Solver App for Android and iPhone
Strategies for Number Puzzles of all kinds
Latest Apps
Page:2022 1 2 :2021 2020 2019 2018 2017 2016 2015 2014 2013 2012 2011 2010 2009 2008 2007 2006

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. I'm also directing Str8ts feedback here. Please feel free to drop me a note on the side of the page. 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!

Wednesday 26-Oct-2022

... by: Michael Hayes, Florida

I notice that in your Sudoku solver, for some tests, e.g. Hidden Singles, you will find as many as exist from searching all the rows columns and boxes, before returning to 'check for solved cells'. Yet, for other tests, e.g. NakedQuad, after finding one, you will return immediately to 'check for solved cell'. In your example for NakedQuad, after removing the 1/5/6/8 bits in Box 1, there is now another Naked Quad, 1/3/7/9 in the row B - in fact there several more, e.g. 2/5/6/8 in column 2, 1/4/5/9 in Row A. My question is, why not continue to search all the rows, columns and boxes for a particular test vs immediately returning to 'check for solved'? I'm guessing it's because when you discovered the NakedQuad in Box 1, you had already done your search of the Row and Column blocks, and you only search each block once for a particular test, rather than say, in a while loop until no more are found.

Second question: Is there any reason to search blocks in any particular order, e.g. boxes first, then rows, then columns?

Andrew Stuart writes:

Good questions. For the very easiest searches it would be too tedious to only return them one at a time. Often it's a question of balancing information displayed versus how much to click. For almost all other strategies it is necessary to return the first (or best) since it would be a muddle to display two or more. Also there can be knock on effects that might solve the puzzle if the same strategy is allowed to recurse and be exhaustive and then the user wouldn't know what had happened. Some strategies such as the chaining ones do search for all instances, store them in a list and sort them by some rules to get the "best". In a future version I'd like to allow the user to cycle through them.

Searching is 1 to 9, top left to bottom right. Mostly this does not matter as can be shown by rotating the puzzle, but sometimes can. Not a lot of choice when proceeding linearly and thoroughtly.

Sunday 23-Oct-2022

... by: Michael Becher, Germany / Munich

Load Sudoku:

Hi Andrew,

yer nice page and solver. Hard to find some sudokus where your solver does not find a solution...
I have attached one.

Best regards

Andrew Stuart writes:

Good one. Where from?
The candidate density is high. That's a sign of difficult or unsolvable.

Thursday 13-Oct-2022

... by: Brian Mottershead, United States

Thank you. I am looking forward to trying some of the "classic" ones.

Monday 10-Oct-2022

... by: Maarten, The Netherlands


First of all, I learned a lot of sudoku through your website, so thank you a lot! I was able to create my own solver and creator of puzzles thanks to your explanation of the different strategies.

In you mention that you use a brute force method to see if there is a single solution while you are creating puzzles. Why don't you use the solving strategies for it? Is it because of performance reasons?

Kind regards,

Maarten van den Hoek

Andrew Stuart writes:

Sorry for the late reply. Did you see this article? Goes into some of this. Happy to answer follow ups

Wednesday 5-Oct-2022

... by: Brian Mottershead, United States


I love the site. Thanks so much for doing it.

On the web, I often run into references or links to the David FIlmer Unsolvables, usually in the context of a discussion of what is the "hardest" sudoku". For example "David Filmer Unsolvable #28" is often mentioned as the hardest sudoku. Another contender for "hardest" is #49.

However, if there is a link (generally in the form,) it does go to the "Weekly Unsolvable" page, but only showing what is ever the current Weekly Unsolvable. Today, thae link for puz=49, shows number 516.

In the "archive" pulldown on the page. various previous numbers are listed, but not older ones. For example, right now, you can only look at numbers 477 to 516.

I am guessing that at some point, you changed this page so that the archive only shows the most recent 40, Or perhaps #28 and #49 are in the current archive, but the numbering has changed.

