The New Sudoku Players' Forum

[jraggio]

I am new to Sodoku and have a few questions. I appologize as I'm sure much of this has probably been discussed before, but I was unable to find a sticky or FAQ here.

1) Is this one of the better discussions boards or forums for Sodoku? It's one of the first that showed up in Google and seems to have some good conversations and recent posts. Are there any others I should check out?

2) Are there puzzles that can only be solved with some of the more advanced techniques like X-Wing and Jellyfish (or whatever they are called)? Or are these techniques simply names for things that can be done without learning the techniques using common sense and logic? Is it he case that these techniques make the puzzle easier to solve or is it the case that they are necessary? I hope it is not the case. I don't like the idea of needing to learn them all to be able to complete the harder puzzles. At what level woudl the techniques become mandatory, if ever?

3) Are there any puzzles that require trial and error? In other words are there some that no techniques could help.

4) Is it true that every legitmate puzzle should have only one solution?

Thanks,
John

[fermat]

I am new to Sodoku and have a few questions. I appologize as I'm sure much of this has probably been discussed before, but I was unable to find a sticky or FAQ here.

1) Is this one of the better discussions boards or forums for Sodoku? It's one of the first that showed up in Google and seems to have some good conversations and recent posts. Are there any others I should check out?

2) Are there puzzles that can only be solved with some of the more advanced techniques like X-Wing and Jellyfish (or whatever they are called)? Or are these techniques simply names for things that can be done without learning the techniques using common sense and logic? Is it he case that these techniques make the puzzle easier to solve or is it the case that they are necessary? I hope it is not the case. I don't like the idea of needing to learn them all to be able to complete the harder puzzles. At what level woudl the techniques become mandatory, if ever?

3) Are there any puzzles that require trial and error? In other words are there some that no techniques could help.

4) Is it true that every legitmate puzzle should have only one solution?

Thanks,
John

1. The best.

2. Many, many and millions, even trillions. It is all logic, people developed the ideas, you could figure out all the methods already figured out, it just takes time.

3. I think there may be, as we learn new ideas, probably none.

4. Oh yes, I got in some embarrasment about this. A valid Sudoku has only one solution.

[tso]

2) Are there puzzles that can only be solved with some of the more advanced techniques like X-Wing and Jellyfish (or whatever they are called)?

Yes and no. Yes, there are puzzles that require more complicated tactics than what are required to solve nearly all the puzzles published in all the newspapers, books and magazines. But if you stick to puzzles in newpapers, puzzle magazines and books -- with few exceptions, you'll rarely ever need anything beyond the most basic tactics. And even if you *do* come across more difficult puzzles, if you are clever enough, you may be able to solve them without studying anything in this forum, just as you can become quite a good chess player without ever picking up a book or knowing what a "Ruy Lopez" is. You very well might enjoy the learning process more anyway.

Or are these techniques simply names for things that can be done without learning the techniques using common sense and logic?

Logic yes, common sense, probably not. There's a lot of smarty-pants types posting here. I've been solving Sudoku much longer than most -- since before they were called Sudoku. I thought I had mastered everything there was to know about them a *decade* ago -- but some of the tactics discussed here are beyond me -- far beyond me.

Is it he case that these techniques make the puzzle easier to solve or is it the case that they are necessary?

There are puzzles that are very likely to stump you unless you practice or study Sudoku for a very long time ... the learning process will be shortedn by reading what others have discovered ... whether you chose to shand on the shoulders of those standing on other's shoulders so that you can crack the harder puzzles sooner, or chose to figure it all out your self is your choice. Both ways have advantages. But again, at least for the time being, nearly all puzzles published in newspapers, magazines and books rarely require much beyond the basic tactics.

I hope it is not the case. I don't like the idea of needing to learn them all to be able to complete the harder puzzles.

I never understand that attitude. Should all puzzles be easy enough that all players will be able to solve them within minutes of learning the rules? Should some require an hour of practice? Shouldn't some require a week of practice, etc? Should I be unhappy not to be able to complete a challenger diagrammless crossword puzzle without *years* of practice on easier puzzles? Can't you swim just fine in the shallow end of the pool knowing that the deep end is very, very deep?

