Main Page - Back |

## What's NewFrom sudokuwiki.org, the puzzle solver's site |

21st April 2016

I have been working on improving the display of strong links in the solvers. Before I release this I'd appreciate comments. On the current versions the strong link display is tiresome to keep turning on and its very transient. My improvement is currently on the Sudoku X solver and there is a beta version to test. You will see a new tick box under "show candidates". Turn that on while stepping through the solver. Use in combination with the String Link tick boxes under the list. The links update when you edit a cell - if "show strong links" is ticked. Thoughts appreciated.

26th March 2016

Upgraded the Windoku Solver to take advantage of the full nine windows (not just the four in the puzzle definition). Confused? Check out this page. Edit 27th March: I have added in proper Pointing Pairs and Line/Box Reduction for windows. I am recalibrating the grades as a result and will provide new examples in the solver example list.

I have been working on improving the display of strong links in the solvers. Before I release this I'd appreciate comments. On the current versions the strong link display is tiresome to keep turning on and its very transient. My improvement is currently on the Sudoku X solver and there is a beta version to test. You will see a new tick box under "show candidates". Turn that on while stepping through the solver. Use in combination with the String Link tick boxes under the list. The links update when you edit a cell - if "show strong links" is ticked. Thoughts appreciated.

26th March 2016

Upgraded the Windoku Solver to take advantage of the full nine windows (not just the four in the puzzle definition). Confused? Check out this page. Edit 27th March: I have added in proper Pointing Pairs and Line/Box Reduction for windows. I am recalibrating the grades as a result and will provide new examples in the solver example list.

28th November 2015

Upgraded the KenKen Solver to match the Killer solver, ie two years of incremental changes all applied at once. Some gains on solving the hardest grade of KenKen.

12th November 2015

I couldn't have a Sudoku X solver without considering it's sister variant Windoku - and so finally it is released! Whereas Sudoku X has two additional constraints the next step us is four (if one likes symmetry). Windoku comes from the four new boxes spread like windows on a child's drawing of a house (one presumes). Most of the Sudoku strategies apply, just be careful when using Uniqueness strategies - I only apply to cells outside the 'windows'. Next I'd like to offer a Daily Windoku puzzle.

29th August 2015

I've added a check at the start of the strategy list to remove candidates in a Killer 'dog-leg'. The

Killer Cage Convention states that no cage can contain duplicate numbers. Although the solver does get those candidates eventually it's much better to remove them early on when they are obvious. An example is 6 in cell C9 in this example. Thanks to Jacqueline Elemans for giving me this idea. Please reload the Killer Solver until it says version 2.03.

2nd July 2015

After a lot of testing to see exactly how useful Bi-value Universal Grave is I've decided to re-instate it into the solver. The best place is just before XY-Chains because every instance of BUG can be solved by an XY-Chain (is my conjecture) but it's a simpler pattern to spot. Any more examples appreciated.

15th February 2015

Merged Simple Coloring and 3D Medusa Rules 4 and 5 together since it has been observed that they are both off-chain eliminations and the old Rule 5 was a super-set of Rule 4. Nice simplification.

Upgraded the KenKen Solver to match the Killer solver, ie two years of incremental changes all applied at once. Some gains on solving the hardest grade of KenKen.

12th November 2015

I couldn't have a Sudoku X solver without considering it's sister variant Windoku - and so finally it is released! Whereas Sudoku X has two additional constraints the next step us is four (if one likes symmetry). Windoku comes from the four new boxes spread like windows on a child's drawing of a house (one presumes). Most of the Sudoku strategies apply, just be careful when using Uniqueness strategies - I only apply to cells outside the 'windows'. Next I'd like to offer a Daily Windoku puzzle.

29th August 2015

I've added a check at the start of the strategy list to remove candidates in a Killer 'dog-leg'. The

Killer Cage Convention states that no cage can contain duplicate numbers. Although the solver does get those candidates eventually it's much better to remove them early on when they are obvious. An example is 6 in cell C9 in this example. Thanks to Jacqueline Elemans for giving me this idea. Please reload the Killer Solver until it says version 2.03.

2nd July 2015

After a lot of testing to see exactly how useful Bi-value Universal Grave is I've decided to re-instate it into the solver. The best place is just before XY-Chains because every instance of BUG can be solved by an XY-Chain (is my conjecture) but it's a simpler pattern to spot. Any more examples appreciated.

15th February 2015

Merged Simple Coloring and 3D Medusa Rules 4 and 5 together since it has been observed that they are both off-chain eliminations and the old Rule 5 was a super-set of Rule 4. Nice simplification.

27th December 2014

Happy Christmas All. On the strategy pages for Tough and some Diabolical strategies (eg X-Wings I've added what I call Exemplars. These are puzzles that require the strategy pertaining to that page AND exclude all other strategies. In most cases the hardest additional step is one Naked Pair. The purpose of this to list the easiest puzzles that contain an example of the strategy making it a useful practice puzzle. In theory there is no "best" exemplar because where there is one example an alternative with two examples might yet be found. I don't like to publish these sorts of puzzles in a newspaper since they make poor puzzles - trivial all the way apart from one complex bottleneck. I try and filter them out. But for this purpose I'll be keeping an eye on the puzzles I create. If you find a better one, let me know!

12th December 2014

I have now added a page showing all the Jigsaw Shapes

9th December 2014

Got a great tip from Caleb Andrews to capture more Cage Unit Overlaps in Killer Puzzles. Coded up and working now.

23rd August 2014

Updated the solvers to include a clue count, plus on Sudoku and Sudoku X, you can highlight bi-value cells, as suggested by JcB in Australia.

6th August 2014

The 15 Up puzzle I wrote a generator for, has come out as an app. Android fans, check it out.

15th July 2014 - Responsive Design

I'm currently attempting to re-write the HTML and CSS of this web site into a format known as 'Responsive Design', which means it should appear with a different and more suitable layout when viewed on a cell device. Might be some formatting issues while I progress this.

6th July 2014 - Core Engine Update

Updated the Core Engine for all solvers. Thanks to Kyle for a very tough Sudoku X example I've done two things: Ensure that diagonals are fully used for eliminations and Group Cells when detecting Alternating Inference Chains and secondly, allowing back-to-back Group Cells. This means that two or more Group Cells could be chained together for a chain. Going to do some further testing to see what the impact of this has, could be little, could be a lot. Not sure yet. Anyway, Kyle's excellent example now solves.

Last month I completed the new beta Core Engine replacement which uses Breadth First Searching on AICs. On tests of large libraries of extreme Sudoku puzzles, eg Ruud's 50k, I got a five fold increase in speed. However, I'm not happy with individual times when testing the solver web-service. It also uses 1/2Mb more memory, which doesn't sound a lot, but when multiplied across many users, could be impractical. It's all to hold new data structures. So I'm not ready to release just yet. The new engine does solve fractionally more puzzles, but now have to duplicate the work in the previous paragraph to really compare.

Happy Christmas All. On the strategy pages for Tough and some Diabolical strategies (eg X-Wings I've added what I call Exemplars. These are puzzles that require the strategy pertaining to that page AND exclude all other strategies. In most cases the hardest additional step is one Naked Pair. The purpose of this to list the easiest puzzles that contain an example of the strategy making it a useful practice puzzle. In theory there is no "best" exemplar because where there is one example an alternative with two examples might yet be found. I don't like to publish these sorts of puzzles in a newspaper since they make poor puzzles - trivial all the way apart from one complex bottleneck. I try and filter them out. But for this purpose I'll be keeping an eye on the puzzles I create. If you find a better one, let me know!

12th December 2014

I have now added a page showing all the Jigsaw Shapes

9th December 2014

Got a great tip from Caleb Andrews to capture more Cage Unit Overlaps in Killer Puzzles. Coded up and working now.

23rd August 2014

Updated the solvers to include a clue count, plus on Sudoku and Sudoku X, you can highlight bi-value cells, as suggested by JcB in Australia.

6th August 2014

The 15 Up puzzle I wrote a generator for, has come out as an app. Android fans, check it out.

15th July 2014 - Responsive Design

I'm currently attempting to re-write the HTML and CSS of this web site into a format known as 'Responsive Design', which means it should appear with a different and more suitable layout when viewed on a cell device. Might be some formatting issues while I progress this.