In any case, would there be a way to bring back the "classic" David FIlmer Unsolvables, with the old numbers, so that the old links work?

Best regards,

Brian Mottershead

Andrew Stuart writes:

Yes eventually I'd like to include them all as some sort or premium service but that's way down the line.
However I like your idea so I've added David Filmers ones where they were in the series 1 to 200

Refresh the page and check the list

Thursday 16-Jun-2022

... by: ROY WATTS, USA

I thought I could enter it prior to entering a puzzle but alas, I was wrong. Would it be possible to get it added to your site. I am very much interested in creating jigsaw puzzles and rely heavily on your solvers to insure they work. This particular shape gives new life to leftovers! Excellent site, btw.

Andrew Stuart writes:

I like your pattern and I've added it to the Jigsaw solver. Last in the list. You need to give me a name for it ;)

I've included four examples. I've not run the generator for very long so the first three are rather vanilla. Number 4 is an extreme.

If you donít see the new pattern used [CRTL]+[F5] to reload the page

It's a long standing job of mine to reorder the solver to allow user-defined maps. I really must get on with that some day

Monday 30-May-2022

... by: James McGinn, Ireland

May 22 puzzle packs are listed as May 20

Andrew Stuart writes:

Fixed! Thank you

Friday 27-May-2022

... by: Reuben, BC

Hey Andrew -
I met Jeff years ago and he introduced me to str8ts. I still play every day. I had a novel idea that I thought you might want to consider. When someone plays they could toggle on or off the option to see how their solution time compares to the times of others. I'm constantly competing against myself, but it would be fun to have the option of seeing how fast others are completing the puzzles.


Andrew Stuart writes:

Yes, I'd like to introduce more competitive components like that. Relatively difficult tech however, all that asynchronous communication. Somehow need to make a little more money from the site to feed that sort of development. Mainly that's the pinch at the moment.

Friday 20-May-2022

... by: Stratos, Greece

Dear Andrew
I am using your sudoku solver for many years and helped me a lot. SO many thanks for that and congrats for your job. However I have a question for you. I am a poor human solver. What I see is that all those Diabolical, Extreme strategies and even many of the Tough ones, can only be used by a computer. I can only use some of the simpler Tough's like X-Wing for instance, but nothing more. Do you really believe that a human being is able of using strategies like Jellyfish, Fireworks, Exocet, etc. without using a computer ??
I am just curious if my mind's capabilities are small and if there are genius people who can use by heart all those strategies you have invented.
I would really appreciate your answer.
Many thanks in advance

Andrew Stuart writes:

It is true all strategies are easy for computers - if you can code them right - which is the fun part for me and other programmers. I was interested in exploring this area and seeing what crazy stuff could prove useful. My big divide is between "pattern" based strategies and "trial and error" - whether there is a logic based on the state of play or whether some puzzles need a plug-and-test approach. Humans use that but I find trial and error unsatisfying since it doesn't tell you "why". So for most people the most complex strategies are not really for humans. That said, someone had to invent them and they came out of thinking about a problem, not asking a computer to find a strategy. Maybe an AI can go meta and think up strategies but I donít know how to set that up. So some people - at least the inventors - were mentally exploring these spaces. And I do know some people will pen-and-paper complex chaining, but I donít and most don't and they still crack them. I can't duplicate human intuition. Also, I would never torture an audience in a newspaper with an extreme and I believe my puzzles are harder than most. Diabolicals are my commercial maximum.

Friday 13-May-2022

... by: James Leonard, USA

The archived Sunday puzzles for May 01 and May 08 are blank and the date shown for them is December 30 1899 ! Cheers.

Andrew Stuart writes:

Thanks for the alert. Looks like I uploaded new stock for all the grades except the Sunday extremes. I've put them up now, have a look

Page:2022 1 2 :2021 2020 2019 2018 2017 2016 2015 2014 2013 2012 2011 2010 2009 2008 2007 2006