Talk:Killer sudoku

The text 'below' the sample solution was appearing to the right of it - I didn't know how to fix that properly so there's an unsightly line of s in there instead. --Last Malthusian 00:32, 6 January 2006 (UTC)

Though there is a link at the bottom to my page of tips on solving these puzzles, I was displeased to see that an entire paragraph of that page was pasted, without a credit, into Wikipedia, with a few words changed to disguise its origins. I have now credited it to myself here--sorry to be the egotist, but if someone wants to write a NEW, ORIGINAL paragraph they are welcome to get rid of my self-quotation here. --ND 05:03, 13 January 2006 (UTC)

Per Ndorward's request, I wrote a new paragraph, and erased the direct quote. Nuj 06:47, 13 January 2006 (UTC)

THanks! A lot more space-efficient to boot.... Anyone want to take a crack at improving the rest of the page....? --ND 06:51, 13 January 2006 (UTC)

The other problem here is that the discussion of solving strategies is kind of useless: in particular, the omission of the "45-sum rule" leaves the solver without one of the basic tools needed to solve puzzles efficiently, & the lengthy chart of combinations is misleading: that is by far the worst, most inefficient & mind-numbing way of solving puzzles. The choice of puzzle for display/example here is unfortunate because it doesn't really lend itself to demonstrating the 45 rule, let alone others (udosuk's "subtraction combo" [a key technique for bypassing combination-crunching], J.-C. Godart's "overlap"). If someone wants to fix this page go ahead: I'm not in the mood to do it myself at the moment, given that I don't want to just paste in material I've already discussed at length on my own site. --ND 05:27, 13 January 2006 (UTC)
 * I've had a go at explaining the 45-rule. The trouble is that our sample puzzle doesn't given an immediate example of the traditional application, though it allows a more advanced application which I've included in the initial analysis. Plus I'm sure someone can explain it more clearly than I can. If I have time tonight I'll try creating a two-nonet example of the simple application. What would be best is if we had an easier sample puzzle to which we could apply all the simple techniques. --Malthusian (talk) 10:00, 13 January 2006 (UTC)

"Sumdoku" is something different?
I recently saw the term "Sumdoku" (or it might not have had the "d", and was just "sumoku"?) used in a book for the first time in my memory--but they used it for a slightly different type of puzzle. All the killer sudoku rules applied (including no repetition of any number in any cage), except for one--there were no 3x3 cells in the puzzle, and numbers could therefore appear more than once in those regions. (But the rule of no number appearing more than once in any row or column still applied.)

I didn't buy the book, and I don't remember who published it. But if anyone else has seen it or can reference it, it may be worth noting... —Preceding unsigned comment added by 67.233.237.161 (talk) 07:51, 23 April 2009 (UTC)

Suggestion
Here's a quick idea, I'm curious as to what others think of it. I play regular sudoku, not the killer version. However, it seems to me there is one important (or interesting) distinction between the two, that is not mentioned in the current article: regular sudoku is a purely logical puzzle, whereas killer sudoku is not. (I do not intend this distinction to deingrate either puzzle.) In regular sudoku, the numbers are merely symbols -- their numerical properties do not count (oops, pun!). It would be possible to play regular sudoku using any set of nine unique symbols, without changing the rules or solution strategies. However, this is not true of killer sudoku, which combines logic and math. AlistairLW (talk) 16:15, 2 January 2009 (UTC)


 * Actually, I disagree to some extent. Killer is a purely logical puzzle because you can solve it purely by logical deduction. Also, I think (but I haven't tested this) that you can invert the numbers, by which I mean you can replace a "17 in 2" with a "3 in 2", a "23

in 3" with a "7 in 3" etc. along with replacing 9 with 1, 8 with 2, 7 with 3 and so on. Thus it is not completely dependent on one set of numbers. And perhaps there are other transformations possible (maybe someone who's handy with group theory might be able to work this out). But I do see your point - you could replace traditional sudoku numbers with apples, oranges etc. - I've even seen this in some children's puzzle book.--A bit iffy (talk) 19:45, 2 January 2009 (UTC)

Sum tables
I deleted these for the moment. They take up too much space, they're not that useful, and they're ugly. Er... in my opinion :-). --Malthusian (talk) 10:00, 13 January 2006 (UTC)

Yes, they should be deleted I think: they are misleading because they suggest that that is a good solving strategy to run mechanically through tables, whereas that is a strategy of last resort. -- I think a new, more useful puzzle should be substituted on the page. Maybe someone with a puzzle-generating program can post one up that has some good examples of the 45 rule? (I suggest a computer-generated one because then we won't get into copyright problems.) --ND 15:01, 13 January 2006 (UTC)

I have created a complete sum table here (also used on the Cross_Sums page). Maybe the sum tables could be replaced by this link? --Yaaaay 12:20, 9 March 2006 (UTC)

External links to sites offering puzzles
Following the example of the Sudoku article, I have removed the external links to sites promting their own Killer puzzles. This is because it's becoming clear there are more and more of them turning up, and they are easy to find anyway. --A bit iffy 10:12, 28 March 2006 (UTC)

"No number can appear more than once in a cage."
I've just looked at three puzzle books that include instructions for Killer Sudoku. Not one of them mentions this rule. Does this mean that the puzzle setters believe the rule doesn't exist, or that they merely forget to mention it?

