The New Sudoku Players' Forum

[Pappocom]

The Daily Telegraph's sudoku puzzles are sudoku-lookalikes but they are not Pappocom puzzles.

1. DT puzzles expect you to guess or use trial-and-error. Indeed, many of them cannot be “solved” unless you do guess.

2. They have been known to have alternative or multiple solutions (12 Feb 2005, four solutions).

3. The difficulty label (Moderate, Tough, etc.) is not a very reliable guide to the puzzles' difficulty.

Pappocom does not supply puzzles to the Daily Telegraph (UK).

[Guest]

Hi

Have sent email to daily telegraph saying how it is so annoying that their puzzles cannot be solved other than by guessing. In fact, we put the numbers into your programme by dubbing and it would not verify it - that is why we were having problems solving it!!

M

[Guest]

Dt puzzles are sloveable without the use of guessing. I have solved all the diabolical and tough puzzles which i have tried to do and have managed to slove them without the use of trial and error, just using a slightly higher level of logic...

[Sue De Coq]

Two points:

The (quite useful) notes published by Michael Mepham - the Daily Telegraph composer - at http://www.sudoku.org.uk specifically state that guesses will be necessary in order to solve the Tough and Diabolical puzzles.

The existence of multiple solutions dictates that it must be necessary to guess in order to solve the puzzle. To solve a puzzle by 'pure logic' involves the traversal of a sequence of moves that lead to some fixed final position.

[Guest]

There was a single puzzle with multiple solutions, I believe (no.3), before prizes were introduced. Clearly once a prize was introduced, it was necessary to have a unique solution, and all subsequent puzzles have had, and will have, a unique solution.

Since, in fact, there is a unique solution, any puzzle can be solved by logic! Some of the logic can get quite deep in the DT puzzles, so that effectively one may have to postulate a possibility and follow through the consequences rather further than in Pappocom puzzles. Sometimes this makes a nice change; at other times, it is not always completely satisfying (but it usually isn't necessary). It depends on my frame of mind!

[Sue De Coq]

Much of the debate on the virtues of 'guessing' versus 'pure logic' flounders on inexact definitions of the two terms. I believe that Pappocom would take

to postulate a possibility and follow through the consequences

to mean a guess. Others would simply regard it as reductio ad absurdum - or logic.

I don't quite agree with your statement:

Since, in fact, there is a unique solution, any puzzle can be solved by logic!

Even when there are multiple solutions, they are obtainable by logic, provided that the reductio ad absurdum reasoning frowned upon by Pappocom is considered 'logic'. (You've stated that you do consider it logic). An Ariadne's thread solver, which is driven by the reductio ad absurdum/guess procedure, is guaranteed to find all possible solutions to a Sudoku puzzle. How could the four solutions to DT Puzzle 3 have been found if not by logic? What alternative mechanism could have been employed?

I think the answer is that you define the terms 'logic' and 'guess' differently to how I do.

[shakers]

The DT seem to have set up a competing forum - this post particularly caught my eye http://www.sudoku.org.uk/discus/messages/1/9.html?1114453863 - Mr Mepham sounds terribly defensive, and as for starting a slanging match, it seems quite clear where the first 'slang' has originated from!

[Guest]

Hmmm. Seems to be getting a little heated!

[Frenchie]

Being a recent player, but already hooked, I find Mr Mepham particularly defensive towards his Su Doku and aggressive towards his potential customers... I cannot picture myself having to pay the Torygraph, only to face such aggressivity (even) in the Games section. A pity!

Ah the theme of the "non-uniqueness" of the solution... Since I work in numerical modelling it is a theme that is always most interesting...

[Guest]

Sorry, I find the Telegraph one much more challenging and enjoyable than the Times.

[ArchieB]

Thought I'd try the Telegraph's on-line puzzles for some extra practice so I printed out no. 110 of June 16 2005 and reached a dead end with 15 cells to fill. Using "bifurcation" I then found there were 2 possible valid solutions, something the DT apparently denies about its puzzles.

Wot a swiz!!!!

I'm sticking to the Grauniad from now on!

[RFB]

In true Grauniad style you have almost certainly made a transcription error in copying the clue numbers.

Only one (very early) Telegraph puzzle has had duplicate solutions due to a bug in the generator program.

If you can't spot your error post both of your solutions.

[Finlip]

The only thing I need in a su doku puzzle is that it must have a unique solution.

[frazer]

The DT puzzles really do have unique solutions, with the one early exception, as RFB says. So if you have two solutions, it's you who have made an error! It is sometimes necessary to bifurcate to solve them, but solutions are unique.

[tso]

Thought I'd try the Telegraph's on-line puzzles for some extra practice so I printed out no. 110 of June 16 2005 and reached a dead end with 15 cells to fill. Using "bifurcation" I then found there were 2 possible valid solutions, something the DT apparently denies about its puzzles.

Wot a swiz!!!!

I'm sticking to the Grauniad from now on!

A) You've made a mistake -- that puzzle has one and only one solution.
B) It can be easily solved without bifurcation, guessing or trial and error by forcing chains, coloring or related tactics.

It is not an exceptionally difficult puzzle. On the contrary, it is just slightly beyond the most common methods -- as you say, only 15 cells left to fill, not enough to confuse the issue.

One can *always* solve puzzles published in the Telegraph (so far) without bifurcation, regardless of what Melphan says to the contrary.

Here's the puzzle:

Code: Select all

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



You got stuck here:

Code: Select all

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



Here are just the candidates in the 15 cells left with the rest left blank for clarity:

Code: Select all

 .   .   . | .   .   . | .   .   .
 .   .   . | .   .   . |24  49  29
 .   .   . | .   .   . | .   .   .
 ----------+-----------+----------
 .   .   . | .   .   . | .   .   .
 .   .   . | .   .   . |45  14  15
 .   .   . | .   .   . | .   .   .
 ----------+-----------+----------
89   .   . |68   .   . |56   .  59
 .   .   . | .   .   . | .   .   .
89   .   . |68   .   . |26  19  129


There are many ways to solve from here. The what I found:

Code: Select all

1) If r7c9=5, then r5c9<>5, then r5c7=5.
2) If r7c9=9, then r2c9=2, then r2c7=4, then r5c7=5.
3) Therefore, r5c7=5, and the rest solves itself.


Code: Select all

 .   .   . | .   .   . | .   .   .
 .   .   . | .   .   . |24   .  29
 .   .   . | .   .   . | .   .   .
 ----------+-----------+----------
 .   .   . | .   .   . | .   .   .
 .   .   . | .   .   . |45   .  15
 .   .   . | .   .   . | .   .   .
 ----------+-----------+----------
 .   .   . | .   .   . | .   .  59
 .   .   . | .   .   . | .   .   .
 .   .   . | .   .   . | .   .   .


The DT's puzzles only get marginally more difficult than this.