Kelly, I generally solve sudokus using a multiple pass method:
1) Start with the number 1, and exclude all the places where a 1 can be in each of the nine sections. Fill in the places where its the only option. Do this for all 9 numbers, over and over again, until you go through all of 1-9 and don't fill in any more places. 2) For every empty position on the board, see if only one number is allowed in that location by looking in the row, column, and 3x3 box the empty position is in. If you find a number this way, go back to step one and start over. 3) Look for and mark down any pairs, that is any set of two boxes that can only contain the same two numbers. For example a row with five empty boxes and two boxes that can either 5 or 7. If you find any pairs, go back to step one and start over. About half of what I'd consider 'hard' grade puzzles can be solved with the above three methods. The reason you start over a lot is because you don't want to go past step three unless you absolutely have to: 4) Set breaker pass. Mark out all of the allowed values for each empty square. For each of the three constraint classes (row, column, and 3x3 box) search for positive and negative disjoint sets. If you find one, mark off the excluded values and, go back to step one. Probably around 90% of 'hard' grade puzzles can be solved using the first four methods. 5) Chain breaker pass. Search the puzzle for pairs or triples that force other values elsewhere on the board. Keep forcing values until you find a contradiction. This is a generalization of the ywing strategy and yields results quicker in my opinion; I still have difficulty finding ywing structures easily. If you find a contradiction, mark out the excluded values and go back to step one. Pretty much all hard grade puzzles can be solved using these first five methods. For extreme or ridiculous puzzles, there's not a lot left you can do, but here's the next set of options: 6) Grid search pass. This step is fairly quick but usually pointless, because if you've done step 1 properly it almost never occurs in the wild. I've only used this as to make forward progress perhaps twice out of five hundred puzzles. Search for N rows containing only N instances of a number, then use that to exclude other column possibilities. Search for N columns containing only N instances of a number, then use that to exclude other row possibilities. I've only ever seen this work for N=2 and N=3, so I wouldn't bother looking at values 4 and higher. If you find something, go back to step 1. 7) Stare at it until you see something you missed in a previous step. If you get here, odds are good it's because you haven't done the previous steps perfectly. If you find something, go back to step 1. 8) Guess values and work forward, looking for contradictions. At this point, you're not really using logic anymore, and you've probably picked puzzles that are near the complexity limit and not worth messing with. Dennis Towne Alter Aeon MUD http://www.alteraeon.com On Wed, Jan 9, 2013 at 5:30 PM, Kelly Sapergia <[email protected]> wrote: > Hi, > > I've been struggling with solving Sudoku puzzles for years, and still don't > have a clue as to what I'm doing. I got a free book NBP was distributing for > some time, simply called "Sudoku Puzzles", that had ten supposedly easy > puzzles from a book called "Sudoku Easy" by Will Shortz, complete with a > tactile representation of a board with squares. I honestly don't know how > you come up with the right solution, as looking at the ones in the book make > absolutely no sense to me. I know these are logic puzzles, but so far my > logic doesn't seem to be working, and I end up putting random numbers in the > blanks. I've also got the free Sudoku game from Spoonbill Software, and even > an iPod game called Sudoku4All, but I'm not sure those are helping either. > My mom, who usually can figure out these things, thinks that a board with > pegs would be better as I'd have a tactile representation, so she adapted a > small set for me, but I haven't had much luck with it. As a result, I've > pretty much given up, but if anyone has any hints on solving these puzzles, > please let me know. > > Thanks. > > Yours Sincerely, > Kelly John Sapergia --- Gamers mailing list __ [email protected] If you want to leave the list, send E-mail to [email protected]. You can make changes or update your subscription via the web, at http://mail.audyssey.org/mailman/listinfo/gamers_audyssey.org. All messages are archived and can be searched and read at http://www.mail-archive.com/[email protected]. If you have any questions or concerns regarding the management of the list, please send E-mail to [email protected].