At what level woudl the techniques become mandatory, if ever?

When you find a puzzle you cannot solve by the tactics you know.

3) Are there any puzzles that require trial and error? In other words are there some that no techniques could help.

On the one hand, even the simplest tactics are trial and error -- they differ only by degree with the most complicated tactics. On the other hand, maybe the question is -- are there puzzles so hard than any use of tactics by anyone other than a savant will reach the solution no more quickly using basic tactics and guessing when blocked? Absolutely -- puzzles much harder than that are disected here. That doesn't at all take away from the fascination of finding new and more complicated -- and sometimes simpler -- techniques and patterns.

[4) Is it true that every legitmate puzzle should have only one solution?

Sudoku's are defined as having a unique solution -- unless otherwise specified. If a puzzle is specifically labeled as having 2 solutions, that's perfectly valid. There dozens of variants that break one or more of the basic rules of the puzzle.

[fermat]

WOW, tso, a lawyer for sure.

[tso]

WOW, tso, a lawyer for sure.

I'd make the *worst* lawyer. I couldn't live in the legal world. The law *pretends* to be about logic -- but then it wouldn't matter which judge you got, or how expensive your lawyer was, right? I don't think I'd be happy if the rules of mathematics were decided by 5-4 decisions.

Ob-puzzle for John:

Here's an example of a puzzle that is requires a tactic that you will rarely require in newspaper, magzine or book puzzles (The LA Times *does* include puzzles like this.) The puzzle is *not* difficult, taken as a whole, but if you don't know how to apply that last tactic, you will either give up or have to take a guess. But you may find that this last logical step is by far, the most fun part of the puzzle and that you'll want to find other puzzles that require this type of logic. Try to solve it before you read the hints below.

Code: Select all

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


1) You should be able to solve all but 19 cells by basic methods -- "singles".


2) A "forcing chain" is the advanced tactic that solves the puzzle -- a 5 cell long xy-type forcing chain to be specific.


3) Here's what the puzzle will look like after all the singles have been filled in and you've written in candidates in the remaining open cells:

Code: Select all

+-------------+-------------+-------------+
| 478 789 5   | 6   3   89  | 1   2   47  |
|+48  1   2   | 7   5   89  |+49  6   3   |
| 67  679 3   | 1   2   4   |+79  5   8   |
+-------------+-------------+-------------+
| 3   67  1   | 8   46  2   | 5   9   467 |
|-678 2   4   | 9   1   5   |+78  3   67  |
| 5   68  9   | 3   46  7   | 48  1   2   |
+-------------+-------------+-------------+
| 2   3   7   | 4   9   1   | 6   8   5   |
| 9   5   8   | 2   7   6   | 3   4   1   |
| 1   4   6   | 5   8   3   | 2   7   9   |
+-------------+-------------+-------------+

Take a look at the four cells marked with a plus sign. I think you can convince yourself that all values in any of these four cells eliminate the '8' from r5c1. For example, if r2c1=4, then r2c7=9, then r3c7=7, then r4c7=8, then r5c1 cannot also be 8. On the other hand, if r2c1=8, then r5c1 *also* cannot be 8. Since r2c1 has to be one or the other -- then though we don't know which it is, we do know that r5c1 cannot be 8. That's a simple forcing chain.

[ravel]

3) Are there any puzzles that require trial and error? In other words are there some that no techniques could help.

A form of trial and error is needed for all harder puzzles (like for crosswords), but no "guessing" (setting a number without having a proof, that it must be there).

To get a feeling, how hard puzzles can be, you can look at the solutions of the 2 hardest there. You will see, that it will even take some time to be able to understand them.

[Carcul]

Hi Jraggio. Welcome to this forum.

1) Is this one of the better discussions boards or forums for Sodoku? It's one of the first that showed up in Google and seems to have some good conversations and recent posts. Are there any others I should check out?