6th July 2014 - Core Engine Update

Updated the Core Engine for all solvers. Thanks to Kyle for a very tough Sudoku X example I've done two things: Ensure that diagonals are fully used for eliminations and Group Cells when detecting Alternating Inference Chains and secondly, allowing back-to-back Group Cells. This means that two or more Group Cells could be chained together for a chain. Going to do some further testing to see what the impact of this has, could be little, could be a lot. Not sure yet. Anyway, Kyle's excellent example now solves.

Last month I completed the new beta Core Engine replacement which uses Breadth First Searching on AICs. On tests of large libraries of extreme Sudoku puzzles, eg Ruud's 50k, I got a five fold increase in speed. However, I'm not happy with individual times when testing the solver web-service. It also uses 1/2Mb more memory, which doesn't sound a lot, but when multiplied across many users, could be impractical. It's all to hold new data structures. So I'm not ready to release just yet. The new engine does solve fractionally more puzzles, but now have to duplicate the work in the previous paragraph to really compare.

6th April 2014

"The Logic of Sudoku" has Sold Out! Last copy went today to a lady in Stayton, Oregon. Many thanks, and thank-you to everyone who's bought a copy and supported the site. I hope you found it enjoyable. Good news is that there will be a second edition, but this time as an e-Book. Bad news is that I've only just started to write it ;)

30th March 2014

We are on a new server! (same ISP). I'm sure the setup need fine tuning over the next week, so any problems, drop me an email at andrew@zo.gs. I may not find issues first. This is a higher spec machine and 64bit, more space etc. Should even be faster.

20th March 2014

I'm very pleased to be hosting a paper by Kevin Gromley - Generalizing Sudoku Strategies. I've written up some of our discussion on this topic. An excellent analysis maths fans will appreciate.

5th March 2014

Solver update to Unique Rectangles. Tom Morrin pointed out that type 3B was missing the eliminations in the box. Quite right.

4th February 2014

I've updated the Str8ts Solver to version 1.08 and introduced Required Digits which makes a big difference to cracking harder puzzles. I'm indebted to Ulrich for his prodding on this point and I realize it was major gap in the solver.

The consequence of this has been to render 36 of my stock of unsolvables now solvable. I have removed any from the coming weekly puzzles but for past puzzles I am putting a note under the puzzle title (as I do with the weekly Sudoku) when flagged as solvable. Most of the remaining are helped by the added strategy so I'll have to begin the search for better unsolvables given this new benchmark. I will also have to recalibrate the grades of normal puzzles as well.

"The Logic of Sudoku" has Sold Out! Last copy went today to a lady in Stayton, Oregon. Many thanks, and thank-you to everyone who's bought a copy and supported the site. I hope you found it enjoyable. Good news is that there will be a second edition, but this time as an e-Book. Bad news is that I've only just started to write it ;)

30th March 2014

We are on a new server! (same ISP). I'm sure the setup need fine tuning over the next week, so any problems, drop me an email at andrew@zo.gs. I may not find issues first. This is a higher spec machine and 64bit, more space etc. Should even be faster.

20th March 2014

I'm very pleased to be hosting a paper by Kevin Gromley - Generalizing Sudoku Strategies. I've written up some of our discussion on this topic. An excellent analysis maths fans will appreciate.

5th March 2014

Solver update to Unique Rectangles. Tom Morrin pointed out that type 3B was missing the eliminations in the box. Quite right.

4th February 2014

I've updated the Str8ts Solver to version 1.08 and introduced Required Digits which makes a big difference to cracking harder puzzles. I'm indebted to Ulrich for his prodding on this point and I realize it was major gap in the solver.

The consequence of this has been to render 36 of my stock of unsolvables now solvable. I have removed any from the coming weekly puzzles but for past puzzles I am putting a note under the puzzle title (as I do with the weekly Sudoku) when flagged as solvable. Most of the remaining are helped by the added strategy so I'll have to begin the search for better unsolvables given this new benchmark. I will also have to recalibrate the grades of normal puzzles as well.

15th January 2014

Expanded the use of Innies and Outies and Cage Splitting by using Locked Sets to split cages. See docs for examples; Added some more checks to the Killer Solver popup designer. It should now correctly post an alert if the clues are not in the right cells. (They must always be in the top left most cell of the cage).

1st January 2014

Wishing you all a Happy New Year!

Added some reporting to the grader/solve path when the sudoku puzzle has errors.

19th November 2013

Added Unique Rectangles Type 3b to the solver. Many thanks to Hervé Gérard in France for the example. I've also updated how the cells are coloued when the results are returned for Types 2 and 3.

17th November 2013

Starting this November there are brand new Killer Sudoku Puzzle Packs, fifty of each grade and one mixed pack of five different grades. Hoping to add more puzzles in addition to Sudoku and Killer Sudoku for the new year. I have recalibrated the Killer Sudoku puzzles and starting tomorrow, these will replace all the Daily Killers and feed puzzles. Recalibration means the bands that they fall into will have moved.

8th October 2013

Re-vamp of Aligned Pair Exclusion to find large ALSs and non-aligned versions. These changes greatly increases the chance of finding an APE although the simplest instances overlap with Y-Wings and XYZ/WXYZ in some instances.

7th October 2013

There will be a Letterlicious iPhone/iPad App released very soon! Please have a look at this page for screen shots and further information. If you want to get your hands on a pre-release version and test it out - you can!.

Expanded the use of Innies and Outies and Cage Splitting by using Locked Sets to split cages. See docs for examples; Added some more checks to the Killer Solver popup designer. It should now correctly post an alert if the clues are not in the right cells. (They must always be in the top left most cell of the cage).

1st January 2014

Wishing you all a Happy New Year!

Added some reporting to the grader/solve path when the sudoku puzzle has errors.

19th November 2013

Added Unique Rectangles Type 3b to the solver. Many thanks to Hervé Gérard in France for the example. I've also updated how the cells are coloued when the results are returned for Types 2 and 3.

17th November 2013

Starting this November there are brand new Killer Sudoku Puzzle Packs, fifty of each grade and one mixed pack of five different grades. Hoping to add more puzzles in addition to Sudoku and Killer Sudoku for the new year. I have recalibrated the Killer Sudoku puzzles and starting tomorrow, these will replace all the Daily Killers and feed puzzles. Recalibration means the bands that they fall into will have moved.

8th October 2013

Re-vamp of Aligned Pair Exclusion to find large ALSs and non-aligned versions. These changes greatly increases the chance of finding an APE although the simplest instances overlap with Y-Wings and XYZ/WXYZ in some instances.

7th October 2013

There will be a Letterlicious iPhone/iPad App released very soon! Please have a look at this page for screen shots and further information. If you want to get your hands on a pre-release version and test it out - you can!.

26th September 2013

There's been a little bit of confusion over how the solver, grader and solve path handle Pairs and Triples. Sometimes the solve path would produce a slightly different set of eliminations from the step-by-step solver. I've smoothed this out now by imposing some rules on Naked/Hidden Pairs and Triples. The solver uses javascript for these strategies as they are common eliminators and relatively easy to code, but it's quite old code, going back to the first years of the solver. Only Quads and beyond call the server for some hard-core number crunching. So discrepancies arose over time. One was how to handle knock-on effects of one elimination creating new eliminations in the same strategy. I dislike this as it is confusing to work out what’s going on in one step and I'd done something about this some years ago, eg with Naked Pairs, for instance, but not all the 'sets' strategies had been tidied up. So, now, no more knock-on effects. However, within the same step, the solver will report all instances of pairs and triples -**provided they are independent of each other**. This slightly speeds up the solver but removing clicks and shouldn’t be visually cluttering like the knock-on eliminations. More complex strategies continue to report only the first instance of the first strategy used, as before.

5th September 2013

22th July 2013

Sudoku solver now updated to 1.97. Some of the analysis I've been working on - and many of the strategies in this solver - require looking at chains. I've added a tool under the strategy list to show Strong Links - the basic component of chains. Over the years I've had many requests to build a feature like this. These might help anyone stuck on a difficult Sudoku to work out the next step without asking the solver. Currently it does not interfere with the step-by-step solver but it will clear any drawing on the board from a step.

