View previous topic :: View next topic |
Author |
Message |
cgordon
Joined: 04 May 2007 Posts: 769 Location: ontario, canada
|
Posted: Thu Aug 23, 2007 9:45 pm Post subject: Al Escargot - Hardest Sudoku Ever |
|
|
I came across a 2006 article in USA Today referring to a Finnish mathematician who claimed he had created the hardest Sudoku ever (named Al Escargot - the puzzle not the Finn). I downloaded it from this blog site:
http://photos1.blogger.com/photoInclude/blogger/3235/1499/1600/Al%20escargot.jpg
I started it but it's brutal - after a few moves I got totally stuck - though I will persevere. I hope it's a genuinely solvable and one solution problem.
Try it out. But be afraid !! |
|
Back to top |
|
|
Marty R.
Joined: 12 Feb 2006 Posts: 5770 Location: Rochester, NY, USA
|
Posted: Thu Aug 23, 2007 10:33 pm Post subject: |
|
|
Well Craig, a few moves is more than I could do. I'll be the first to embarrass myself and say that after writing in the pencil marks I couldn't eliminate one candidate. |
|
Back to top |
|
|
Asellus
Joined: 05 Jun 2007 Posts: 865 Location: Sonoma County, CA, USA
|
Posted: Fri Aug 24, 2007 4:49 am Post subject: |
|
|
There is a finned swordfish on 9 that eliminates four <9>s in B8. After that, I am, so far, stuck. |
|
Back to top |
|
|
cgordon
Joined: 04 May 2007 Posts: 769 Location: ontario, canada
|
Posted: Sat Aug 25, 2007 9:20 pm Post subject: |
|
|
I looked at all the popular techniques discussed on this site and couldn't get past a few preliminary steps (though there might be some advanced techniques I've heard mentioned e.g. Sue Coq). Anyway I did finally solve the thing by guesswork - and yes, there is a unique solution.
But I had hoped that all puzzles were solvable by existing techniques, without guesswork. Maybe they still are. |
|
Back to top |
|
|
Ruud
Joined: 18 Jan 2006 Posts: 31
|
Posted: Sat Aug 25, 2007 11:21 pm Post subject: |
|
|
There's more info on this puzzle at the Player's Forum.
It is no longer considered the hardest Sudoku ever. Here's a list of the hardest puzzles. Don't try these at home. |
|
Back to top |
|
|
Johan
Joined: 25 Jun 2007 Posts: 206 Location: Bornem Belgium
|
Posted: Sat Aug 25, 2007 11:26 pm Post subject: |
|
|
cg,
If this puzzle has a unique solution there must be some logical solving technique,(probably somewhere hidden and yet to be discovered?), for cracking this tough nut(which is an understatement).
In my opinion anyone who can solve this one with only logical solving techniques, has find the ultimate solution path for every SUDOKU, even the hardest one ever designed, like this one.
Chapeau for the one who can solve this without T&E.
Johan |
|
Back to top |
|
|
keith
Joined: 19 Sep 2005 Posts: 3355 Location: near Detroit, Michigan, USA
|
Posted: Sun Aug 26, 2007 2:59 am Post subject: |
|
|
Johan,
You may want to look at the Sudoku Programmer's Forum:
http://www.setbb.com/phpbb/?mforum=sudoku
As I recall, you can determine that a particular puzzle does / does not have a unique solution without "solving" it in human terms.
As I understand it, there are Sudokus which cannot be solved other than by trial-and-error. The nice term for this is "exhaustive search": Try every possibility.
At the risk of irritating the computer scientists who read this: Exhaustive search is how you find the next Mersenne Prime, or prove the four-colour map theorem.
Or, for readers of this forum: The only way to prove a puzzle is valid (has a unique solution) is to solve your puzzle using only methods blessed by Glassman.
Best wishes,
Keith |
|
Back to top |
|
|
Mogulmeister
Joined: 03 May 2007 Posts: 1151
|
Posted: Sun Aug 26, 2007 8:02 pm Post subject: |
|
|
Indeed - and most computer solvers make use of this algorithm although you need the brute force capabilities of a computer to do it. |
|
Back to top |
|
|
cgordon
Joined: 04 May 2007 Posts: 769 Location: ontario, canada
|
Posted: Sun Aug 26, 2007 9:12 pm Post subject: |
|
|
Quote: | ...there must be some logical solving technique,(probably somewhere hidden and yet to be discovered |
I'm with you Johan. I have faith that one day someone will come along with an exquisitely simple solution that will not only make all other solutions redundant - it will make Sudokus themselves become too simple to bother with.