I must agree with Fermat: as far as I am concern, this is the best Sudoku forum available. However, you may like to check the following ones:

The Sudoku Progammers Forum
Henk's IntoSudoku Forum
The Sudoku Discussions Forum

Carcul

[jraggio]

Thanks so much for the replies folks. I appreciate your time, especially tso who wrote the very lengthy response. I'd like to clarify my original statement of hoping that it is not the case that you need to learn the advanced techniques to solve even the hardest puzzles. What I meant to say was that I hope that I would be able to solve a very hard puzzle without learning what an x wing or jellyfish was, but rather apply some of that logic without knowing the official tactic. I don't mind being challenged, or even frustrated, but I don't think it is right that one needs to learn a jellyfish to solve a puzzle. I agree that is is cool if it makes it easier, but not mandatory. I don't like the idea of there being the possibility for a dead end unless I learn the advanced tactics.

One more thing, what class are the puzzles that generally require the advanced techniques? Are they "devilish" and "fiendish" only? Are there standard ratings available that I coould use as a guide. i basically want to avoid starting a puzzle that I can't solve (eventually anyway)

Thanks,
John

2) Are there puzzles that can only be solved with some of the more advanced techniques like X-Wing and Jellyfish (or whatever they are called)?

Yes and no. Yes, there are puzzles that require more complicated tactics than what are required to solve nearly all the puzzles published in all the newspapers, books and magazines. But if you stick to puzzles in newpapers, puzzle magazines and books -- with few exceptions, you'll rarely ever need anything beyond the most basic tactics. And even if you *do* come across more difficult puzzles, if you are clever enough, you may be able to solve them without studying anything in this forum, just as you can become quite a good chess player without ever picking up a book or knowing what a "Ruy Lopez" is. You very well might enjoy the learning process more anyway.

Or are these techniques simply names for things that can be done without learning the techniques using common sense and logic?

Logic yes, common sense, probably not. There's a lot of smarty-pants types posting here. I've been solving Sudoku much longer than most -- since before they were called Sudoku. I thought I had mastered everything there was to know about them a *decade* ago -- but some of the tactics discussed here are beyond me -- far beyond me.

Is it he case that these techniques make the puzzle easier to solve or is it the case that they are necessary?

There are puzzles that are very likely to stump you unless you practice or study Sudoku for a very long time ... the learning process will be shortedn by reading what others have discovered ... whether you chose to shand on the shoulders of those standing on other's shoulders so that you can crack the harder puzzles sooner, or chose to figure it all out your self is your choice. Both ways have advantages. But again, at least for the time being, nearly all puzzles published in newspapers, magazines and books rarely require much beyond the basic tactics.

I hope it is not the case. I don't like the idea of needing to learn them all to be able to complete the harder puzzles.

I never understand that attitude. Should all puzzles be easy enough that all players will be able to solve them within minutes of learning the rules? Should some require an hour of practice? Shouldn't some require a week of practice, etc? Should I be unhappy not to be able to complete a challenger diagrammless crossword puzzle without *years* of practice on easier puzzles? Can't you swim just fine in the shallow end of the pool knowing that the deep end is very, very deep?

At what level woudl the techniques become mandatory, if ever?

When you find a puzzle you cannot solve by the tactics you know.

3) Are there any puzzles that require trial and error? In other words are there some that no techniques could help.

On the one hand, even the simplest tactics are trial and error -- they differ only by degree with the most complicated tactics. On the other hand, maybe the question is -- are there puzzles so hard than any use of tactics by anyone other than a savant will reach the solution no more quickly using basic tactics and guessing when blocked? Absolutely -- puzzles much harder than that are disected here. That doesn't at all take away from the fascination of finding new and more complicated -- and sometimes simpler -- techniques and patterns.

[4) Is it true that every legitmate puzzle should have only one solution?

Sudoku's are defined as having a unique solution -- unless otherwise specified. If a puzzle is specifically labeled as having 2 solutions, that's perfectly valid. There dozens of variants that break one or more of the basic rules of the puzzle.

