Strategies for Number Puzzles of all kinds
Page:2014 2013 2012 2011 :1 2 3 4 5 6 7 :2010 2009 2008 2007 2006 Search:
Latest Apps
Str8ts - it's not Sudoku
1 TO 25 - quick but tricky, play now
Feedback and Questions
I've received a lot of interesting comments and questions from Sudoku fans over the last few years and this page is where I try to answer them. Please feel free to drop me a note on the side of the page or try the Facebook comment box. Or you can email me directly at .

Post a Comment or Question here...
Your Name or 'Handle'

Part of the World you're from

Email Address (optional) so I can reply directly if necessary (it will not be displayed here)

Any Sudoku you want to publish here for easy loading into the solver - 81 characters, use '.' or '0' for unknowns.

public - comment/question will be added to feedback column
private - email comment/question directly to Andrew Stuart, don't display here

Comment, question or feedback:

Enter these letters
Please enter the
letters you see:
Many thanks to all the people who have helped improve the solvers and strategies with their feedback!

Thursday 2-Jun-2011

... by: Cruetsonly, USA

Is it possible to show only possible candidates that I enter myself? I want to be able to enter guesses, as I would do with paper and pencil, but this does not seem to allow it. For example, I want to be able to show that I have eliminated two squares down to 5 & 6 without showing all the other candidates elsewhere on the board. Other than that, I love the site.

Andrew Stuart writes (3-Jun-2011):
The solver needs to know what candidates are available at any given stage; so an empty cell is wrong since it contains no options. It would create too many ambiguous states to have the cells partially filled by the solver and partially the player. However, you can remove candidates if you are confident of your eliminations, by using the "Candidates can be Edited or Highlighted" options above the main board. The Puzzle Players, eg the Daily or Weekly puzzles allow you to enter positive candidates rather have negative removals, but I don't have a system whereby you can import your own puzzle into a player, only a solver. I'll think about what can be done.

Sunday 15-May-2011

... by: Rodney, Cyprus

Have you considered getting an apps on iTunes for this excellent program?

Andrew Stuart writes (15-May-2011):
It's definitely something I'd like to do. We have access to some good developers who I'm hoping will give me a shell I can program the solvers into and update myself.

Saturday 30-Apr-2011

... by: Nick the Geek, San Francisco Bay Area,

Load Sudoku:

Outstanding website and sudoku solver! Over the last 6 months or so, I've run across a few really interesting puzzles. Here's on of my favorites- I hope I've sent it correctly. It is particularly difficult: 6 bingos,4 quads, a pattern overlay, a huge ALS, 29 extreme chaining techniques, and finally 7 3d Medusas! (There are a few other techniques of course, but these are the most hideous) I am able to solve this puzzle, but only by guessing, which I'm learning to avoid. I think you're correct: solving by logical means instead of just picking something and see where it leads is not the true point of solving these, although I have a long ways to go. I highly reccomend the xsudo puzzle generator! Thanks for everything!

Andrew Stuart writes (30-Apr-2011):
Wow that is seriously off the chart, just ran it through my offline solver, great fun. Its down the candidate density. Most of the cells and units have 3 or more possibles for X so it greatly reduces the availability of chaining techniques and other strategies that reply on bi-value and bi-location candidates, but this *just* has enough. On the boundary, very nice. Thanks for sharing :)

BTW, I left a bug on the submission form, didn't save your email address, if you put one in.

Thursday 28-Apr-2011

... by: Donald Waddell, US

This is the best site for Sudoku that I have found by far. I am able to input the puzzles that I am totally stuck on and get a step by step solution. This enables me to see how to correctly apply the techniques you outline. Thanks.

Tuesday 19-Apr-2011

... by: Jimmy Young, Silicon Valley, California

Love your step by step solving Sudoku. I have two comments:
1. It would be very nice if you can make the most current solved entries different color, high lighted or blinking. So that we can easily find what item you just solved.
2. It would be very helpful if we can reverse the solution process, step by step. I tried but it got stuck and did not allow me to go back. Not everytime.
Thank you for your great learning tool. Jimmy

Andrew Stuart writes (21-Apr-2011):

Most strategies do highlight the eliminated candidates (in yellow), and the chaining strategies are the most clear since I use red and green indicate off/on and they also draw lines. The basic strategies could use some work since I donít show which candidates cause the eliminations, mainly because I show all eliminations of the basic strategy type, which can get a bit crowded. I'm looking to improve what's being drawn, especially as I have access to a simple graphics library.