There's been a little bit of confusion over how the solver, grader and solve path handle Pairs and Triples. Sometimes the solve path would produce a slightly different set of eliminations from the step-by-step solver. I've smoothed this out now by imposing some rules on Naked/Hidden Pairs and Triples. The solver uses javascript for these strategies as they are common eliminators and relatively easy to code, but it's quite old code, going back to the first years of the solver. Only Quads and beyond call the server for some hard-core number crunching. So discrepancies arose over time. One was how to handle knock-on effects of one elimination creating new eliminations in the same strategy. I dislike this as it is confusing to work out what’s going on in one step and I'd done something about this some years ago, eg with Naked Pairs, for instance, but not all the 'sets' strategies had been tidied up. So, now, no more knock-on effects. However, within the same step, the solver will report all instances of pairs and triples -

5th September 2013

I'm pleased to see released a new Sudoku puzzle player (Sue Doku) for iOS platforms, designed by James Forrest and Michael McWatters, containing four grades of puzzles from our stock. This has a glossy, minimalist and professional look and feel and is just 99c on the Apple store. Do check it out.

22th July 2013

Sudoku solver now updated to 1.97. Some of the analysis I've been working on - and many of the strategies in this solver - require looking at chains. I've added a tool under the strategy list to show Strong Links - the basic component of chains. Over the years I've had many requests to build a feature like this. These might help anyone stuck on a difficult Sudoku to work out the next step without asking the solver. Currently it does not interfere with the step-by-step solver but it will clear any drawing on the board from a step.

20th July 2013

Sudoku solver now on 1.96. Greatly expanded WXYZ-Wings detection. Finally got round to incorporating insights by SudoNova and StrmCkr. See article for more.

3rd June 2013

It's not often I get a brand new strategy that goes to the top of the solver list, or at least immediately after the basic strategies. Brain Hobson has been very kind to send me the idea of doubling up one box/one line Intersection Removals. So, for Jigsaw Sudoku, we now have Double Pointing Pairs and Double Line/Box Reduction and it works brilliantly! Sadly it cannot be used in normal Sudoku since this requires the odd shapes present in Jigsaw, but it is elegent and easy to spot.

28th May 2013

A good example from Hans Berkenkamp has allowed me to improve the Killer Solver again, this time looking at pseudo-cages created by Innies and Outies. The solver now considers the way each available combination can fit in the same way normal cages are checked. Had a good impact only previously unsolvable puzzles.

25th May 2013

Fixed some bugs in the Killer solver and improved the text returned from Cage Combinations. Thanks to Robin Thornton for spotting the example. I also made sure the 'last remaining cell' in a cage was correctly being identified and now shows on the Hidden Singles stage.

22th May 2013

Over the weekend I've been looking again at the Forcing Chains family and improved the detection. Last night some of those improvements created a bug (in the formating of the output string) which stopped the solver. Fixed today. It was quite a difficult thing to test as I need examples of all cases. Documentation updates to follow.

17th May 2013

A new 3D Medusa rule has been discovered! Languishing in the comments and known to you all, I'm sure, it finally clicked for me and its now in the solver. Many thanks to Anton Delprado for posting his insight.

Sudoku solver now on 1.96. Greatly expanded WXYZ-Wings detection. Finally got round to incorporating insights by SudoNova and StrmCkr. See article for more.

3rd June 2013

It's not often I get a brand new strategy that goes to the top of the solver list, or at least immediately after the basic strategies. Brain Hobson has been very kind to send me the idea of doubling up one box/one line Intersection Removals. So, for Jigsaw Sudoku, we now have Double Pointing Pairs and Double Line/Box Reduction and it works brilliantly! Sadly it cannot be used in normal Sudoku since this requires the odd shapes present in Jigsaw, but it is elegent and easy to spot.

28th May 2013

A good example from Hans Berkenkamp has allowed me to improve the Killer Solver again, this time looking at pseudo-cages created by Innies and Outies. The solver now considers the way each available combination can fit in the same way normal cages are checked. Had a good impact only previously unsolvable puzzles.

25th May 2013

Fixed some bugs in the Killer solver and improved the text returned from Cage Combinations. Thanks to Robin Thornton for spotting the example. I also made sure the 'last remaining cell' in a cage was correctly being identified and now shows on the Hidden Singles stage.

22th May 2013

Over the weekend I've been looking again at the Forcing Chains family and improved the detection. Last night some of those improvements created a bug (in the formating of the output string) which stopped the solver. Fixed today. It was quite a difficult thing to test as I need examples of all cases. Documentation updates to follow.

17th May 2013

A new 3D Medusa rule has been discovered! Languishing in the comments and known to you all, I'm sure, it finally clicked for me and its now in the solver. Many thanks to Anton Delprado for posting his insight.

6th May 2013

I've changed the article comment system to allow me to move comments into a separate "talk" section. Comments in talk pertain the text itself (corrections and so on) rather than the subject of the article. I am always very grateful for corrections but as some comment lists are quite long and the correction has been made, I wanted to keep the credits, but push these into a new section.

6th April 2013

I'd like to take a swipe at Royal Mail for shafting its customers - again. The price of a small packet airmailed overseas has risen from £4.16 in 2011 to £5.54 (+33%) in 2012 to £7.20 (+30%) in April 2013. Disgusting.

4th April 2013

Been a fair few crashes since the last update but when I've tried the puzzles myself its been fine. Howeverm today I've found an instance that reproducable - Sue-de-Coq didnt have enough temporary space. I'm hoping the current version is stable now.

28th March 2013

Reloaded the Letterlicious puzzles (English version) with new stock. I've tweaked the generator to stop clusters of the same letter (usually vowels) appearing in the tray. The rule is - a letter can't be added if it is already on the board in the three cells above or the one cell to the left. This will make the puzzle slightly harder in my opinion. First of the new series tomorrow.

20th March 2013

Big update to the Sudoku, Sudoku X and Jigsaw solvers (now v1.91) - i've added a new yellow button called Solve Path. This attempts to solve the currently loaded puzzle and place the whole set of results in a new window or tab. This feature has been requested by a number of people and it's nice to get it off the job queue. Sudoku X and Jigsaw also get a Grader button for the first time.

(I've tested it in all the major browsers - all will ask you if you want to allow "pop-ups" and you should be able to allow it for just this site - without compromising your security elsewhere. The only exception is Google Chrome - which doesn't ask. You can permit this site by following these instructions. When you enter the domain exception put the full "http://www.sudokuwiki.org".)

Because most of the interesting puzzles are extremes and unsolvables, these can be quite a hit on the server and I dont know how to throttle the CPU. So, to prevent overloading (and bots) there is a 5 second timer between requests for the 'Grader'/'Solve Path'.

Please let me know if you find any bugs - internally I've had to recode all the strategy message to be in HTML instead of plain text. I've also put the strategy name as a name before the results as it is with the Apps.

4th January 2013

Happy new Year everyone. I fixed a problem with the "5 Daily Puzzle" app and "Sun Puzzles" app - both read a fetching script to pull in today's Killer puzzles. Unfortunately, when I loaded the puzzles for 2013 I didnt tell the script the so it was rejecting date requests in 2013. Doh. Sorry if you couldn't play Killers but glad it is easy to fix.

I'm looking for testers for the Sudoku Solver for Android. If you have a smartphone and you want to try the solver I'd like to hear from you. If you have a tablet - give me a day or so, I'm working on the larger tablet layouts at the moment.

I've changed the article comment system to allow me to move comments into a separate "talk" section. Comments in talk pertain the text itself (corrections and so on) rather than the subject of the article. I am always very grateful for corrections but as some comment lists are quite long and the correction has been made, I wanted to keep the credits, but push these into a new section.

6th April 2013

I'd like to take a swipe at Royal Mail for shafting its customers - again. The price of a small packet airmailed overseas has risen from £4.16 in 2011 to £5.54 (+33%) in 2012 to £7.20 (+30%) in April 2013. Disgusting.

4th April 2013

Been a fair few crashes since the last update but when I've tried the puzzles myself its been fine. Howeverm today I've found an instance that reproducable - Sue-de-Coq didnt have enough temporary space. I'm hoping the current version is stable now.

28th March 2013

Reloaded the Letterlicious puzzles (English version) with new stock. I've tweaked the generator to stop clusters of the same letter (usually vowels) appearing in the tray. The rule is - a letter can't be added if it is already on the board in the three cells above or the one cell to the left. This will make the puzzle slightly harder in my opinion. First of the new series tomorrow.

