![]() |
![]() |
|||||
|
Sudoku GuessworkCall it what you will backtracking, guessing, labyrinth or trial and error, Sudoku players are split as to whether this is a good technique to use. The strategy is simplicity itself and can be used on its own to solve every Sudoku puzzle. You just choose one available possibility for a square, remember what you have chosen and then proceed to solve the remaining squares. If you made the wrong choice then the puzzle will become un-solvable. A Sudoku puzzle becomes un-solvable because an incorrect allocation has been made. This shows up when two squares in the same region have the same number (breaking the rules) or as a square in a region unable to take any number at all (again breaking the rules). If this happens you have to back-track all the allocations until you come back to the last square you guessed, and then choose one of the other alternatives, you have now knocked out one of the alternatives. If you are solving with pencil and paper this technique often becomes a real mess of marks and rubbings out as choices are made and unmade. With the help of a program such as Sudoku Dragon this is not a problem as it lets you easily undo and redo previous square allocations, it also lets you put a note on a square to indicate where and what guess you made. For most puzzles the standard set of strategies enable you to make a definitive choice without needing to guess. Many players hate having to make an allocation
on the off chance it is correct, they want to work it out entirely be logic as to what goes where for sure.
[For more dicsussion on this topic see "Farewell to guessing" One of the other reasons that guessing may be best is that it makes solving a puzzle quicker than looking for the more esoteric solution strategies. By the time you have looked for x-wings, swordfish, xy-wings and alternate pairs you might just as well have solved a square by using trial and error. Ensuring that a puzzle can be solved using strategies and does not involve requiring a guess is hard for puzzle generating software. If you look on the Internet you will find that most puzzle generators do use a lot of trial and error to work out if a puzzle is solvable, they do not follow the techniques that a human solver would use. To see how you can solve a whole Sudoku puzzle just using trial and error take a look at this external web site that has an animated example : Animated Sudoku solver
With trial and error you choose a square with only two possibilities as then the chances are just 50:50 of the guess being correct. Look at the highlighted green square Dd, it can take only a 2 or a 6. So let's try a 2. This immediately opens things up a lot, 6 must then go in Fe (only choice in region Dd) a 2 in He (only square in column e) and a 6 in Di (only choice in row D).
To download this puzzle
Sometimes it will take many steps to prove one option was an incorrect guess and that can make it a slow and laborious process. Solving without guessingSo if you are to solve this puzzle without guessing, how is it to be done? This is indeed a hard puzzle. If you look for a square to solve with the simpler strategies 'only choice', 'single possibility' and 'only square' nothing is evident. If you extend this to looking for 'sub group exclusions', 'twins', 'triplets' and 'chains' still no luck. This has all taken quite some time to do. Now on to the more esoteric solution strategies. An X-Wing is present, if you look at where '2's are possible you will spot a box in Row D and Row I, however this only knocks out a '2' occurring in Hd and does not allow a square to be solved. So what next? The last recourse is to the more advanced solution strategies. The key is to use the alternate pairs coloring of where '6's can occur. If you do this you will find that quite a few squares that looked as if they could take a '6' can not. So square Be can not be a '6' so it must be a '5' and the '6' in region Ad must go in Ae. Once this is done all the remaining squares can now be easily solved. So who said solving Sudoku puzzles is easy? Sometimes making a guess can save a lot of time. Mazes
The most familiar example is finding your way through a maze. You can always escape from a maze as long as you run through all the alternative paths in a consistent, logical fashion. If turning right leads to a dead end or leads back to somewhere you have been before, retrace steps and try straight ahead, if that fails try left. In this way all of the possible alternative routes are tried out in turn - until you find the path that lets you escape the maze. Unless you are careful you can end up literally going round in circles. To stop this you need to record where you have been and what route you took. This can be done by unravelling a thread; if you come across the thread again later on you have done a circuit and must try another route (no longer following the thread). More practical alternatives are a bag of rice or a stick of chalk, anything that can mark where you have tried so far. If you choose rice, put a grain at each junction to mark the turning that you took, if you find a grain ahead of you then you've repeated yourself and must try an alternative path. Computers can use this technique to solve any maze problem with 'brute force'. The algorithm just runs through each possibility in turn, e.g. 'turn left at each junction until reach dead-end backtrack and then instead of turning left, turn right'. The analogy with Sudoku is strong as 'trial and error' is using very much the same idea. When navigating through mazes an experienced human can make educated guesses on the most promising path rather than slavishly trying each option in turn. Similarly a skilled Sudoku player can dismiss unpromising options to reach the solution much quicker.
The LabyrinthIn ancient Greek myth the Minotaur
Copyright © 2005-2009 Sudoku Dragon |