Reversing would be nice but its would require a considerable amount of memory and variable management and I'm at the limit of what I can do with javascript. The "<<" button will always work on the last successful operation and you can toggle take step with "<<" and turn off strategies etc.

Wednesday 13-Apr-2011

... by: Jerry Foil, Virginia

Hi Andrew,
Two questions on Sudoku theory and terminology.
1. Does your terminology make a distinction between "conjugate pair", "strong link", and "bi-location pair"? I see these terms used to note a pair of cells where a given value can only be. I understand that "strong link" is used in x-cycles and grid coloring. Is there a logical distinction? Just wondering so I'll use the terms right in the future.
2. Do puzzle designers made the givens the "minimum set" for the puzzle, meaning that if you removed any given, the puzzle would have multiple solutions? I tested this on one of your very easy examples and it seems to be true.
- Jerry

Andrew Stuart writes (14-Apr-2011):
A conjugate pair and a bi-location pair is the same thing, but since the start of Sudoku people have drifted to a variety of terms including those two. I prefer bi-location since we can also talk about "bi-value" in a single cell - which is also a conjugate pair, so its a finer distinction.

A strong link is only used in an inference chain were you turn "off" one candidate forcing the other to be "on".

I can't speak for other puzzle publishers, but my generator stops removing clues when there are multiple solutions so all my puzzles are minimal. It then grades it. But clue density is no indicator of grade except when statistically graphing many thousands of puzzles.

Wednesday 23-Mar-2011

... by: Scott Baird, Nottingham England

Thanks for taking the time to put this site together, it has saved me from going out of my mind. Was doing a medium problem on my phone, and having used your tool understand why I could not even start it. Solvable, but not by me, just yet... Ever probably.

Cheers Scott

Sunday 20-Mar-2011

... by: Marcel Cox, Luxembourg

Killer Sudoku solver

Lately, I have been testing your Killer Sudoku solver on a couple of sample problems. Unfortunately, I found that unlike your regular Sudoku Solver which goes very far in strategies it uses anc solves every human solvable Sudoku, the Killer solver often fails at moderate puzzles like this one:,170000140607180000130000000000001500150010001611001000001400060014070000050008002400141000110000000000001600002600000011000000100024200000190014000000000000000000

The hint which would allow the program to solve this Killer Sudoku (at the point the solver fails) is to see that in one of the 2 cells A5 B5, there must be a 1. So the other most be a 5.

Have you ever thought of enhancing the Killer Sudoku solver to handle this kind of problems?

Andrew Stuart writes (22-Mar-2011):
That is quite an oversight. I only get a fraction of the strategy feedback for killers - even though I think they are more interesting, but even so, I am surprised my searching doesn't find them. I've done some work on this and some testing and this type will now show up. I've created a new item in the list called "cage/unit overlap" - and as its relatively simple and useful put it ahead of most other strategies. It works fine on all rows, columns and boxes

I believe I can extent it further, since a partial cage - providing the number is unique in all the cells visible to the unit and the cage together, it should also apply. I'll need more time for that extension though, but your example works. Where did it come from

If its ok with you, I've added your name to the credits (on and will also on the document page when I'm done fishing out good examples.

If you find any further exceptions, do let me know

Sunday 20-Mar-2011

... by: Art, New York

Shouldn't "Unique Rectangles" be categorized under merely "Tough" strategies, and not "Diabolical" because they can so often be spotted quite easily?

Love the site!

Andrew Stuart writes (22-Mar-2011):
Possibly. There are many variations and possible types of UR though, so overall its quite complex. My order and categorisation can only be subjective. What I'd really like to offer is a user enabled re-ordering but it's tough without making the user interface even more complex

Sunday 20-Mar-2011

... by: Werner Cejnek, Austria

Since yesterday, March 19, solution count shows "more than 500 solutions" for the daily sudoku. I think this is a bug in the counter

Andrew Stuart writes (22-Mar-2011):
Appreciate the report. There was a bug in the grid initialisation which caused this, fixed now.

Thursday 17-Mar-2011

... by: zweibieren, Pittsburgh PA USA

On the kenken solve, the solver does very little and then I cannot get it to do more. It says to remove certain choices, but it does not remove them.
And if I remove them, it puts them back.

Using Firefox 3.6.15

Having scroll bars within the page is awkward. My browser also has scroll bars and the two collaborate to make navigation difficult.

Andrew Stuart writes (27-Mar-2011):
I've upgraded the solver this weekend and also discovered a bug which stopped the server side calls working in the embedded page, I think that was the main problem. All fixed now, do have a go

Page:2014 2013 2012 2011 :1 2 3 4 5 6 7 :2010 2009 2008 2007 2006