Nice Loop general

ivoc · Posted: Fri Mar 05, 2010 6:55 am Post subject: Nice Loop general

I Like to discuss about Method Nice Loop. But first we should make a temporary Understanding, what is "The best Sudoku Solver"? Take this:

1. The Solver is using only logical Methods
2. The Solver can solve a certain Sudoku Grid, all other known Solvers NOT (without Guessing)
3. The Speed is on the third Place (within the acceptable Borders, of course)

If OK:

NOT many Solvers use the logical Nice Loop Method. If your Solver do know (use) it, you will agree, this Method is (combined with his Sisters Double- and Triple-Implication) the most powerfull one, the most demanding, but other Site - it uses only pure basic Sudoku Definition direct.

Well, this Time I would only ask you, can your Sudoku Solver solve next Grids (found on NET) without Guessing:

1.......9..7.8...6...3...5.....4.8..7....1....96........82.....5......63.6..7.4..
...7.....8.1..6.....9..54.829......3..8..39..6..2...7.9.........8...9.25.32.6.7..
8.....9...4..6..3...61....8..15....7.8..7....9....3...5....23...6.....4...86..7.5
9..7..1...2..1...7...5..6........56......32..1...5...824..813...1542...6..9.....2
7..9....6....6..83.....32..8....9....5.....3...24....1..8...6...6..1...741...5.9.
.4.3..9...7...5.2.8........6.....352..5....9...4...6.....9....8....42.7..9.16....
.6.....7.87.3..59...54..2....9.564.....9...8.2.78.....7....4..2..8...71.........9
.32..........2..4.9.7..5....1...3.98...2..6...68.1..2....53....57......3....471..
.3.....82..2.9......4..67......43....1....5..7..6.5.2..9.2....6..65....9......31.
..3.4..1....192......3....7..4.6....5...34..2.8.9.........7...67......58.29....3.
..23.....85..4....97.5.1...4.....89..2.91...3.....4..6......5.9.....5.1...7.3...8

Next time about Method WeddingRing, a simple Upgrade of Nice Loop, the same Logic - other Philosophie, wich solves all the Grids above (without Guessing).

Regards IvoC

hobiwan · Posted: Fri Mar 05, 2010 7:59 am Post subject: Re: Nice Loop general

ivoc · Posted: Sun Mar 07, 2010 12:40 pm Post subject: Re: Nice Loop general

hobywan,

Would you be so kind please and solve (only) one of written grids and write step by step the used methods. A "logical" solver shows it automatically if requested! Maybe I ask too much? Sorry!

However, I'll be very verry obligated to you!

Regards IvoC

P.S. Now I know! You inserted a link. Thanks, but it is why my (quoted) answer are not accepted!

hobiwan · Posted: Sun Mar 07, 2010 9:00 pm Post subject:

Sure, but I have to warn you, some of the steps are pretty ugly.

ivoc · Posted: Mon Mar 08, 2010 6:07 am Post subject:

Sorry, hobywan, I'm deeply sorry! I don't know, how it has happened.

Writting my solver, I have had as a "model" the best accessible sudoku solver I found on NET, named InToSudoku (BTW, now this site is "half alive" only, no forum more and mail contact adress is not valid. What a loss!).

By mistake I have written the grids - the solver InToSudoku do need guess, my solver doesn't. Of course, all this grids are pretty ugly! To solve them in addition to standard methods the logical method WeddingRing is implemented and used, a powerful upgrade of Nice Loop.

However, my solver really doesn't solve next grids without guess:

.....56....5.....76..9...3...7.....42...6.9...8.....6.1.28...9......3...3....7..5
.6.9....35....4.2.....8.4..8......5...3...7...9......1..1.5.....7.3....99....2.4.
..29......3..8....4....1.5...9.7...2.1.....3.6.....4.......3..1....6..8...72..9..

and many others, of course. From all 633 grids you have sugested my solver doesn't solve 11 without guess. Pretty good, isn't it?

Thanks for your kindness and regards IvoC

hobiwan · Posted: Tue Mar 09, 2010 9:52 pm Post subject:

If your solver really can solve all but 11 puzzles with SE > 9.5 from the Patterns Game then I am really impressed!

I dont understand why it cant solve the first of the three last puzzles you posted though: It doesnt even need a Forcing chain.

ivoc · Posted: Wed Mar 10, 2010 1:54 pm Post subject:

strmckr · Posted: Fri Mar 12, 2010 5:59 am Post subject:

ivoc · Posted: Fri Mar 12, 2010 7:38 am Post subject:

hobiwan · Posted: Fri Mar 12, 2010 6:30 pm Post subject:

Hi ivoc,

my solver (http://hodoku.sourceforge.net) uses

for the puzzle in question. Perhaps you could post the state of the puzzle where your solver uses NoLogic the first time?

I think what strmckr meant is, that it would be nice to get an explanation what "Wedding Ring" is (perhaps with a few examples). I at least have no idea what this technique is and how it works.

ivoc · Posted: Sat Mar 13, 2010 8:25 am Post subject:

It seems, nobody but us is interested in this theme. Not to boring the others and because my solver shows the steps/situations in graphical and PDF modus only, I suggest to continue in private mail modus.

If you agree, ivo.coz@gmail.com

Regards IvoC

tarek · Posted: Sat Mar 13, 2010 2:20 pm Post subject:

ivoc,

if you want to conduct a private conversation then you should initiate it. There is the PM button next to each user if needed.

I have no problem with threads that have even only 1 person posting in them ... I guess if nobody else states otherwise then you should assume that everybody feels the same.

tarek

hobiwan · Posted: Sat Mar 13, 2010 7:08 pm Post subject:

ivoc · Posted: Mon Mar 15, 2010 8:04 am Post subject:

Hi, hobiwan,

I will explaine, of course - no problem!

But thanks to you, I found in my solver a bug, in Unique Rectangle method, and I fixed it. Now from all 633 grids you suggested, running my solver only 7 need guess - thanks to method Wedding Ring, I hope so!

Voila:

070060900006002030020000006003100040000200008050034000900003020000700105008000000
010020300004003050060000001005700060000800002070012000400005090000400805007000000
000012300000400000105006700706000020800050003090000806008300609000008000004560000
000012300000400000506007100201000080800030007090000501008100706000008000004750000
000012300000400000401005600705000010600020007030000806003600109000001000006590000
000000000001234500020605030014000260050000010073000890040709020002413700000000000
010000020003000400000105000004060700020000080007040300000203000006000900070000010

I still think, the method Wedding Ring is a powerful method. But to "certify" my conviction I need the parallel tests. How many grids with NoLogic finds hodoku?

Would you do it using your solver, or anybody other kind person, please?

Regards IvoC

ivoc · Posted: Mon Mar 15, 2010 8:13 am Post subject:

*There is the PM button next to each user if needed.*

I know! But is NOT polite to send to unknown person such a message without his permission. The concrete result is evident!

Regards IvoC