Recently, I have asked a question and it doesn't seem to get an answer and is remaining on hold. I have edited the question and removed any content that I believed was unnessary. I believe the reason being that it is taken as one of those "P vs NP" proof questions. This is not my primary intentions. My intentions are to ask if the algorithm generates Sudoku grids of any arbitrary size following a circular shift language.

To be honest, I did indirectly imply a potential proof existed. I didn't state it because I wanted to avoid downvote and get put on hold. Not just for that reason, but for the reason that I'm still learning complexity and algorithms. I don't have many credentials, knowledges, or experiences to say so. If I was highly confident, I would have a proof, I would write a reduction of circuit-sat into my algorithm and publish it. Otherwise, I would continue learning more and more.

I am vaguely aware of the P vs NP policy, because many cry wolf and unfortunately genuine attempts at anything can be closed. An answer in the prior question says proofs should be short and have few points and that its cleanly phrased.

Overall, in what ways can I avoid the implication of "cry proof" while growth is expanding into my project? And, how do I get my question off of hold?

P vs NP, policy about such questions

My post that is on hold

  • $\begingroup$ your current post edit doesnt refer to P vs NP & think its acceptable to stay open, did vote for that but such reversals tend to be rare on SE sites in general. there is some academic research into Sudoku and yeah matrix theory can be used to construct valid Sudokus & think this all relates to computer science. iirc Sudoku has also been proven NP complete. etc not sure how your construction relates to P vs NP. suggest dropping by Computer Science Chat for extended discussion. $\endgroup$
    – vzn
    Commented Apr 27, 2019 at 3:46
  • $\begingroup$ @vzn No, the posts are closed because of clarity not because of PvsNP attempt. If you prefer discussion, please come to chat.stackexchange.com/rooms/92911/… $\endgroup$
    – Evil
    Commented Apr 27, 2019 at 13:38

2 Answers 2


Closure reasons

Your posts are closed because are unclear, not self-contained, using non-standard definitions and mixing terminology. P = NP attempt in the background is not a problem in your posts, because you did not throw your attempt with peer review request, but it is easy to connect pieces together. You were given advises, learning path plans and background required to such attempt and rules that governs this site. Please reiterate through them once more.

Consecutive posts

At very first post you have indeed started to attempt P vs NP problem, with tags and comments indicating this and to be honest I do not think it is cranky or wrong, it is just a wrong place and I all the time insist that you need some proper tools (definitions, understanding) to start a journey or realize the attempt is not going into good direction. It took whole day for me and ryan to extract underneath question and drop heavy implications from it. Overall it was 29 times edited, so I have asked you to wait and do less editing. You did, so thank you for that.

Here was missing post, but it literally asked the first question once more with question whether it would work, in the comments we have come to conclusion that it is a very long journey from your solver to proving P = NP, and exchanged advises to read about definitions (P, NP, NP-Compete, Sudoku, solver etc.).

Then you asked second question and got definition of Sudoku generalization ($k^2 \times k^2$) and info about what is NP-Hard.
Take a moment to look at this post, it has one edit, it is self-contained, you have not been asked what is going on, posts should look like this.

Your third post has wrong construction, it links previous question and once more shows your permuting/shifting operator which takes half of your post.
Answer suggests reading the definitions. It was clear at this point that you haven't read them.

Another post got closed, it took 11 edits and you still ask if something that works will work. By the way, 12 x 12 is not $k^2 \times k^2$ grid, is it?
In the comments you have given me definition of NP, and it was wrong.
Please read my comment and definitions, this time for real.
The construction of your post is wrong, you have referenced previous question, shown some 12 x 12 grid, formatted it twice, which I have asked you several times to drop and then asked whether your idea generalizes. After edit there is hand-waving argument about checking your equation at Wolfram and it works. I cannot follow your attempt, but if you are convinced that it works (and it was shown at least twice to you earlier), the post is not needed. If it is, you cannot show formatted consecutive natural numbers, format them into matrix and ask whether it generates Sudoku grid, parts of posts clearly do not stick together.

Your last post still suggests that you are confused with 3SAT and grid generator.


You have used over 5 posts to show your grid generator, but people still cannot get your idea, please stick to one definition (created by ryan), drop all irrelevant steps, and read carefully What is the definition of $P$, $NP$, $NP$-complete and $NP$-hard? then Is there a system behind the magic of algorithm analysis? and then read all https://cs.stackexchange.com/search?tab=votes&q=sudoku
This links aim at helping you to sort out definitions, get grasp of what is important to you.


You need to create real Sudoku solver, not mapper to continue your attempts, an algorithm that solves any valid Sudoku 9 x 9 grid. Measure its complexity, then generalize to $k^2 \times k^2$ and then, with your knowledge after reading given posts you may continue.

What will be closed in the future

You cannot throw another grid cyclic creator, insist it is solver, use half space of the post to show it generates some numbers and connect it to DNF, 3SAT or any other thing asking us to connect dots for you in the way you did.


Please, once again, start using standard terminology.

Sudoku puzzle is uniquely (or pairwise not uniquely) solvable grid with hints/clues (more than 9 for 9 x 9 grid). (look at https://www.websudoku.com/ what you see is Sudoku instance), whole grid filled with numbers abiding Sudoku rules is solved instance. If you want to describe solver, it will get encoded instance and return solved grid.

Grids generated by your shifting are subset of all possible puzzles, so you literally cannot infer anything from this.

It doesn't make sense to try instances 50 x 50 or 100 x 100, since you get result that it will always work. It works, yet still nothing follows from this.

One box (3 x 3) is not solved, it is filled. It is not enough to provide solution.


Since I have put your post on hold, I will first explain why I did so. I have not done this because of the topic of the question. The reason why I put it on hold is because

  1. At the time I did, your question was unclear to me.
  2. The long comment chain, people asking for clarifications there and suggesting you to read up on standard terminology, led me to believe your question was not clear to others either.
  3. The post is not self-contained. While you should link to your other post, I would at least expect a summary of the algorithm in question, as that is central to this post. Currently, I really don't know what your question is even about without referring to other material.

If these points are improved, then I think your post can be reopened. It may help to try and refine a few things in chat first, but that is up to you.

Often, I inform people how they should improve their post to get it reopened. It did not do so here, because other users were already working with you on improving/clarifying the question.

For completeness, let me explain how you would reopen a question in general. A question can get closed after having received 5 close votes from users with at least 4K rep, or singlehandedly by a moderator. (there is also some abilities for gold tag-badge users, but those are rare here)

A question can get reopened by 5 reopen votes from users with at least 4K rep, or again singlehandedly by a moderator. Whenever a closed question is edited or a reopen vote is cast on it, it gets send to the reopen queue, where users can decide to cast a reopen vote as well.

So, in principle, questions can get closed and reopened without moderator involvement at all. However, on Computer Science, there are simply too few active >4K rep users to do this effectively, meaning that the moderators are more actively involved in this process. This means that a user can be more pro-active and should ask the one who closed the question for clarification. I think this can be mostly done in the comments, though, but a meta question also works.


You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .