The New Sudoku Players' Forum

[Guest]

|...|52.|...|
|.9.|..3|..4|
|...|...|7..|
|---+---+---|
|.1.|...|.4.|
|.8.|.45|3.1
|6.4|.1.|..8|
|---+---+---|
|7.2|...|...|
|1.8|...|.32|
|.4.|.8.|.17|

I can't find next move
Can you help me ?:P

[Karyobin]

You seem to have posted an asymmetric and uncommonly difficult sudoku. I have no doubt some of our more tenacious residents will apply all manner of techniques to arrive at a solution but for myself, I have to try to get a job and so will honourably bow out of this little challenge.

[tso]

Here is one "next logical step". There may very well be others. This is just what I found first.


You can do a "Nisho" in 3's.

When performing a Nishio, only one digit is considered, in this case, the digit '3'. The solver is looking to see if placing the digit in a particular cell will still allow placement of the rest of the matching digits. For those who don't like the Nisho tactic -- remember, he's the one who invented Sudoku in it's present form and is most responsible for it popularity in Japan. He also invented the "paint-by-numbers" aka "griddlers" aka "a hundred other names" puzzles. (Odd that that puzzle has a hundred names around the world and this has only two.)

See if you can solve it without reading further. If you have Simple Sudoku, SadMan or other software that will allow filtering, try it on 3's.






1) If the 3 in box 8 is r9c4,
2) then the 3 in box 5 will be r4c5, AND the 3 in box 7 will be r7c2.
3) There is no now place for a 3 in box 4. This is a contradiction.

Therefore, we can eliminate the candidate 3 from r9c4.

Another way to look at the same situation is a triple forcing chain.
There are only three cells in box 5 that can contain a 3:

1) r4c4=3 => r9c4<>3
2) r4c6=3 => then r9c4<>3
3) r4c5=3 => r4c13<>3 => r6c2=3 => r7c2<>3 => r9c13=3 => r9c4<>3
Therefore, r9c4<>3.

If you are using Simple Sudoku, filter on 3's and either of these is easy to see. The Nishio is simple enough for most people to do in your head without filtering.

After this, there you will get far before you hit another wall.

[Pappocom]

...remember, he's the one who invented Sudoku in it's present form and is most responsible for it popularity in Japan.

Neither statement is correct. Perhaps you are confusing Nishio and Nikoli?

- Wayne

[PaulIQ164]

And at any rate, just because you invent something doesn't mean you automatically know what's best to do with it.

"And for those who think shooting people is wrong, remember that the guy who invented the gun thinks it's okay."

[Guest]

tso how did you know to search Nisho in 3's ?
did you use pc solver or you check 1's amd 2's first
i never heard about Nisho so i'm asking
p.s sory for my English

[tso]

...remember, he's the one who invented Sudoku in it's present form and is most responsible for it popularity in Japan.

Neither statement is correct. Perhaps you are confusing Nishio and Nikoli?

- Wayne

I made the mistake I warn others about -- I believed what I read (or at least what was implied strongly) in the Times -- though I'm well aware that article they've written on Sudoku has factual errors, exagerations, hyperbole and typos -- as if the TIMES were having an on-going "find 10 errors in this article" contest.

From http://www.timesonline.co.uk/article/0,,7-1757275_2,00.html

"When he does venture outdoors there are certain things he always takes with him. One is a battered, yellowing copy of a 20-year-old American crossword magazine which published the first Su Doku puzzle that Nishio saw. It was included in the magazine as a curio, and never took off in the US. Nishio not only enjoyed the puzzle — as the pencil markings surrounding the grid attest — but gave up his job writing standard logic puzzles to take the Su Doku idea forward. "

Let's count the errors or false implications in this one paragraph from a much longer article:

1) It wasn't a Sudoku. It was a non-symmetrical Number Place puzzle with 36 clues, four per box. The puzzle required little if any logical skills to fill in.

2) It wasn't a crossword magazine. It was Dell Pencil Puzzles and Word Games, a magazine that has few, if any, crossword puzzles. It was then published in Dell's Math Puzzles and Logic Problems. Though most of Dells magazines contain mostly or entirely crosswords, these to do not.

3) It took off just fine. Dell has been publishing without a break since then. Competitors including, but not limited to, Penny Press, also published the puzzle since then. Dell has 27 books of old puzzles.

4) Really? He never goes out without a 20 year old magazine? Really? What are we, 8 year olds? I'll go out on a limb -- this is just a construct to make the Sudoku world seem more interesting -- like the story you tell about being able to make a Sudoku so hard that it couldn't be solved overnight if failure meant execution -- we challenged you to post such a puzzle, but so far, we're all alive -- or your oft quoted “I think there’s something in the British personality — they like their puzzles hard." At least they like they're puzzles to be*labeled* hard, as the recent dumbing-down of both standard and "killer" puzzles in the Times indicates along with the ratings inflation in the TIMES and the TELEGRAPH. "Fiendish?" "Diabolical?" Any puzzle that can be solved in a lunch hour and have time left over to eat doesn't deserve such a rating. Certainly no one who has competed in the World Puzzle Championships would use these words in this context.

Anyway, here is Tetsuya Nishio -- I don't see the magazine.