I'm working on it .. and when I find it - I'll miss you guys. |
|
Back to top |
|
|
Mogulmeister
Joined: 03 May 2007 Posts: 1151
|
Posted: Mon Aug 27, 2007 8:52 am Post subject: |
|
|
If you look in the Sudoku programmer's forums you will see some conversation with Arto Inkala himself the creator of AI Escargot. |
|
Back to top |
|
|
Swampy
Joined: 27 Mar 2008 Posts: 1
|
Posted: Thu Mar 27, 2008 3:10 pm Post subject: Completed puzzle |
|
|
I found this puzzle about 15min's ago
Have finished it and can confirm a logical solution |
|
Back to top |
|
|
ravel
Joined: 21 Apr 2006 Posts: 536
|
Posted: Thu Mar 27, 2008 5:12 pm Post subject: |
|
|
The solutions i have seen so far for puzzles of this difficulty took me more than an hour to read and verify them (and it was more work than fun).
So what did you do in these 15 minutes ? |
|
Back to top |
|
|
nataraj
Joined: 03 Aug 2007 Posts: 1048 Location: near Vienna, Austria
|
Posted: Fri Mar 28, 2008 9:44 pm Post subject: Re: Completed puzzle |
|
|
Swampy wrote: | I found this puzzle about 15min's ago
Have finished it and can confirm a logical solution |
There is always a "logical solution":
"It is logical that this must be the solution, because otherwise the sudoku rules would have been violated".
Hm. Can't dispute the logic. Doesn't feel right, though.
Can you do better than that? |
|
Back to top |
|
|
waleeed12
Joined: 26 Mar 2016 Posts: 2
|
Posted: Sat Mar 26, 2016 8:49 am Post subject: |
|
|
this site and couldn't get past a few preliminary steps (though there might be some advanced techniques I've heard mentioned e.g. Sue Coq). Anyway I did finally solve the thing by guesswork - and yes, there is a unique solution.
waleeed |
|
Back to top |
|
|
dongrave
Joined: 06 Mar 2014 Posts: 568
|
Posted: Sun Mar 27, 2016 3:38 am Post subject: |
|
|
The link to the hardest Sudoku puzzle ever seems to be broken for me but luckily Gordon mentioned it by name so I was able to find it. I'm guessing that this is the puzzle that everyone is referring to here.
Code: |
+-------+-------+-------+
| 1 . . | . . 7 | . 9 . |
| . 3 . | . 2 . | . . 8 |
| . . 9 | 6 . . | 5 . . |
+-------+-------+-------+
| . . 5 | 3 . . | 9 . . |
| . 1 . | . 8 . | . . 2 |
| 6 . . | . . 4 | . . . |
+-------+-------+-------+
| 3 . . | . . . | . 1 . |
| . 4 . | . . . | . . 7 |
| . . 7 | . . . | 3 . . |
+-------+-------+-------+
|
If this is the right puzzle, here's the grid that I had after basics.
Code: |
+-----------------+---------------------+-----------------+
| 1 2568 2468 | 458 345 7 | 246 9 346 |
| 457 3 46 | 1459 2 159 | 1467 467 8 |
| 2478 278 9 | 6 134 138 | 5 2347 134 |
+-----------------+---------------------+-----------------+
| 2478 278 5 | 3 167 126 | 9 4678 146 |
| 479 1 34 | 579 8 569 | 467 34567 2 |
| 6 2789 238 | 12579 1579 4 | 178 3578 135 |
+-----------------+---------------------+-----------------+
| 3 25689 268 | 245789 45679 25689 | 2468 1 4569 |
| 2589 4 1 | 2589 3569 235689 | 268 2568 7 |
| 2589 25689 7 | 124589 14569 125689 | 3 24568 4569 |
+-----------------+---------------------+-----------------+
|
That's it? I thought they said it was 'hard'? This is the toughest they've got? (snicker.) I'm not going to even bother listing my solution here. Come on, they can't come up with anything tougher than this? (snort.) |
|
Back to top |
|
|
dongrave
Joined: 06 Mar 2014 Posts: 568
|
Posted: Sun Mar 27, 2016 2:36 pm Post subject: |
|
|
Of course I was being sarcastic in my last post. I wasn't able to make ANY progress at all after the basics so even though I hate to, I'm going to have to use a computer to tell me if this thing can be solved using only moves that are commonly used on this site (i.e. basics, wings, fish moves, chains, etc.). And while I'm at it, Keith has inspired me to find the 50th Mersenne prime! (I expect it to only contain somewhere around 30 million digits.) |
|
Back to top |
|
|
dongrave
Joined: 06 Mar 2014 Posts: 568
|
Posted: Sun Mar 27, 2016 8:36 pm Post subject: |
|
|
Well, apparently according to the computer it's possible to begin solving this puzzle by showing that r2c6=9. Now I'm not saying that it's easy - but at least it appears to be possible! I didn't check it for correctness (since that would take me the rest of my life) but it was able to show after a couple dozen advanced chains that r2c6 could not be 1 or 5 so it must be 9.
