I was pleasantly surprised to find out this morning, after getting back from the National Organic Symposium at Duke, that I had made my second pitstop on Wei-Hwa's Google Puzzle Gadget. Since a lot of the development of the ssudoku idea was inspired by the Udoku work from Wei-Hwa, and in fact all ssudoku have been written by Wei-Hwa using a variant of his Udoku engine, he should get a lot of credit for putting this idea of mine into practice. Just as with our earlier tabletop sudoku, I'm excited to see what other puzzles we can create when we have the time to work together. [Edit: The difficulty from the random set is not very well metered. Some of the earlier puzzles (2, 3) are the hardest of the first thirty-five I've tried. If you get discouraged, just skip ahead some puzzles and retry them.]

As history, Udoku was a puzzle type Wei-Hwa brought to the WSC which presented NxN cells with (N+1) boxes with (N-1) cells and 1 box with just a single lone cell. While all rows and columns would still be filled by 1 to N as in a sudoku, the compromised box sizes meant that for the larger boxes, which were one cell smaller than N, you had to fill in all but one of the numbers from 1 to N. This made solving the puzzle quite difficult as hidden singles (the route I'd argue most every sudoku solver uses first, even if they don't know the name) were no longer good inside of boxes. That, and many of the examples Wei-Hwa brought were impossibly hard even if they looked awesome (particularly his 10x10 puzzles that contained within them what looked like a normal sudoku). In general, you were best served to look for naked singles, and target rows and columns but not boxes. There were some properties on where digits would be missing that you could use to make some unique headway on them, but in general they were challenging to solve as scanning for hidden singles was rarely good.

The solving characteristics of Udoku were similar to another variant I'd seen in GAMES called Blackout Sudoku last fall. I wondered, as I couldn't sleep one night, if I could come up with a puzzle that would compromise most naked singles but maintain all hidden singles. There aren't many sudoku variants that specifically compromise some but not all solving methods, but my goal was to compromise nakeds and not hiddens which I'd never seen before. I found that a good way to accomplish this would be to use (n-1) boxes with (n+1) cells and a single lone cell, an obvious counterpoint to Wei-Hwa's udoku set-up, and therefore each box always has all the numbers once, and a single number twice. Ssudoku was the result. As the name of udoku came from dropping the s as boxes are missing one entry, I made the much more awkward suggestion of Ssudoku as the name for this type with a repeated entry. I find them interesting- indeed the hidden/naked properties are as I expected- and am very glad Wei-Hwa put together so many of these puzzles so that people can try out this idea. Head over to his gadget and ssample ssome ssudoku!

As history, Udoku was a puzzle type Wei-Hwa brought to the WSC which presented NxN cells with (N+1) boxes with (N-1) cells and 1 box with just a single lone cell. While all rows and columns would still be filled by 1 to N as in a sudoku, the compromised box sizes meant that for the larger boxes, which were one cell smaller than N, you had to fill in all but one of the numbers from 1 to N. This made solving the puzzle quite difficult as hidden singles (the route I'd argue most every sudoku solver uses first, even if they don't know the name) were no longer good inside of boxes. That, and many of the examples Wei-Hwa brought were impossibly hard even if they looked awesome (particularly his 10x10 puzzles that contained within them what looked like a normal sudoku). In general, you were best served to look for naked singles, and target rows and columns but not boxes. There were some properties on where digits would be missing that you could use to make some unique headway on them, but in general they were challenging to solve as scanning for hidden singles was rarely good.

The solving characteristics of Udoku were similar to another variant I'd seen in GAMES called Blackout Sudoku last fall. I wondered, as I couldn't sleep one night, if I could come up with a puzzle that would compromise most naked singles but maintain all hidden singles. There aren't many sudoku variants that specifically compromise some but not all solving methods, but my goal was to compromise nakeds and not hiddens which I'd never seen before. I found that a good way to accomplish this would be to use (n-1) boxes with (n+1) cells and a single lone cell, an obvious counterpoint to Wei-Hwa's udoku set-up, and therefore each box always has all the numbers once, and a single number twice. Ssudoku was the result. As the name of udoku came from dropping the s as boxes are missing one entry, I made the much more awkward suggestion of Ssudoku as the name for this type with a repeated entry. I find them interesting- indeed the hidden/naked properties are as I expected- and am very glad Wei-Hwa put together so many of these puzzles so that people can try out this idea. Head over to his gadget and ssample ssome ssudoku!

4 comments | Leave a comment