5) I inferred -- incorrectly -- that "take the Su Doku idea forward" meant that he was the one that changed it from Number Place to Sudoku by adding symmetry and reducing the number of clues. My mistake, though it seems as if it was written in such away as to allow, or even encourage, this implication. I assumed that he was the one at Nikoli who refined the puzzle. Mia Culpa. It still isn't clear who made the changes, but it is clear that Kaji Maki at Nikoki was the one who coined the word "Sudoku".

In the same article, they spell the creator of Samunamupure wrong -- It's Miyuki Misawa with an *M* not an *N*. http://www7a.biglobe.ne.jp/~sumnumberplace/93668566/

Spelling someone's name wrong is *not* a small error. The entire article is devoted to her creation -- she's mentioned once, they get her name wrong, and spend the rest of the time talking about Nishio. I wonder, if Nishio were a woman and Misawa were a man, would this be the case?

And of course, it took weeks for the Times to finally get the rules correct.

[tso]

And at any rate, just because you invent something doesn't mean you automatically know what's best to do with it.

"And for those who think shooting people is wrong, remember that the guy who invented the gun thinks it's okay."

Good point. (Moot now, as I was wrong about Nishio being the one to turn Number Place into Sudoku.)

tso how did you know to search Nisho in 3's ?
did you use pc solver or you check 1's amd 2's first
i never heard about Nisho so i'm asking
p.s sory for my English

I used Simple Sudoku to filter for 1's, then 2's, then 3's. The arrangement of the cells that had the candidate 3 in boxes 4 and 7 looked very promising, though I cannot say why other than experience.

If I were solving this for fun, I would have been using paper and pencil and wouldn't have been able to 'filter'. I don't know how I would have handled it. I *think* the situation with the 3's would have been obvious enough to see.

But as to why I would look at 3's instead of 7's -- or anywhere else -- you can *always* ask this question, even if the puzzle is fairly simple. You will look in several spots between each placement of a number and you won't always have an obvious reason for the path you follow.

[angusj]

If you have Simple Sudoku, SadMan or other software that will allow filtering, try it on 3's.

I've fairly recently updated Simple Sudoku to filter on Big Numbers when "Show Candidates" is off. (This makes it much easier to solve those easier puzzles when you don't want to use candidates at all.) It also makes it easier to spot Nishio:


Here's another example:

[Jeff]

When performing a Nishio, only one digit is considered, in this case, the digit '3'. The solver is looking to see if placing the digit in a particular cell will still allow placement of the rest of the matching digits.

Tso, can you help to clarify this one. The diagram below indicates a similar approach starting from r6c6=2 and finally forced a contradiction at r9c2. Since this operation involves 2s and 7s, can this still be called 'Nisho'.

[tso]

1) No. Nishio is specificaly one number. Ask "If I place this number here, will I be able to place the rest of the matching numbers?" If not, eliminated the candidate.

2) Could you write out the implication chain that leads from r6c6=2 to a contradiction at r9c2? I can't see it.

[Jeff]

It's not a chain, but a net.

r6c6=2 => r3c6=7 => r2c2=7 => r9c2=2
r6c6=2 => r7c4=2 => r8c8=2 => r3c7=2 => r2c1=2 (since r2c2=7) => r4c2=2 => r9c2<>2
Therefore r6c6<>2

[tso]

It's not a chain, but a net.

r6c6=2 => r3c6=7 => r2c2=7 => r9c2=2
r6c6=2 => r7c4=2 => r8c8=2 => r3c7=2 => r2c1=2 (since r2c2=7) => r4c2=2 => r9c2<>2
Therefore r6c6<>2

That works -- it's more than I can do in my head. It isn't a Nishio. If you filter on 2's alone, you'll have no trouble placing all nine 2's.

Is there a definition for "net" in this context?

[Jeff]

If you filter on 2's alone, you'll have no trouble placing all nine 2's.

Not without r2c2 having been determined as 7.

Is there a definition for "net" in this context?

Each chain enforces one outcome and has all cells in operation forming a loop or 2 loops for double implication or triple implication respectively. 'Net' is not a defined term; it is used to describe a network of chains interconnecting to force one outcome.

[Pat]

.

Code: Select all

 . . . | 5 2 . | . . .
 . 9 . | . . 3 | . . 4
 . . . | . . . | 7 . .
-------+-------+------
 . 1 . | . . . | . 4 .
 . 8 . | . 4 5 | 3 . 1
 6 . 4 | . 1 . | . . 8
-------+-------+------
 7 . 2 | . . . | . . .
 1 . 8 | . . . | . 3 2
 . 4 . | . 8 . | . 1 7

1) If the 3 in box 8 is r9c4,
2) then the 3 in box 5 will be r4c5, AND the 3 in box 7 will be r7c2.
3) There is no now place for a 3 in box 4. This is a contradiction.

Therefore, we can eliminate the candidate 3 from r9c4.

that's an example of a "net" -
following 2 (or more) consequences of the assumption,
eventually combining the results to reach the contradiction.


by the way, in that example,
a single chain is sufficient:

→ the 3 for box 8 is in r9 (our assumption)
→ the 3 for box 7 is in c2
→ the 3 for box 4 is in r4
→ the 3 for box 5 is nowhere (contradiction)


sorry if i seem to be repeating what's obvious to everyone---

i just prefer the single chain (wherever possible), seems easier to follow

though i do wonder how long it might take to discover such a chain

.