[udosuk]

Another question is, is there a particular finite set of techniques that will enable you to solve all sudoku puzzles? For example, if one only knows hidden/naked singles and forcing chains (or nishio), but knows nothing about pairs/triples/quads/locked candidates/wings/fishes/colors/unique rectangles, could he still solve all puzzles?

We know the particular set used by simple sudoku is not good enough to solve them all...

Think of it this way: a golfer is only allowed 14 clubs in his bag... Which 14 would you be choosing?

[ab]

Another question is, is there a particular finite set of techniques that will enable you to solve all sudoku puzzles?

It's very easy to write a program that solves all 3x3 sudokus. I have as have many others, so arguably the techniques embodied in them fit the bill. However those techniques are just singles and then try to place each candidate in one cell in turn until you find a cell where all but one candidate leads to a contradiction, you can then place that candidate in the cell. and continue in this manner until the puzzle is solved. This technique is easy for computers to employ but not so easy for humans to carry out, whereas the arsenal of techniques collected in Mike Barker's post are relatively easy for humans to use, but may not solve all sudokus.

[jraggio]

Another question is, is there a particular finite set of techniques that will enable you to solve all sudoku puzzles?

It's very easy to write a program that solves all 3x3 sudokus. I have as have many others, so arguably the techniques embodied in them fit the bill. However those techniques are just singles and then try to place each candidate in one cell in turn until you find a cell where all but one candidate leads to a contradiction, you can then place that candidate in the cell. and continue in this manner until the puzzle is solved. This technique is easy for computers to employ but not so easy for humans to carry out, whereas the arsenal of techniques collected in Mike Barker's post are relatively easy for humans to use, but may not solve all sudokus.

The golf bag analogy above was excellent. I think that was the theme of my original question. I could hit the course and complete a round with just a putter, but I would have a real hard time scoring well.

Can you please give me a link to Mike Barker's post? I am not familiar with it.

Thanks,
J

[Cec]

"...Can you please give me a link to Mike Barker's post? I am not familiar with it ..."

I think this is it.. Click Here

Cec

[tso]

One more thing, what class are the puzzles that generally require the advanced techniques? Are they "devilish" and "fiendish" only? Are there standard ratings available that I coould use as a guide. i basically want to avoid starting a puzzle that I can't solve (eventually anyway)

No, there are no standard ratings. Ratings can be usually be useful comparing puzzles from a single source, though it seems as if some newspapers sometimes apply the ratings at random.

If you use the software available from this site, you'll never have to learn any advanced tactics for the puzzles rated Very Easy, Easy, Medium or Hard. The Very Hard Puzzles can be tough to solve unless you are familiar with the X-wing tactic. Many newspapers uses puzzles created by this program. The UK Times "Fiendish" is a Pappocom "Hard"; the "Superior" is a "Very Hard". Very few newspapers, magazines or books publish puzzles beyond the level of Pappocom's Very Hard.

To continue the golf analogy, one set of 14 clubs not only won't be right for every course, they won't be right for every golfer. Each golfer has different strengths and weaknesses, as well as different preferences. Two solvers will often disagree which of two puzzles was harder to solve.

Personally, I think it's very condescending for the publishers to use ratings like "Diabolical", "Evil", "Fiendish", "Insane", "Migraine" "Deadly", etc, especially when often those puzzles can be solved by the typical commuter on a bus with a pen. If I didn't know better, I'd just assume that something with these hyperbolic ratings would only be solvable by a small percentage of the population, and then only after many, many hours. If everyone can solve every puzzle, how hard can they be? And why did the Times decide to call Samunamupure puzzles "Killers" -- even the easy ones? You'll never find a crossword puzzle with these ridiculous ratings -- they seem to have greater respect for the traditional word puzzle world.

[jraggio]

Tso:

I finally took the time to read this sample and try it out. I think I get it, but think that I would have ahard time seeing this on my own.

I am actually still struggling with why r1c9 and r4c9 do not form a hidden pair. I thought they would as those are the only two cells that have 4&7 together. Then I realize that the cells in col 9 actually form a naked triple. Is that correct? Is it true for r1c9 and r4c9 to from a hidden pair that r5c9 could not have had a 7 in it?

