The New Sudoku Players' Forum

by **Max Beran** » Sat Oct 15, 2005 11:23 am

Some while back Tso posed the following in response to someone asking for harder puzzles:

xxx|1xx!7xx
x2x|69x|xxx
9xx|xx3|x82
========
xxx|xxx|46x
64x|xxx|x57
x58|xxx|xxx
========
21x|8xx|xx9
xxx|x16|x7x
xx4|xx2|xxx

Firstly, has a solution (short of tabling) been posted for this one?

Secondly, while chasing round the bilocation/bivalue plot I noticed that there were occasions when one can go from a multi-candidate bilocation cell to a bivalue cell with quite useful looking short term results. Obviously this is directional and in no case actually led to any resolution about a cell but it left me wondering if there are generalisable circumstances where this might be useful, say for non-cyclic conflicting cell cases.

by **Myth Jellies** » Sun Oct 23, 2005 5:15 pm

Working on this using the pattern overlay method. Should have a result shortly.

by **Max Beran** » Sun Oct 23, 2005 6:05 pm

Having acquired Sudoku susser I pasted this puzzle into it and tabling was required to be switched on before the puzzle could be completed.

I don't know nothing about pattern recognition so await results with baited breath as it promises to boldly go ........

by **robertsetiadi** » Mon Oct 24, 2005 5:50 am

Hmmm, I'm a new person here.

[4,3,5,1,2,8,7,9,6]
[8,2,1,6,9,7,5,4,3]
[9,7,6,5,4,3,1,8,2]
[1,9,2,7,3,5,4,6,8]
[6,4,3,2,8,1,9,5,7]
[7,5,8,4,6,9,3,2,1]
[2,1,7,8,5,4,6,3,9]
[5,8,9,3,1,6,2,7,4]
[3,6,4,9,7,2,8,1,5]

<can I post a solution here?>

by **emm** » Mon Oct 24, 2005 7:59 am

Since this looks like it's becoming a trend with you, possum, do you mind if I say that posting the solution is a bit like giving someone a Christmas present already unwrapped and after you’ve had a little play with it. Mostly we like them still in the box with maybe just the smallest of hints about what’s inside.

by **Karyobin** » Mon Oct 24, 2005 12:38 pm

Three solutions have been posted by robertetc... in one day. Is it possible that this person is falling prey to the idea that - 'They're discussing incomplete puzzles - this must mean they can't do them. I'll look really clever if I just post completed grids!'?

by **stuartn** » Mon Oct 24, 2005 1:14 pm

Is it half-term in Oz aswell? ....read my mind

stuartn

by **Myth Jellies** » Tue Oct 25, 2005 4:52 am

The Pattern Overlay Method does find the solution. I've solved it, then restarted the puzzle and used my notes to solve it again. I have to comb through and reorganize about 10 pages of handwritten stuff before I present it though.

by **Myth Jellies** » Sun Oct 30, 2005 7:12 pm

Here is a link to the promised POM solution/tutorial

http://www.sudoku.org.uk/discus/messages/2/323.html?1130695682

by **Sue De Coq** » Sun Oct 30, 2005 8:28 pm

That's very interesting - do you have any source code or did you do all the analysis by hand?

by **stuartn** » Sun Oct 30, 2005 9:59 pm

Blimey MJ - that must've taken HOURS.........I'm sure there's a quicker method - albeit a less ELEGANT one. Triffik work.

stuartn

by **Myth Jellies** » Sun Oct 30, 2005 11:33 pm

Sue,

All of it was done by hand. Some of the steps might be amenable to a computer program, certainly creating the merge grid would be, but I am not sure that all of them could be computerized very easily. There is a fair bit of set theory that just gets used as it comes up here. With POM there are a lot of ways to skin the cat, though, so it may very well be that a few POM algorithms might be all you really need to solve the puzzle.

Stuart,

It definitely took many hours. The way the patterns get eroded away, though, I always felt like I was making progress. I always felt like I had another thing to try and that things would solve if I was just smart enough to put the pieces together. Altogether it was a very satisfying experience.

The New Sudoku Players' Forum

Hard puzzle and combination chain query

Hard puzzle and combination chain query