Then after it determined that r2c6=9, it's apparently possible (after another dozen chains) to eliminate the 4 from r5c3 so it must be 3. And then another dozen chains later it showed that r2c1=5. THEN another dozen chains later it showed that r6c3=8 and after that, it ONLY required a few chains to prove that r7c3=2. It was only at this point that it was finally able to solve it - but NOT by using just basics; it STILL required more advanced chains! This puzzle is absolutely insane!!!
Strangely enough, this was the ONLY solution that it found (with my help) using the common moves used on this site - and there were a number of times that it could have used an advanced move other than a chain (like the swordfish that Asellus mentioned and several wings) but I had it focus on chains. If anyone finds anything else (either with or without the use of a computer), I'd love to hear about it! Thanks, Don. |
|
Back to top |
|
|
keith
Joined: 19 Sep 2005 Posts: 3355 Location: near Detroit, Michigan, USA
|
Posted: Sat Apr 02, 2016 7:56 am Post subject: |
|
|
Zombie thread!
I believe there are puzzles much more difficult for a human solver than this one.
I may still have the poster from Cray Research in my attic that has the 29th Mersenne Prime. That filled a 2x4 foot poster with digits in a fairly small font.
Keith |
|
Back to top |
|
|
dongrave
Joined: 06 Mar 2014 Posts: 568
|
Posted: Sat Apr 02, 2016 6:34 pm Post subject: |
|
|
What's it doing in your attic? Why isn't it on your wall? You know Keith - it'll probably take you close to 1000 posters that size to print out the next one when they find it!
Hey! By the way, when I searched the internet looking for 'Al Escargot', I saw that the guy that created it followed it up with a new one that he says is much more difficult - so now apparently they're referring to this one as the 'hardest Sudoku in the world'. Here it is:
Code: |
+-------+-------+-------+
| 8 . . | . . . | . . . |
| . . 3 | 6 . . | . . . |
| . 7 . | . 9 . | 2 . . |
+-------+-------+-------+
| . 5 . | . . 7 | . . . |
| . . . | . 4 5 | 7 . . |
| . . . | 1 . . | . 3 . |
+-------+-------+-------+
| . . 1 | . . . | . 6 8 |
| . . 8 | 5 . . | . 1 . |
| . 9 . | . . . | 4 . . |
+-------+-------+-------+
|
Of course I'd have NO chance of making ANY progress whatsoever so I used my computer tell me if it could be solved using 'human methods' and surprisingly enough, it can! If anyone wants to see how it's possible, let me know. I'd be glad to post it (but only in general terms of course; the entire solution would take forever). Someone said on another site that he didn't believe that this puzzle could be solved by using human methods so I might post it there.
Another interesting thing that I noticed about this puzzle is that it's possible to solve by using chains if you assume the right first value for either of 2 different cells - but since that's not a 'human method', I disredarded them.
It's perfectly okay to assume a wrong value and then use chains to arrive at a contradiction in order to eliminate it - but assuming a right value and then solving the puzzle is not (at least in my mind).
I guess the 'hardest' possible Sudoku in the world would have to be one that could not be solved by using ANY methods used by humans and could ONLY be solved by pure trial and error - and would include so many branches that only a computer could solve it. I'm sure eventually someone (probably this same Finnish mathematician) will create a puzzle where you'll only be able to proceed by assuming a candidate of any given cell and then it would branch out to a number of possibilities which would then branch out again - and then again, etc. until you finally arrive at all of the bottom branches (except one) leading to contradictions.
Anyway, I've rambled on long enough. |
|
Back to top |
|
|
waleeed12
Joined: 26 Mar 2016 Posts: 2
|
Posted: Wed Apr 06, 2016 9:36 am Post subject: Nice |
|
|
There is a finned swordfish on 9 that eliminates four <9>s in B8. After that, I am, so far, stuck.
== Solitaire == |
|
Back to top |
|
|
|