Thanks
John

WOW, tso, a lawyer for sure.

I'd make the *worst* lawyer. I couldn't live in the legal world. The law *pretends* to be about logic -- but then it wouldn't matter which judge you got, or how expensive your lawyer was, right? I don't think I'd be happy if the rules of mathematics were decided by 5-4 decisions.

Ob-puzzle for John:

Here's an example of a puzzle that is requires a tactic that you will rarely require in newspaper, magzine or book puzzles (The LA Times *does* include puzzles like this.) The puzzle is *not* difficult, taken as a whole, but if you don't know how to apply that last tactic, you will either give up or have to take a guess. But you may find that this last logical step is by far, the most fun part of the puzzle and that you'll want to find other puzzles that require this type of logic. Try to solve it before you read the hints below.

Code: Select all

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


1) You should be able to solve all but 19 cells by basic methods -- "singles".


2) A "forcing chain" is the advanced tactic that solves the puzzle -- a 5 cell long xy-type forcing chain to be specific.


3) Here's what the puzzle will look like after all the singles have been filled in and you've written in candidates in the remaining open cells:

Code: Select all

+-------------+-------------+-------------+
| 478 789 5   | 6   3   89  | 1   2   47  |
|+48  1   2   | 7   5   89  |+49  6   3   |
| 67  679 3   | 1   2   4   |+79  5   8   |
+-------------+-------------+-------------+
| 3   67  1   | 8   46  2   | 5   9   467 |
|-678 2   4   | 9   1   5   |+78  3   67  |
| 5   68  9   | 3   46  7   | 48  1   2   |
+-------------+-------------+-------------+
| 2   3   7   | 4   9   1   | 6   8   5   |
| 9   5   8   | 2   7   6   | 3   4   1   |
| 1   4   6   | 5   8   3   | 2   7   9   |
+-------------+-------------+-------------+

Take a look at the four cells marked with a plus sign. I think you can convince yourself that all values in any of these four cells eliminate the '8' from r5c1. For example, if r2c1=4, then r2c7=9, then r3c7=7, then r4c7=8, then r5c1 cannot also be 8. On the other hand, if r2c1=8, then r5c1 *also* cannot be 8. Since r2c1 has to be one or the other -- then though we don't know which it is, we do know that r5c1 cannot be 8. That's a simple forcing chain.

[tso]

I am actually still struggling with why r1c9 and r4c9 do not form a hidden pair.

Code: Select all

+-------------+-------------+-------------+
| 478 789 5   | 6   3   89  | 1   2   47  |
| 48  1   2   | 7   5   89  | 49  6   3   |
| 67  679 3   | 1   2   4   | 79  5   8   |
+-------------+-------------+-------------+
| 3   67  1   | 8   46  2   | 5   9   467 |
| 678 2   4   | 9   1   5   | 78  3   67  |
| 5   68  9   | 3   46  7   | 48  1   2   |
+-------------+-------------+-------------+
| 2   3   7   | 4   9   1   | 6   8   5   |
| 9   5   8   | 2   7   6   | 3   4   1   |
| 1   4   6   | 5   8   3   | 2   7   9   |
+-------------+-------------+-------------+

A naked pair is two cell each of which have the same two candidates. Cells r1c9 and r4c9 have *three* candidates between them. You are correct that the three cells r145c9 form a naked triple in column 9. No two of those three cells are a naked pair. You may be conflating naked pairs with *hidden* pairs.

In this row, the first two cells are a naked pair. You can exclude the 1's and 2's from the rest of the row:

Code: Select all

Before: [12][12][12345][12678] ...
After: [12][12][xx345][xx678] ...

In this row, the digits 1 and 2 appear ONLY in the first two cells. You can exclude all other candidates from these first two cells:

Code: Select all

Before: [12345][124567][<>1,2][<>1,2][<>1,2][<>1,2][<>1,2][<>1,2][<>1,2]
After:[12x][12xx][<>1,2][<>1,2][<>1,2][<>1,2][<>1,2][<>1,2][<>1,2]