So far, I'm yet to see a Killer Sudoku that violates this rule. Has anyone? Conversely, has anyone seen one that actually requires this rule for the solution to be unique? -- Smjg 18:21, 22 July 2006 (UTC)


 * Killer #74 was the first puzzle in The Times that actually required the rule. Discussion here: http://www.sudoku.com/boards/viewtopic.php?t=2498&highlight=&sid=d01aa1a7d37b155a4d2919782a459cf4 86.172.43.7 (talk) 16:21, 10 July 2009 (UTC)

sudoku.org.uk that does the TIMES puzzles calls this rule "the killer convention." if You go to the site "Daily killer" there is an explaination that shows a puzzle that cant besolved without it. also in the forums someone spotted a killer in the answers to archived puzzles that didnt follow it..

oh the guy just below... nevermind

well the guy two down yes that is the times and Michael Mepham... mind

Petetyj (talk) 04:33, 12 February 2008 (UTC)

Look at the section on the "duplicate cell controversy". There is exactly one puzzle that violates this rule in the Times Killer Sudoku book; every other puzzle I've encountered has held to it. ND 03:17, 23 July 2006 (UTC)

The book Martin Bauer, Michael Mepham: Killer Sudoku, Die neue Sudoku-Dimension does not use this rule. It is not mentioned in the rules and some puzzles violate it. (For example, puzzles 1, 4, 7 and 8.) --Tarvaina 00:20, 2 December 2006 (UTC)

This rule is never mentioned for the puzzles published daily in the UK in the Daily Telegraph, either. Which is, frankly, very annoying. Whilst I've never yet seen an example that violates it, I've never felt able to assume it when attempting a solution, either. Fredd169 (talk) 07:43, 8 April 2015 (UTC)

Combinations
I see that someone's added in the cage combos for 6, 7, and 8 digits. Are these really necessary? As is already stated in the article, the combinations are more easily derived by using the complement of the 1, 2, and 3-cell cage combos. (In particular the 8-digit list seems silly to me: all you need to do to figure out the missing digit is subtract the cage-sum from 45.) ND 01:53, 31 July 2006 (UTC)

6 cell cage 21: 123456 min 22: 123457   23: 123458 123467     24: 123459 123468 123567     25: 123469 123478 123568 124567     26: 123479 123569 123578 124568 134567     27: 123489 123579 123678 124569 124578 134568 234567     28: 123589 123679 124579 124678 134569 134578 234568     29: 123689 124589 124679 125678 134579 134678 234569 234578     30: 123789 124689 125679 134589 134679 135678 234579 234678     31: 124789 125689 134689 135679 145678 234589 234679 235678     32: 125789 134789 135689 145679 234689 235679 245678     33: 126789 135789 145689 234789 235689 245679 345678     34: 136789 145789 235789 245689 345679     35: 146789 236789 245789 345689     36: 156789 246789 345789     37: 256789 346789     38: 356789    39: 456789 max ugly but interesting... i thought the colons were better than the = sign http://groups.yahoo.com/group/Sudokuworld has a copy of my available free sudoku program. which has a help file and a table with a strangely famillar chart.. i thought the min max would be helpful apparently the wiki "Contributor" didnt... or perhaps im just paranoid.. this chart is the output of my also freely available cage sum calculator although with the program in the delphi ide i can make all 1..9 digit cages in one swell foop... you have my permission as the author to use them here... the program allows you to eliminate digits that cant be in a cage or only includes combinations that must have a digit present so it's probably more helpful than the chart and a lot smaller. you have my permission to use it and also if you want some prettier non copyrighted example Pictures of killer sudoku.. my program streches to any size and works from two color to many millions. it also does killerjigsaws.. forum "my sudoku" thread "Killer Jigsaw Septet" http://www.sudoku.org.uk/SudokuThread.asp?fid=1&sid=3948&p1=2&p2=4 gives an example of the images available .. i also drop them in other places espically on that site Petetyj (talk) 04:20, 12 February 2008 (UTC)