20th March 2013

Big update to the Sudoku, Sudoku X and Jigsaw solvers (now v1.91) - i've added a new yellow button called Solve Path. This attempts to solve the currently loaded puzzle and place the whole set of results in a new window or tab. This feature has been requested by a number of people and it's nice to get it off the job queue. Sudoku X and Jigsaw also get a Grader button for the first time.

(I've tested it in all the major browsers - all will ask you if you want to allow "pop-ups" and you should be able to allow it for just this site - without compromising your security elsewhere. The only exception is Google Chrome - which doesn't ask. You can permit this site by following these instructions. When you enter the domain exception put the full "http://www.sudokuwiki.org".)

Because most of the interesting puzzles are extremes and unsolvables, these can be quite a hit on the server and I dont know how to throttle the CPU. So, to prevent overloading (and bots) there is a 5 second timer between requests for the 'Grader'/'Solve Path'.

Please let me know if you find any bugs - internally I've had to recode all the strategy message to be in HTML instead of plain text. I've also put the strategy name as a name before the results as it is with the Apps.

4th January 2013

Happy new Year everyone. I fixed a problem with the "5 Daily Puzzle" app and "Sun Puzzles" app - both read a fetching script to pull in today's Killer puzzles. Unfortunately, when I loaded the puzzles for 2013 I didnt tell the script the so it was rejecting date requests in 2013. Doh. Sorry if you couldn't play Killers but glad it is easy to fix.

I'm looking for testers for the Sudoku Solver for Android. If you have a smartphone and you want to try the solver I'd like to hear from you. If you have a tablet - give me a day or so, I'm working on the larger tablet layouts at the moment.

13th December 2012

I've started to build the Sudoku Solver for the Android platform, for phones and tablets. I'm very excited about this project and i'm hoping its going to be done relatively quickly. I'm starting with all the problem areas and unknowns so I'm clear on what I need to do. I'm also approaching the App with varients in mind from the start. That is, Jigsaw, Sudoku X and Killer ought to be build now rather than bolted on later. Check out the development page for more.

21st November 2012

Update to the Jigsaw Sudoku Solver. I've re-written the identification of Pointing Pairs/Triples as I was missing a big trick that used the odd shapes of boxes. This greatly increases the number of Pointing eliminations. Many thanks to Brian Hobson for the example.

17th October 2012

Update for the Killer Solver today. Chris Queen's questions prompted me take another look at my innies/outies techniques. I've now split that strategy into two. What we had before is now the single cell innie or outie - usually much easier to spot and compute. But the full multiple cell pseudo-cage generalization had such a large impact I've put that nearer the middle of the pack. Full explanations are now written up on the Innies and Outies page. There is now quite a bit of overlap with Cage Splitting and Cage/Unit comparison. When one technique gets a boost it tends to take away from another but this is also helping with bottlenecks on very hard Killers.

8th October 2012

Take a look at the awsome 18 elimination Jelly-Fish on the bottom of this page!

4th October 2012

STR8S.COM is back as the primary domain for that site. Let me know if you can't find a page or site you could before, always helpful to hear other peoples experience. Users of the '5 Daily Puzzles' should be able to get their puzzles.

3rd October 2012

Anyone wishing for a compilation of news, trivia and links to new strategies and discussions can sign up to the new Newsletter. I would also like to include updates to any sites, players, competitions and apps and generally keep people informed through email.

I've started to build the Sudoku Solver for the Android platform, for phones and tablets. I'm very excited about this project and i'm hoping its going to be done relatively quickly. I'm starting with all the problem areas and unknowns so I'm clear on what I need to do. I'm also approaching the App with varients in mind from the start. That is, Jigsaw, Sudoku X and Killer ought to be build now rather than bolted on later. Check out the development page for more.

21st November 2012

Update to the Jigsaw Sudoku Solver. I've re-written the identification of Pointing Pairs/Triples as I was missing a big trick that used the odd shapes of boxes. This greatly increases the number of Pointing eliminations. Many thanks to Brian Hobson for the example.

17th October 2012

Update for the Killer Solver today. Chris Queen's questions prompted me take another look at my innies/outies techniques. I've now split that strategy into two. What we had before is now the single cell innie or outie - usually much easier to spot and compute. But the full multiple cell pseudo-cage generalization had such a large impact I've put that nearer the middle of the pack. Full explanations are now written up on the Innies and Outies page. There is now quite a bit of overlap with Cage Splitting and Cage/Unit comparison. When one technique gets a boost it tends to take away from another but this is also helping with bottlenecks on very hard Killers.

8th October 2012

Take a look at the awsome 18 elimination Jelly-Fish on the bottom of this page!

4th October 2012

STR8S.COM is back as the primary domain for that site. Let me know if you can't find a page or site you could before, always helpful to hear other peoples experience. Users of the '5 Daily Puzzles' should be able to get their puzzles.

3rd October 2012

Anyone wishing for a compilation of news, trivia and links to new strategies and discussions can sign up to the new Newsletter. I would also like to include updates to any sites, players, competitions and apps and generally keep people informed through email.

24th August 2012

Version 1.89 is a minor update to fix a bug occuring in Internet Explorer. Please refresh the solver pages.

24th August 2012

I have added the Extended Unique Rectangles strategy to the Sudoku solver. (v 1.88)

23th August 2012

Killer Sudoku update - added some missing Pointing Pairs strategies within cages. Thank you to Xavier Assfeld in France for showing me those gaps.

Updated the Killer Sudoku solver code to have the changes present in Sudoku and others. This means the Pairs/Triple and so on now highlight in green. Also added automatic saving of the current board position on any change - so you can navigate away. Still allows the normal Sabe/Reload.

29th July 2012

Very interesting paper which I'm reading: The Chaos Within Sudoku - A Richter Scale on a paper by Maria Ercsey-Ravasz and Zoltan Toroczkai at the Faculty of Physics, Babes-Bolyai University, Romania. I'm hoping to apply this method to a scoring for this site but there's a lot of interesting math to digest. Various outlets are summarising the results, for example i-programmer.info

29th July 2012

Printing issues? Refresh the printable popup page - new scripts to allow small (newspaper) sized printing as well as the large print. Don't forget - to remove candidates untick "show candidates" on the solver. Changes in Sudoku, Sudoku X and Colour Sudoku.

I now have a method of flagging those Weekly unsolvables which are now solvable due to improvements in the solver. I've also extended the archive from 30 weeks to 40.

Version 1.89 is a minor update to fix a bug occuring in Internet Explorer. Please refresh the solver pages.

24th August 2012

I have added the Extended Unique Rectangles strategy to the Sudoku solver. (v 1.88)

23th August 2012

Killer Sudoku update - added some missing Pointing Pairs strategies within cages. Thank you to Xavier Assfeld in France for showing me those gaps.

Updated the Killer Sudoku solver code to have the changes present in Sudoku and others. This means the Pairs/Triple and so on now highlight in green. Also added automatic saving of the current board position on any change - so you can navigate away. Still allows the normal Sabe/Reload.

29th July 2012

Very interesting paper which I'm reading: The Chaos Within Sudoku - A Richter Scale on a paper by Maria Ercsey-Ravasz and Zoltan Toroczkai at the Faculty of Physics, Babes-Bolyai University, Romania. I'm hoping to apply this method to a scoring for this site but there's a lot of interesting math to digest. Various outlets are summarising the results, for example i-programmer.info

29th July 2012

Printing issues? Refresh the printable popup page - new scripts to allow small (newspaper) sized printing as well as the large print. Don't forget - to remove candidates untick "show candidates" on the solver. Changes in Sudoku, Sudoku X and Colour Sudoku.

I now have a method of flagging those Weekly unsolvables which are now solvable due to improvements in the solver. I've also extended the archive from 30 weeks to 40.

16th July 2012

Updated the solver to cope with weird non-killer patterns that could cause it to crash. Probably a good idea to have that sort of trapping in anyway. I have added Nishio Forcing Chains replacing the old Nishio which never did anything anyway. This is a variety of Digit Forcing Chain where a contradiction is sought with two chains starting with a candidate turned ON. I have also rewritten part of the Grouped Cells to include many more instances. Fingers crossed its not going to get significantly slower. I've also updated the Jigsaw Solver to bring the basic strategy reporting into line with the solver solvers.

30th June 2012

New Sudoku article: A New Metric for Difficult Sudoku Puzzles?.

