|
||||||||||||||||||||
|
||||||||||||||||||||
|
||||||||||||||||||||
Farewell to Ariadne's thread?THE TIME has come to wave goodbye to Diabolical sudoku puzzles that appear to require guessing - at least from the mainstream that appear in newspapers and books. They have served their purpose and will now be available only to solvers who require sources for developing new logic strategies.There was some method behind my madness when I introduced these puzzles just eight months ago. Perhaps I'd better explain. Back in January of 2005 there was only one person in the world outside Japan producing sudoku in any quantity, and that was Wayne Gould of Pappocom. His business was (and still is) selling a sudoku program off the web and he also produced puzzles for newspapers, such as the London Times. He had spent six years or so developing his program to compile sudoku and grade their difficulty. With newspaper deadlines hanging over me, I could afford to spend only one week on my program's development. When it came to setting grades of difficulty I needed to compare my levels with those of the competition and at the time that was only the Times. So I used Gould's program to compare his grades of sudoku to my grades. However, I soon discovered that many of my puzzles were just too difficult for the Pappocom program to solve. My own solver worked through the basic logic strategies, but would sometimes be forced to take a guess and then resort to the Ariadne's thread method of finding a solution. Logical solutions My own "manual" analysis of some of these puzzles revealed that there was indeed a logical solution, but the basic routines I had programmed were just not up to the job. I figured that the Pappocom program was failing for the same reason. It was at this point that Pappocom and I differed in the grading of our puzzles. Gould's standpoint was "if my logic can't solve it, then it's either too difficult or unfair" while mine was "I know that some of these puzzles will require guessing, but for many there will be a logic strategy that will work". I felt that if the sudoku was valid and had a unique solution why should I scrap it simply because my program's logic wasn't up to the job? My decision to publish these few Diabolical puzzles turned out to be controversial. Some solvers who had thought that the Times's Fiendish puzzles were the ultimate, were faced with a new level of difficulty in the Daily Telegraph's Diabolical sudoku. And it didn't go down well at all with them. Despite the fact that most of these puzzles could be solved with logic, the forums began to receive complaining messages about "unfair" and "unsolvable" sudoku. In fact, less than three per cent of my sudoku output was at a Diabolical level, and of those puzzles, two in every three could be solved with logic without resorting to trial and error. Open minds Not everybody complained, though. Many, extremely bright people with open minds, sharpened their pencils and set about the task of finding the logic schemes that might solve them without resorting to Ariadne's thread. And it worked. Now, just eight months on, strategies like remote pairs, swordfish and X-wings have been developed by enthusiasts and can be used in solving Diabolical sudoku. These complex schemes and more have been incorporated into the grading software for my puzzles and so I can say with confidence that all of my Diabolical puzzles can be solved with logic alone.* However, the story should not end there, because there is still more to learn from sudoku. Out of my so-called impossible puzzles, new schemes have been developed, so these puzzles will continue to be published here at sudoku.org.uk. They will be in a format that allows for easy import into our Sudokulist software and who knows what new schemes will emerge in the future. ![]() *Please don't write to tell me that you've discovered a puzzle that cannot be solved with logic alone. There will be some sudoku puzzles out there that were created before I made this decision and will be for some time to come. ![]() |
|
||
|