Even harder puzzles

Programs which generate, solve, and analyze Sudoku puzzles

Even harder puzzles

Postby Guest » Sat Apr 23, 2005 10:38 pm

Wayne,

Would you consider releasing a version of your program that can produce puzzles requiring techniques harder than x-wings? I know you said you have the code to produce Nishio patterns, perhaps there are others too?

If you don't want to release the program, would you be willing to publish the occasional "Super hard" puzzle on this site?

The x-wing really opened my eyes to possibilities, and having mastered it, I find myself hankering after a tougher challenge.

Addicted? Me?:D
Guest
 

Eh?

Postby lunababy_moonchild » Tue Apr 26, 2005 8:22 am

Would somebody please tell me what X-wings and Nishio are, or where it is I'll find an explanation?

Thanks very much *hate to see things getting past me*

Luna
lunababy_moonchild
 
Posts: 659
Joined: 23 March 2005

Postby Animator » Tue Apr 26, 2005 8:24 am

There is a topic about x-wings in the Times-part of this forum... (even though it isn't required too solve the puzzles in the Times)

And you can also look at the hints given for the 'Very hard' puzzles in Solving techniques
Animator
 
Posts: 469
Joined: 08 April 2005

Postby simes » Tue Apr 26, 2005 10:14 am

simes
 
Posts: 324
Joined: 11 March 2005
Location: UK

Postby lunababy_moonchild » Tue Apr 26, 2005 4:46 pm

Thanks very much!

Luna
lunababy_moonchild
 
Posts: 659
Joined: 23 March 2005

Postby Guest » Wed Apr 27, 2005 8:19 am

Still no word from Wayne on the original question...
Guest
 

Postby Pappocom » Wed Apr 27, 2005 9:18 am

Sorry for no reply until now. Since the BBC Online Magazine published an article last Friday (April 22) I have been run ragged. (For the article, if you are interested, see http://news.bbc.co.uk/1/hi/magazine/4469719.stm)

My program (Sudoku) can produce "Very Hard" puzzles, as you know. I guess you are asking for a grade which is one level up from that.

If a puzzle is harder than Very Hard but still solveable with logic, I call it "Unfair". That's because there's no way a (non-superhuman) human can solve an Unfair puzzle without writing pencilmarks everywhere and doing a lot of tedious, time-consuming, mechanical plotting of numbers - which is no longer fun.

My program can solve Unfair puzzles (as long as they are solveable by logic, of course). That's so that my program can let you play an existing Unfair puzzle if it is dubbed in from a published 3rd-party source.

On the other hand, my program does not make up Unfair puzzles itself. It does not present original Unfair puzzles to players. It has the power to do so, but chooses not to.

I don't want to validate Unfair puzzles, so I guess the answer to both your questions is "no". I won't be releasing a version of the program which can produce Unfair puzzles. Further, I don't think Unfair puzzles should be recognized as valid, so I won't be publishing any myself.

- Wayne
Pappocom
 
Posts: 599
Joined: 05 March 2005

Postby Guest » Wed Apr 27, 2005 10:24 am

I had a feeling you might say that, but I had to ask :o)
Guest
 

Postby Sue De Coq » Wed Apr 27, 2005 12:36 pm

Like IJ, I'm very interested in the patterns that might exist beyond X-Wings. First of all, I'd like to confirm that I understand the Pappocom terminology. I believe you define puzzle descriptions as follows:

'Very Hard' = Features logical patterns up to and including the complexity of X-Wings.
'Valid but arguably unfair' = Requires use of the Nishio technique.
'Invalid' = Can't be solved by Nishio, so requires a pure guess.

I'd be grateful if you could describe any patterns you know other than X-Wings and Nishio on the forum, even if you don't incorporate then within the software. You mentioned elsewhere that you read about the Nishio on a Japanese website. Perhaps you could provide links?