And I've buckled under the deluge - so an article on Arto Inkala's puzzle (now in the solver list).

And since it got burried, here is the link to Escargot.

I've updated the Jigsaw Sudoku to show box numbers in a diagram. The strategy results were sometimes unclear as to the boxes referred because of the odd shapes. Thanks to Peter Chris for this idea.

Finally, I've put together a comprehensive list of all the books Jeff and I have compiled or written over the years, including the new ones out later this year! (actually there's half a dozen Telegraph books I contributed to in the period 2006-2007 which I can't find, but as I never got a cent in royalties for any of them, it's a case of ho hum).

22nd June 2012

Tweaked the Grouped X-Cycles detection to prefer loops with fewer Grouped Cells if the lengths and other preferred factors are the same. Updated and added to the documentation.

8th June 2012

New in version 1.85. I had a lot of emails from people using the solver to enter a puzzle to print it off. The recent changes meant that 1 to 9 candidates were also being shown in the print page. The answer was to untick 'show candidates' but few people realised that so I have now reverted to the previous behaviour where "Clear" produced blank cells - but only if 'Auto Clear' is unticked (which it is by default). I have added an option to allow the user to determine if number entry is for clues or solutions. This only effects the color of the numbers but its important for the new saving mechanism I have created.

Updated the solver to cope with weird non-killer patterns that could cause it to crash. Probably a good idea to have that sort of trapping in anyway. I have added Nishio Forcing Chains replacing the old Nishio which never did anything anyway. This is a variety of Digit Forcing Chain where a contradiction is sought with two chains starting with a candidate turned ON. I have also rewritten part of the Grouped Cells to include many more instances. Fingers crossed its not going to get significantly slower. I've also updated the Jigsaw Solver to bring the basic strategy reporting into line with the solver solvers.

30th June 2012

New Sudoku article: A New Metric for Difficult Sudoku Puzzles?.

And I've buckled under the deluge - so an article on Arto Inkala's puzzle (now in the solver list).

And since it got burried, here is the link to Escargot.

I've updated the Jigsaw Sudoku to show box numbers in a diagram. The strategy results were sometimes unclear as to the boxes referred because of the odd shapes. Thanks to Peter Chris for this idea.

Finally, I've put together a comprehensive list of all the books Jeff and I have compiled or written over the years, including the new ones out later this year! (actually there's half a dozen Telegraph books I contributed to in the period 2006-2007 which I can't find, but as I never got a cent in royalties for any of them, it's a case of ho hum).

22nd June 2012

Tweaked the Grouped X-Cycles detection to prefer loops with fewer Grouped Cells if the lengths and other preferred factors are the same. Updated and added to the documentation.

8th June 2012

New in version 1.85. I had a lot of emails from people using the solver to enter a puzzle to print it off. The recent changes meant that 1 to 9 candidates were also being shown in the print page. The answer was to untick 'show candidates' but few people realised that so I have now reverted to the previous behaviour where "Clear" produced blank cells - but only if 'Auto Clear' is unticked (which it is by default). I have added an option to allow the user to determine if number entry is for clues or solutions. This only effects the color of the numbers but its important for the new saving mechanism I have created.

28th May 2012

~~I have created a beta version of the Sudoku solver to allow anyone interested to test updates before they are released. This will hopefully ensure smooth roll outs. The changes are stated on the beta version.~~ Beta is now closed. (Didn't work, didn't get any feedback lol).

12th May 2012

On the small board for number entry I have added an option that automatically clears off candidates as numbers are added. Also, changes to the small board are automatically saved. Really appeciate the alerts and feedbacks.

11th May 2012

I have redesigned the way cookies are stored and puzzles loaded on the solvers for Sudoku, Sudoku X, Jigsaw and Colour. You still have a manual save and reload but the solver now automatically saves the board every time it changes. Should you loose the page it will restore the puzzle you were working on. Cookies also retain the difference between clues and solved cells as well.

12th March 2012

New stats from a large stock have allowed me to update The Relative Incidence of Sudoku Strategies article.

5th March 2012, 5pm

Sorry about today's downtime. The entire ISP was down due to a spam attack which broke the router computers at the server farm. Appears to be all solved now.

4th March 2012

Testing a new back end tonight. Last weeks one had to be reverted because of occasional crashes. This version improves handling of some strategies in Sudoku X solver.

27th February 2012 - Documentation Update and todays Interruptions

The Sudoku Solver has been up and down today. I think I have got the problem, but as usual, some trivial sounding updates didn't go in as smoothly as I planned. Not that I plan, I just go ahead. This update will improve the reporting on certain advanced strategies. The document update is almost complete. I have replaced many old examples which have been superceded by a better solver and I am changing all the diagrams to PNG to get over the fuzziness of jpg.

25th January 2012 - Documentation Spring Clean

In preparation for the release of the iPhone Sudoku Solver I am going through all the documents and tidying up the examples and text. All the site documentation (that's relevant) will be available off-line on the app. Mostly this is style sheet stuff but I am hoping to update the oldest pages and a few missing strategies. If anyone has any suggestions (anything from whole articles to typos) please let me know.

I have added a new Getting Started article as the site has been a bit weak for beginners.

12th May 2012

On the small board for number entry I have added an option that automatically clears off candidates as numbers are added. Also, changes to the small board are automatically saved. Really appeciate the alerts and feedbacks.

11th May 2012

I have redesigned the way cookies are stored and puzzles loaded on the solvers for Sudoku, Sudoku X, Jigsaw and Colour. You still have a manual save and reload but the solver now automatically saves the board every time it changes. Should you loose the page it will restore the puzzle you were working on. Cookies also retain the difference between clues and solved cells as well.

5th May 2012 Now in the Apple store! SudokuWiki Solver iPhone App Click for views and features |

12th March 2012

New stats from a large stock have allowed me to update The Relative Incidence of Sudoku Strategies article.

5th March 2012, 5pm

Sorry about today's downtime. The entire ISP was down due to a spam attack which broke the router computers at the server farm. Appears to be all solved now.

4th March 2012

Testing a new back end tonight. Last weeks one had to be reverted because of occasional crashes. This version improves handling of some strategies in Sudoku X solver.

27th February 2012 - Documentation Update and todays Interruptions

The Sudoku Solver has been up and down today. I think I have got the problem, but as usual, some trivial sounding updates didn't go in as smoothly as I planned. Not that I plan, I just go ahead. This update will improve the reporting on certain advanced strategies. The document update is almost complete. I have replaced many old examples which have been superceded by a better solver and I am changing all the diagrams to PNG to get over the fuzziness of jpg.

25th January 2012 - Documentation Spring Clean

In preparation for the release of the iPhone Sudoku Solver I am going through all the documents and tidying up the examples and text. All the site documentation (that's relevant) will be available off-line on the app. Mostly this is style sheet stuff but I am hoping to update the oldest pages and a few missing strategies. If anyone has any suggestions (anything from whole articles to typos) please let me know.

I have added a new Getting Started article as the site has been a bit weak for beginners.

9th January 2012 - There is no 16 Clue Sudoku

I have written up a 'news article' about the new paper offering a brute force proof that There is no 16 Clue Sudoku

4th January 2012 - Solver crashes

Back in the UK now and trying to figure out what in the last update is causing the solver to crash.~~The usual culprit is a stray key request or print out used in the offline test program. The server side program waits for a key that will never be pressed~~. Many thanks for all the alerts. 7th January 2012 - Figured it out from the log files. The new strategy addon was not appropriate for certain types of Jigsaw puzzle with too many 'boxes' in a row or column. Should be stable now.

1st January 2012 - Grouped Cell AIC Update

I've had some very interesting posts recently and just one small aside in one of these - thanks David PB - has made me take another look at the Grouped Cells used in X-Cycles, AICs and Forcing Chains. What I realised was that in all cases an OFF candidate was turning ON all the candidates in the group - the group then pointed along the row and column to the next link in the chain. I hadn't tried the opposite where an ON candidate in a row or column turned OFF all the candidates in the group and then to continue the link this might turn ON a candidate in the same box as the group. I'll expand the docs later to make this clearer. Testing with Ruud's 50k set I am now up to 99.2% and 14,000 of these types of links were used. So a big boost. One side effect is to solve half of the unsolveables on the Weekly Unsolvable. It will be more difficult to produce these puzzles from now on, but that's progress.

