The New Sudoku Players' Forum

[PaulIQ164]

For what it's worth, I did Monday's Killer again, and I'm pretty certain that none of the filled-in digits were necessary to solve the puzzle. The 1s and 2s looked like they were in that annoying combination where you can swap each pair round and still have a valid solution, but they were all in separate 3v3 boxes so they weren't really.

[Christofer]

From today's Times (12/9/05):

'Within each dotted-line "shape", a digit CAN be repeated if the normal row, column and 3x3 box rules are not broken.'

What they mean is that a shape can contain duplicate digits if the shape crosses over more than one line, row or 3x3 box.

Since English isn't my native language I have difficulties explaining this. I'll try with a little picture instead...

XXXXXXXXX
XXXXXXXXX
XXAAXXXXX
XXXAXXXXX
XXXXXXXXX
XXXXXXXXX
XXXXXXXXX
XXXXXXXXX
XXXXXXXXX

A=6 (for instance)

Then a solution can be:

XXXXXXXXX
XXXXXXXXX
XX14XXXXX
XXX1XXXXX
XXXXXXXXX
XXXXXXXXX
XXXXXXXXX
XXXXXXXXX
XXXXXXXXX

I.e. a shape with duplicate digits...

[Enigma]

I agree... you didn't need any of the filled in numbers to solve the puzzle. You can place the 1 and 3 in the top left box immediately, and the placement of the 1s and 2s down the bottom right can come from needing 3 of 1,2,3 and 5 to make up the enclosure totalling 10 in box 8. The left over 1 goes in r7c6.

I wonder what comments the Times would get if they were to publish a crossword with some of the answers already filled in ?

Paul

[CathyW]

And today's puzzle was too easy as well. Definitely not tricky, though it might have been if they hadn't given some of the numbers

I thought the whole point of "Killer Sudoku" was that no numbers were given and we had to work them out from the sums as well as the normal sudoku rules. Perhaps I'll have to get the book for some properly tricky puzzles. Has anyone written to The Times about the extra clues?

[roger888]

I thought the whole point of "Killer Sudoku" was that no numbers were given and we had to work them out from the sums as well as the normal sudoku rules.

On this specific point, when the Times introduced them, they quoted Nishio thus:

"The really satisfying part ... is that you can design a Samunamupure puzzle in a way that you do not need to insert any starting numbers in the grid.”

This implies to me that both variants fit the general definition of a well-formed samunamupure.

[tso]

I sent the following email to the TIMES:

=====================================

Three questions regarding ongoing errors with your Sudoku column:


1) In your cover story "Too good for Fiendish? Then try Killer Su Doku" http://www.timesonline.co.uk/article/0,,7-1757275_3,00.html August 31st, the rules for solving Samunamupure (Killer SuDoku) are laid out, followed by these hints:

"... in the case of three joined squares, if the printed number is 6, the only combination possible is 1, 2 and 3; if the number is 24, 7, 8 and 9..."

Having solved Samunamupure for years in Japanese publications, this seems to be correct; no duplication of digits is allowed within the numbered areas surrounded by dotted lines. (The puzzle is a combination of two of the most popular puzzles in Japan -- Sudoku aka Number Place and Cross Sums, as the name suggests. The "no duplicates" rule is intrinsic to Cross Sums.) If the solver ignores this rule, s/he may find there are multiple solutions. However, you do not state this rule explicitly as you should have, instead forcing us to infer it. Then, you've started printing the following contradictory disclaimer underneath the Killers:

"Within each dotted-line “shape”, a digit CAN be repeated if the normal row, column and 3x3 box rules are not broken."
http://www.timesonline.co.uk/article/0,,18209-1776891,00.html

How are we to reconcile these two statements? If "a digit CAN be repeated" within a shape, then if the printed number is 6, there are TWO combinations possible, not one: 1-2-3 and 1-4-1. Same goes for 24: 7-8-9 or 9-6-9. For example:

Code: Select all

. . . | . . . | . . .
. . . | . . . | . . .
. . 1 | 4 . . | . . .
------+-------+------
. . . | 1 . . | . . .
. . . | . . . | . . .
. . . | . . 9 | . . .
------+-------+------
. . . | . . 6 | 9 . .
. . . | . . . | . . .
. . . | . . . | . . .
 

