either have a one and a zero, two ones, or two zeroes since those are the only ways to ensure a unique solution. So all puzzles have the same difficulty which is why it is 'medium'.--108.162.238.162 08:28, 28 July 2015 (UTC)

- I don't follow your sentence "you'll either have a one and a zero, two ones, or two zeroes", please can you try and explain more clearly. Essentially though, there are two possible solutions to the puzzle:

0 1 or alternatively 1 0 1 0 0 1

- The only thing which can make the difficultly change is how many cells are pre-filled. If 0 are filled then either both solutions are correct, or it is impossible to know which solution is correct. If 1 cell is filled, then it is easy to complete the rest of the grid. If 2 are filled, then it is even easier to complete, and easier again with 3 filled. --Pudder (talk) 11:01, 28 July 2015 (UTC)