25th December 2011 - Happy Christmas!

Thanks everyone for such a great year. It's been fascinating corresponding with so many people and I have a job queue as long as my arm - so many new ideas across a range of puzzles.

I've put a screen shot of the solver app on the App Development page. Yesterday I released a Puzzle Pack of the first fifty Extreme Str8ts puzzles. A very nice collection indeed.

I have written up a 'news article' about the new paper offering a brute force proof that There is no 16 Clue Sudoku

4th January 2012 - Solver crashes

Back in the UK now and trying to figure out what in the last update is causing the solver to crash.

1st January 2012 - Grouped Cell AIC Update

I've had some very interesting posts recently and just one small aside in one of these - thanks David PB - has made me take another look at the Grouped Cells used in X-Cycles, AICs and Forcing Chains. What I realised was that in all cases an OFF candidate was turning ON all the candidates in the group - the group then pointed along the row and column to the next link in the chain. I hadn't tried the opposite where an ON candidate in a row or column turned OFF all the candidates in the group and then to continue the link this might turn ON a candidate in the same box as the group. I'll expand the docs later to make this clearer. Testing with Ruud's 50k set I am now up to 99.2% and 14,000 of these types of links were used. So a big boost. One side effect is to solve half of the unsolveables on the Weekly Unsolvable. It will be more difficult to produce these puzzles from now on, but that's progress.

25th December 2011 - Happy Christmas!

Thanks everyone for such a great year. It's been fascinating corresponding with so many people and I have a job queue as long as my arm - so many new ideas across a range of puzzles.

I've put a screen shot of the solver app on the App Development page. Yesterday I released a Puzzle Pack of the first fifty Extreme Str8ts puzzles. A very nice collection indeed.

4th December 2011 - Naked and Hidden Singles

I'm trying to bring some consistency between the online solver and the new one we've building for phones. It's making me think again about some stuff I did very early on and not reflected on since. So now you will see the Hidden Singles show as green (yellow I'm trying to reserve for eliminated cells consistently). I have also rolled the row and column singles into one with boxes. At the same time I have stopped 'knock-on' eliminations which occurred because of previous eliminations. I've also removed the knock on effects in Naked Pairs. Additionally I have cleaned up the list of eliminations by putting into one line whether the Hidden Single was unique in a Row and/or a Column and/or a Box.

30th October 2011 - New puzzles for the KenKen player

new puzzles have been created and loaded into the ">6x6 KenKen Player. These will come in four grades throughout the week: Diabolicals on Monday and Tuesday going down to a Gentle on Sunday. If you are struggling with the solution you can load previous puzzles into the solver.

3rd October 2011 - Solvers updated to version 1.78

I have looked again at the basic strategies, something I haven't done for a while. For Pairs, Triples, Quads, Pointing Pairs and Box/Line Reductions - these all now show in green the candidates that contributed to the pattern. This should make searching for them from the textual description redundant and more 'at a glance'. Previously only the eliminated candidates were being show (ones in yellow).

30th September 2011 - Lots of new Books!

Harper Collins in the UK have commission three difficult Sudoku books, a small strategy book to be called "The Little Book of Sudoku Secrets" (which follows their current 'secrets' range for Scrabble and Crosswords) and a Str8ts book. These will be available in shops in the autumn of 2012. Our four German Str8ts books published by Süddeutsche Zeitung are on their second reprint and they have commissioned two more Str8ts books.

We will also be co-branding our new Sudoku Solver App with Harper Collins.

8th September 2011 - Jigsaw Sudoku Player

I'm pleased to release the latest puzzle players for this site: The Daily Jigsaw Sudoku Puzzle contains one of four graded Jigsaw Sudoku puzzles - and in the iPad version you can play four grades! Let me know if there are any issues. I have also released a 6x6 KenKen Player courtesy of www.kenken.com

16th August 2011 - Jigsaw Sudoku updated

Added some new 'shape' patterns to the Jigsaw Sudoku solver - see the end of the example pull down list. These are used in the new Jigsaw Sudoku Puzzle Packs.

10th August 2011 - Chains and Links Articles

I have added the article Introducing Chains and Links and connected it to the article Weak and Strong Links.

7th August 2011 - Solver version 1.76

Last three days have seen numerous crashes with the solvers. I suspected some kind of automated scripting attack which was hitting the server side too fast. But I have also found I can crash the killer solver is cages bigger than 10 are passed. As these are illegal I never though they would be entered, but thanks to feedback I've identified this problem and added a sanity check to killers. I am hoping this is the last of the crashings for awhile. Let me know if you can consistently break the solver - but please don't go out of your way to try :)

I have also taken the opportunity to add another Killer strategy. It is not documented yet, but I will do so asap.

21st June 2011 - Updated "Take Step"

I'm pleased to release version 1.74 of the solvers. I have sped up the "Take Step" process by skipping failed strategies and going straight to the first successful one. This applies to the basic strategies which are client-side and should save a great deal of clicking. If you get "script taking too long" messages from your browser, let me know. It's a bit risky for slower machines and I'd like to find out if this works for everyone. To help with speed I have moved the Naked/Hidden Quads to the server side. Please clear your cache if you have problems.

19th May 2011 - New 'Colour' Sudoku Solver

This new Sudoku variant solver is now released. Please let me know of any bugs, the version is beta until it has been tested more thoroughly. Share and Enjoy.

3rd May 2011 - New Weekly 'Unsolvable' Sudoku Puzzle

I am pleased to announce a new weekly puzzle. This is an extreme Sudoku slightly off the charts - in that I cannot solve it logically myself using the armory of strategies I have, or if I can I need to use trial and error at some point. I invite you to have a go and post any ideas in the discussion tab. Best of luck!

1st May 2011 - Bowman's Bingo Documentation

I have created an article to explain Bowman's Bingo. Not a strategy of choice since I regard it as falling to the sins of trial and error, but worth explaining.

I have added the article Introducing Chains and Links and connected it to the article Weak and Strong Links.

7th August 2011 - Solver version 1.76

Last three days have seen numerous crashes with the solvers. I suspected some kind of automated scripting attack which was hitting the server side too fast. But I have also found I can crash the killer solver is cages bigger than 10 are passed. As these are illegal I never though they would be entered, but thanks to feedback I've identified this problem and added a sanity check to killers. I am hoping this is the last of the crashings for awhile. Let me know if you can consistently break the solver - but please don't go out of your way to try :)

I have also taken the opportunity to add another Killer strategy. It is not documented yet, but I will do so asap.

21st June 2011 - Updated "Take Step"

I'm pleased to release version 1.74 of the solvers. I have sped up the "Take Step" process by skipping failed strategies and going straight to the first successful one. This applies to the basic strategies which are client-side and should save a great deal of clicking. If you get "script taking too long" messages from your browser, let me know. It's a bit risky for slower machines and I'd like to find out if this works for everyone. To help with speed I have moved the Naked/Hidden Quads to the server side. Please clear your cache if you have problems.

19th May 2011 - New 'Colour' Sudoku Solver

This new Sudoku variant solver is now released. Please let me know of any bugs, the version is beta until it has been tested more thoroughly. Share and Enjoy.

3rd May 2011 - New Weekly 'Unsolvable' Sudoku Puzzle

I am pleased to announce a new weekly puzzle. This is an extreme Sudoku slightly off the charts - in that I cannot solve it logically myself using the armory of strategies I have, or if I can I need to use trial and error at some point. I invite you to have a go and post any ideas in the discussion tab. Best of luck!

1st May 2011 - Bowman's Bingo Documentation

I have created an article to explain Bowman's Bingo. Not a strategy of choice since I regard it as falling to the sins of trial and error, but worth explaining.

30th April 2011 - Comment Boxes

I've added Facebook comment box to the pages with content and re-vamped the feedback page. I hope the colour scheme is okay wih most people. I have been taking advice from this page :)

19th April 2011 - Added Social Networking links

Finally figured out how to do this properly thanks to the excellent service provided by Add This. I've also added Open Graph protocol tags which will make facebook and search engine optimisation even more optimised. I am keeping social network links confined to just the pages with juicy content, which are mostly the strategy pages.

I have also ported the html and css code of the site and the solvers to XHTML 4.01 Transitional. This was a pain but it cleaned up a lot of tags and makes it much stricter. But Facebook tags still dont validate, hey ho.