This has caused great confusion and argument among solvers. PLEASE print a correction, one way or the other. It would be nice if someone posted an explanation to sudoku.com as well, or at least inform Wayne Gould at Pappocom. As it stands, your statements about the rules to this logic puzzle are illogical.

2) Also in the same article you say:

"But as you can see, in Killer Su Doku there are no traditional clue numbers — only cells linked by dotted lines with a number printed in the top left-hand corner."

However, both Killer #9 and #10 DO have clue numbers with cells. Again, I've solved Samunamupure for years and have never seen this. It seems as if you are "dumbing down" the puzzle for the English audience, who, to quote the supplier of your standard Sudoku, supposedly "like their puzzles hard". I suppose you had complaints from people who couldn't solve the puzzles -- must every puzzle be so easy that 100% of the public can bash it out on their bus ride to work?

3) For 10 puzzles in a row, you've printed a copyright notice "Puzzles by Pappocom" beneath the Killer, though Pappocom does not supply that puzzle.


Thank you for your attention to this matter.

[Karyobin]

Top stuff tso. Let's see 'em wriggle out of that one.

Did you do the spinning/bouncing thing?




Just wondered.

[tso]

Top stuff tso. Let's see 'em wriggle out of that one.

Did you do the spinning/bouncing thing?




Just wondered.

Spinning/bouncing thing? If you are refering to placing a rubber ball between my thumb and middlefinger and snapping the ball forward with a great deal of reverse spin (aka "english" in the american vernacular) so that the ball bounces back to me, surprising my dog, -- uh, no, that didn't seem appropriate.

[Karyobin]

Fair enough.

[jfm]

The latest variant of the rules on the timesonline website is even odder:

Within each dotted-line "shape", a digit CANNOT be repeated if the normal
row, column and 3x3 box rules are not broken.

Ermm, hang on a sec...

Still, all we have to do now is get rid of the patronising clue numbers.

[roger888]

The latest variant of the rules on the timesonline website is even odder:

Within each dotted-line "shape", a digit CANNOT be repeated if the normal
row, column and 3x3 box rules are not broken.

This is truly bonkers

Still, all we have to do now is get rid of the patronising clue numbers.

Out of interest, I re-solved today's without any of the clue numbers, and it was perfectly possible. Does anyone know whether this is always true? Have the clue numbers ever been needed for a unique solution?

Incidentally, there's also a letter in today's paper:

"Sir, Has The Times gone soft? My delight at the introduction of the Killer Su Doku (letters, September 9, etc) has been tinged with disappointment this week when you not only started to fill in the grid but are also publishing a key, albeit upside down. This changes everything. With my wife competing to get to the Fiendish version before me, my refuge in the Killer now looks shaky."

[katymeredith]

I've been (metaphorically) up a Tibetan mountain for a bit, so today's Killer was the first for me with the "well, there's unique and there's unique" dotted shape rule.

I have to admit it felt really odd at first, but as I progressed the additional challenge did start to appeal to me, especially the 12 near the top - having to bear in mind the possiblity that it could be a double-six was quite intriguing.

I just hope they don't start sticking in starfish shapes with 5 nines in them just because they can now.

And what's the point of the key? I mean, really? For how many people would the key make the difference between being stuck and actually being able to complete the puzzle? How many lost sales have they prevented by printed that?

And don't get me started on the clue numbers

If they want to give away clues, at least have the good grace to do it by way of single cell shapes.

[zebedeezbd]

especially the 12 near the top - having to bear in mind the possiblity that it could be a double-six was quite intriguing

Am I missing something? I think everyone agrees that it still has to work as a su doku, doesn't it? So a 2-box summing to 12 could never be a double-6. (But a 3-box summing to 6 could, under the right conditions, be a 1-4-1, say.)

How many lost sales have they prevented by printed that?

Lost sales or lost soles?

[katymeredith]

especially the 12 near the top - having to bear in mind the possiblity that it could be a double-six was quite intriguing

Am I missing something? I think everyone agrees that it still has to work as a su doku, doesn't it? So a 2-box summing to 12 could never be a double-6. (But a 3-box summing to 6 could, under the right conditions, be a 1-4-1, say.)

How many lost sales have they prevented by printed that?

Lost sales or lost soles?

Oops! I had a hard day in the office today ....... !!!

[sudoker]

Oops! I had a hard day in the office today ....... !!!

LOL