You might have followed the discussions on this and the Sudoku Programmers forum about the Generalized X-Wings (or Swordfish) technique. The consensus is that this method is logically equivalent to X-Wings, yet the Pappocom software classifies X-Wing puzzles as 'Very Hard' while it classifies Swordfish puzzles as 'Arguably Unfair', which suggests that the software fails to recognize the Swordfish pattern. Instead, it appears to categorize the Swordfish pattern as a Nishio (of which it, like X-Wings, is a special case). I'd be interested to hear your opinion on Generalized X-Wings techniques.

It seems to me that, regardless of your thoughts on the Generalized X-Wings technique, the boundaries between 'Very Hard'/logically-positive, 'Arguably Unfair'/limited-proof-by-contradiction and 'Invalid'/full-proof-by-contradiction techniques will vary over time as new patterns are documented. Is it possible that you would add new techniques to the Pappocom software in the future?

TIA
Sue De Coq
 
Posts: 93
Joined: 01 April 2005

Postby Spok » Tue Aug 08, 2006 2:08 pm

Pappocom Wrote:
If a puzzle is harder than Very Hard but still solveable with logic, I call it "Unfair". That's because there's no way a (non-superhuman) human can solve an Unfair puzzle without writing pencilmarks everywhere and doing a lot of tedious, time-consuming, mechanical plotting of numbers - which is no longer fun.


I find this very annoying. I am not a superhuman by any means and yet I regularly solve much harder puzzles than the Very Hard ones produced by the Pappocom program. I always solve without any help from the computer and if I go wrong I start again. And it is actually a great deal of fun and extremely rewarding.

Puzzles that require techniques such as X-wing, Swordfish, XYZ-wing, Colours, Chains etc., are all still valid. They absolutely have just one solution and can be solved by using logic alone. You do not need to resort to guessing. Just because the people at Pappocom and their program cannot solve these puzzles is no reason to call them unfair or invalid. They are not. They are just harder and more challenging than the ones they produce. For serious Sudoko-ists, like myself, Very Hard is just not hard enough.

Spok.
Spok
 
Posts: 6
Joined: 03 January 2006

Postby Pat » Tue Dec 11, 2007 3:46 pm

Sue De Coq (2005.Apr.27) wrote:Like IJ, I'm very interested in the patterns that might exist beyond X-Wings. First of all, I'd like to confirm that I understand the Pappocom terminology. I believe you define puzzle descriptions as follows:

  • Very Hard = Features logical patterns up to and including the complexity of X-Wings
  • Valid but arguably unfair = Requires use of the Nishio technique
  • Invalid = Can't be solved by Nishio, so requires a pure guess
I'd be grateful if you could describe any patterns you know other than X-Wings and Nishio on the forum, even if you don't incorporate then within the software. You mentioned elsewhere that you read about the Nishio on a Japanese website. Perhaps you could provide links?

You might have followed the discussions on this and the Sudoku Programmers forum about the Generalized X-Wings (or Swordfish) technique. The consensus is that this method is logically equivalent to X-Wings, yet the Pappocom software classifies X-Wing puzzles as Very Hard while it classifies Swordfish puzzles as Arguably Unfair, which suggests that the software fails to recognize the Swordfish pattern. Instead, it appears to categorize the Swordfish pattern as a Nishio (of which it, like X-Wings, is a special case). I'd be interested to hear your opinion on Generalized X-Wings techniques.

It seems to me that, regardless of your thoughts on the Generalized X-Wings technique, the boundaries between Very Hard/logically-positive, Arguably Unfair/limited-proof-by-contradiction and Invalid/full-proof-by-contradiction techniques will vary over time as new patterns are documented. Is it possible that you would add new techniques to the Pappocom software in the future?

TIA


hey Sue De Coq -- you know Pappocom hasn't posted here in a while -- well, neither have you -- but anyway, in response to your question, he has now released the Sudoku 1.1 update, and it does allow Swordfish !!
User avatar
Pat
 
Posts: 3438
Joined: 18 July 2005


Return to Software