Edit: I've just discovered that "Like" links on Facebook return parameters to my site. This screwed up the page identification and gave everyone a 404 error. Fixed now but it explains why I had so many 404 messages. Doh.

23rd March 2011 - Added a strategy to Killer Sudoku Solver

I have added a new strategy, Cage/Unit Overlap, - a fairly obvious one - thanks to an example from Marcel Cox. The existing strategies didn't take into account this kind of overlap between cages and units, and its sufficiently common and useful to warrent its own entry. Documentation to follow.

I've added Facebook comment box to the pages with content and re-vamped the feedback page. I hope the colour scheme is okay wih most people. I have been taking advice from this page :)

19th April 2011 - Added Social Networking links

Finally figured out how to do this properly thanks to the excellent service provided by Add This. I've also added Open Graph protocol tags which will make facebook and search engine optimisation even more optimised. I am keeping social network links confined to just the pages with juicy content, which are mostly the strategy pages.

I have also ported the html and css code of the site and the solvers to XHTML 4.01 Transitional. This was a pain but it cleaned up a lot of tags and makes it much stricter. But Facebook tags still dont validate, hey ho.

Edit: I've just discovered that "Like" links on Facebook return parameters to my site. This screwed up the page identification and gave everyone a 404 error. Fixed now but it explains why I had so many 404 messages. Doh.

23rd March 2011 - Added a strategy to Killer Sudoku Solver

I have added a new strategy, Cage/Unit Overlap, - a fairly obvious one - thanks to an example from Marcel Cox. The existing strategies didn't take into account this kind of overlap between cages and units, and its sufficiently common and useful to warrent its own entry. Documentation to follow.

13th March 2011 - New Rotate button on Sudoku Solver

In response to a query about how a Sudoku puzzle solved differently when rotated, I have added a couple of rotate buttons on the Sudoku and Sudoku X solvers. Infact the original Sudoku was broken and nothing could be said about it, but in general, rotating a puzzle can lead to new insights, especially when solving by pen and paper. My solver will in almost all cases, solve the puzzle in the same way whatever rotation or symetrical transformation is applied. There are exceptions due to the way patterns are searched for - often the search begins in the top left corner and ends in the bottom right - and the first instance of a pattern is returned. I imagine that for some advanced strategies, a rotation may return a different elimination - it depends how many opportunities there are.**I also re-wrote some internal code, so there may be new bugs. Hopefully now, but all reports greatly appreciated**

24th January 2011 - New Email button on Killer Solver

I've added an Email button to ease transfering a Killer definition and loading it into the solver. If you want to link to the solver with a puzzle, the definition is this: www.sudokuwiki.org/killersudoku.htm?bd= followed by the colour map. This is an 81 string number containing the numbers 1 to 4 which are the four colours. This is the most compact way to express the cages. Then a comma. Then the clues. The clue number in the top-left-most cell is required. All clues are two digits padded with a zero if necessary. Non-clue cells are '00'. The string will be 162 numbers long. See the Email button for an example when any board is loaded.

20th January 2011 - New navigation on data entry board

I've replaced the navigation (arrow keys, tabbing and so on) on the small board for Sudoku, Sudoku X and Jigsaw. It should be a lot easier and faster to enter numbers now. Its quite a tricky bit of code with all the validation and so on, so please report any weird behaviour.

24th November 2010 - Apperance on Dragon's Den

My colleague Jeff Widderich is appearing Dragon's Den! To be shown on Canada on CBC, next Wednesday

Click on this link for pictures and further announcements

In response to a query about how a Sudoku puzzle solved differently when rotated, I have added a couple of rotate buttons on the Sudoku and Sudoku X solvers. Infact the original Sudoku was broken and nothing could be said about it, but in general, rotating a puzzle can lead to new insights, especially when solving by pen and paper. My solver will in almost all cases, solve the puzzle in the same way whatever rotation or symetrical transformation is applied. There are exceptions due to the way patterns are searched for - often the search begins in the top left corner and ends in the bottom right - and the first instance of a pattern is returned. I imagine that for some advanced strategies, a rotation may return a different elimination - it depends how many opportunities there are.

24th January 2011 - New Email button on Killer Solver

I've added an Email button to ease transfering a Killer definition and loading it into the solver. If you want to link to the solver with a puzzle, the definition is this: www.sudokuwiki.org/killersudoku.htm?bd= followed by the colour map. This is an 81 string number containing the numbers 1 to 4 which are the four colours. This is the most compact way to express the cages. Then a comma. Then the clues. The clue number in the top-left-most cell is required. All clues are two digits padded with a zero if necessary. Non-clue cells are '00'. The string will be 162 numbers long. See the Email button for an example when any board is loaded.

20th January 2011 - New navigation on data entry board

I've replaced the navigation (arrow keys, tabbing and so on) on the small board for Sudoku, Sudoku X and Jigsaw. It should be a lot easier and faster to enter numbers now. Its quite a tricky bit of code with all the validation and so on, so please report any weird behaviour.

24th November 2010 - Apperance on Dragon's Den

My colleague Jeff Widderich is appearing Dragon's Den! To be shown on Canada on CBC, next Wednesday

Click on this link for pictures and further announcements

25th September 2010 - New players for iPhone and iPad

I'm pleased to add to this site some Sudoku and Str8ts players for the iPhone and iPad.

These will work on your PC or Mac if you use the Safari browser, but they are designed for those Apple devices specifically.

You can also switch between English and German. The links are:

And in German:

8th August 2010 - Expanded X-Wing docs

1st May 2010 - Killer Solver fixed

There was a bug that prevented modified or newly entered Killer puzzles being sent to the solver when using Internet Explorer. Seems to work for me now, let me know if there are still problems. Clear your cache before reporting. Also fixed a bug that sometimes presented chains in the killer solver when it shouldn't have.

5th Apr 2010 - Sue-De-Coq documented

A bit overdue but the documentation for Sue-De-Coq has been completed.

13th Mar 2010 - Digit Forcing Chains

I have added an extension to Forcing Chains I am dubbing Digit Forcing Chains - since they act on a single digit. It considers the implications through chains of having a candidate turned ON and turned OFF. This is the simplest type of Forcing Chain.

I'm pleased to add to this site some Sudoku and Str8ts players for the iPhone and iPad.

These will work on your PC or Mac if you use the Safari browser, but they are designed for those Apple devices specifically.

You can also switch between English and German. The links are:

- http://www.str8ts.com/ipad/sudoku.html?lang=en - iPhone Sudoku
- http://www.str8ts.com/ipad/str8ts.html?lang=en - iPhone Str8ts
- http://www.str8ts.com/ipad/sudoku-full.html?lang=en - iPad Sudoku
- http://www.str8ts.com/ipad/str8ts-full.html?lang=en - iPad Str8ts

And in German:

- http://www.str8ts.com/ipad/sudoku.html?lang=de - iPhone Sudoku
- http://www.str8ts.com/ipad/str8ts.html?lang=de - iPhone Str8ts
- http://www.str8ts.com/ipad/sudoku-full.html?lang=de - iPad Sudoku
- http://www.str8ts.com/ipad/str8ts-full.html?lang=de - iPad Str8ts

8th August 2010 - Expanded X-Wing docs

1st May 2010 - Killer Solver fixed

There was a bug that prevented modified or newly entered Killer puzzles being sent to the solver when using Internet Explorer. Seems to work for me now, let me know if there are still problems. Clear your cache before reporting. Also fixed a bug that sometimes presented chains in the killer solver when it shouldn't have.

5th Apr 2010 - Sue-De-Coq documented

A bit overdue but the documentation for Sue-De-Coq has been completed.

13th Mar 2010 - Digit Forcing Chains

I have added an extension to Forcing Chains I am dubbing Digit Forcing Chains - since they act on a single digit. It considers the implications through chains of having a candidate turned ON and turned OFF. This is the simplest type of Forcing Chain.

10th Mar 2010 - Editing bugs fixed

Many thanks for the feedbacks, especially bug alerts. I was aware of candidate editing in Internet Explorer was buggy - it wouldn't redraw a cell when you exited the edit box. But the other solvers were worse. I've gone through all four solvers in IE and Firefox and fixed the editing so it is now clear and displays properly. As ever, please clear your cache before reporting a problem - 90% of the time problems are caused by mixing old and new scripts.

Many thanks for the feedbacks, especially bug alerts. I was aware of candidate editing in Internet Explorer was buggy - it wouldn't redraw a cell when you exited the edit box. But the other solvers were worse. I've gone through all four solvers in IE and Firefox and fixed the editing so it is now clear and displays properly. As ever, please clear your cache before reporting a problem - 90% of the time problems are caused by mixing old and new scripts.

7th Mar 2010 - Strategy docs updated

See

3D Medusa

Cell Forcing Chains

Unit Forcing Chains

Some strategies have been dropped from the solver and marked as depreciated in the documentation. These are Guardians, Multivalue X-Wing and BUG.

See

3D Medusa

Cell Forcing Chains

Unit Forcing Chains

Some strategies have been dropped from the solver and marked as depreciated in the documentation. These are Guardians, Multivalue X-Wing and BUG.

6th Mar 2010 - New Release of all Solvers!

I am pleased to be releasing a new version of the solvers (all four). This is the first of two installments. Today's release contains several new strategies: 3d Medusa and a much more powerful set of Forcing Chains. 3D Medusa on a single number is identical to Simple Colouring. On more than one number it replaces Multi-Coloring - but the new implementation is more powerful.

My orginal Forcing Chains algorithm was very weak. I have completely re-written it to look at Dual, Triple and Quad Forcing chains both in a single cell (2, 3 or 4 candidates in a cell) and across a unit (2, 3 or 4 of X in a unit). Both these strategies are in the process of documentation. The success rate of these strategies is very high so I have also had to recalibrate my grader.

I have replaced all the example puzzles in the Sudoku solver and expanded the list. This should help anyone interested in finding an example.

I am moving away from the formal terms "strong (=)" and "weak (-)" links in favour of + and - symbols indicating a candidate is presumed to be ON or OFF - which I am now colouring on the board in green and red. This applies to all chaining strategies. This will I hope make the contraditions and eliminations much clearer and explanatory. I have a great deal of documentation to update because of these changes.

The second part of the release will be a re-design of the way the solver returns solutions. What I want to provide is a list of the best (easiest) eliminations at any one stage. This will apply to diabolical strategies and above. This will be the first time we will be able to look breath-wise across the solution space - rather than depth-wise as at the moment. Currently the solver returns the first elimination of a sucessful strategy. The reason for this was speed. The new version, I hope, will return a list of results from which preferences can be chosen and sorted. This will allow anyone to move through the solution using techniques they are most confortable with. It will also allow us to better understand a difficult problem when different solution can be looked at at each stage.

Enjoy

I am pleased to be releasing a new version of the solvers (all four). This is the first of two installments. Today's release contains several new strategies: 3d Medusa and a much more powerful set of Forcing Chains. 3D Medusa on a single number is identical to Simple Colouring. On more than one number it replaces Multi-Coloring - but the new implementation is more powerful.

My orginal Forcing Chains algorithm was very weak. I have completely re-written it to look at Dual, Triple and Quad Forcing chains both in a single cell (2, 3 or 4 candidates in a cell) and across a unit (2, 3 or 4 of X in a unit). Both these strategies are in the process of documentation. The success rate of these strategies is very high so I have also had to recalibrate my grader.

I have replaced all the example puzzles in the Sudoku solver and expanded the list. This should help anyone interested in finding an example.

I am moving away from the formal terms "strong (=)" and "weak (-)" links in favour of + and - symbols indicating a candidate is presumed to be ON or OFF - which I am now colouring on the board in green and red. This applies to all chaining strategies. This will I hope make the contraditions and eliminations much clearer and explanatory. I have a great deal of documentation to update because of these changes.

The second part of the release will be a re-design of the way the solver returns solutions. What I want to provide is a list of the best (easiest) eliminations at any one stage. This will apply to diabolical strategies and above. This will be the first time we will be able to look breath-wise across the solution space - rather than depth-wise as at the moment. Currently the solver returns the first elimination of a sucessful strategy. The reason for this was speed. The new version, I hope, will return a list of results from which preferences can be chosen and sorted. This will allow anyone to move through the solution using techniques they are most confortable with. It will also allow us to better understand a difficult problem when different solution can be looked at at each stage.

Enjoy

25th Jan 2010 - Vastly superior Chaining illustrations on solver

I'm very pleased with a new graphical element I've added to Chaining strategies. Where chains are used the Sudoku solver will now draw those chains on the board. Previously I could only highlight the cells which were part of the chain but this was difficult to follow. Now Strong and Weak links are plotted between candidates. The new illustrations can also plot grouped ALS cells, as in the example to the right.

This example, from sudoku.com.au, has numerous instances of chaining strategies and tests the new illustrations very well - try loading this puzzle and stepping through.

I am using a very simple but effective javascript library by Walter Zorn (www.walterzorn.de/en).

I'm very pleased with a new graphical element I've added to Chaining strategies. Where chains are used the Sudoku solver will now draw those chains on the board. Previously I could only highlight the cells which were part of the chain but this was difficult to follow. Now Strong and Weak links are plotted between candidates. The new illustrations can also plot grouped ALS cells, as in the example to the right.

This example, from sudoku.com.au, has numerous instances of chaining strategies and tests the new illustrations very well - try loading this puzzle and stepping through.

I am using a very simple but effective javascript library by Walter Zorn (www.walterzorn.de/en).

24th Jan 2010 - New strategy added

Hidden Unique Rectangles have been expanded to include the new Type 2B.

Hidden Unique Rectangles have been expanded to include the new Type 2B.

26th Sep 2009 - New Daily Puzzle available

I've added a new Daily Str8ts puzzle at www.str8ts.com - an asymmetrical version of Str8ts. Really just to explore the fun of a board with less symmetry. Enjoy.

I've added a new Daily Str8ts puzzle at www.str8ts.com - an asymmetrical version of Str8ts. Really just to explore the fun of a board with less symmetry. Enjoy.

4th Sep 2009 - Str8ts Lite available on the iPhone

I'm very pleased to announce the release of the FREE version of Str8ts, a 20 game version of the fun 6x6 Str8ts game. More information here. This compliments the 9x9 Str8ts game already available on iTunes. Enjoy :)

I'm very pleased to announce the release of the FREE version of Str8ts, a 20 game version of the fun 6x6 Str8ts game. More information here. This compliments the 9x9 Str8ts game already available on iTunes. Enjoy :)

27th Aug 2009 - Load Daily Sudoku into Solver

Had a number of requests for this, so I have added a button to allow you to load the Daily Sudoku into the solver, either if you are stuck or if you wish to print the puzzle with the candidates shown.

Had a number of requests for this, so I have added a button to allow you to load the Daily Sudoku into the solver, either if you are stuck or if you wish to print the puzzle with the candidates shown.

17th Aug 2009 - Perfect 3-3-3 SwordFish

Klaus Brenner has found a lovely 3-3-3 Swordfish. You can view and load it from the end of the Sword-Fish page.

Klaus Brenner has found a lovely 3-3-3 Swordfish. You can view and load it from the end of the Sword-Fish page.

17th Aug 2009 - Brain Stretcher!

David Filmer has passed onto me a crackingly difficult Sudoku puzzle he has made himself. You can

load this puzzle into the solver.

David Filmer has passed onto me a crackingly difficult Sudoku puzzle he has made himself. You can

load this puzzle into the solver.

14th Aug 2009 - KenKen and KenDoku solvers released!

These are beta (for testing and evaluation) - I can't claim they are entirely bug free or are perfectly complete for all cage combinations, but I've tested quite a few. Do enjoy and let me know if you can see any improvements. Links in main menu. I have been commissioned by the folks at www.kenken.com to create these solvers and I hope to see a version of them on their web site soon.

**Keeping a track of changes on this page starting from mid August 2009**

These are beta (for testing and evaluation) - I can't claim they are entirely bug free or are perfectly complete for all cage combinations, but I've tested quite a few. Do enjoy and let me know if you can see any improvements. Links in main menu. I have been commissioned by the folks at www.kenken.com to create these solvers and I hope to see a version of them on their web site soon.

24th March 2009 -

Read my response to J.F.Crook's paper "A Pencil-and-Paper Algorithm for Solving Sudoku Puzzles.

Read my response to J.F.Crook's paper "A Pencil-and-Paper Algorithm for Solving Sudoku Puzzles.