The New Sudoku Players' Forum

[jonallen1966]

The article in the times about Tetsuya Nishio,
"the undisputed grand “puzzle master” of Su Doku"

http://www.timesonline.co.uk/article/0,,7-1757275_1,00.html

was very interesting. He reckons he can make them so complicated
even a champion would take 6 hours to do it!
I've never had to time to finish one yet.

[Jules]

some days tricky are easy and others not so same with the normal sudokus - regardless of category

usually find the so called easier ones occasionally take longer due to not concentrating as much and making a pigs ear of them, while a difficult/fiendish automatically focuses the brain cells more

would be nice to have a harder than tricky killer to tackle

[Karyobin]

He reckons he can make them so complicated even a champion would take 6 hours to do it!

Wasn't this quote with regard to 'Killer' sudoku?

[zundevil]

To those of you who have received your copies of the Killer Sudoku book (advertised on the Times' page) -- are these all new problems? I was afraid of buying it only to see that it was a collection of the killers that had been published in the Times already!

Also, if you'd like to see the most difficult killer I've ever tried (much much harder than any I've seen from the Times -- YMMV), try the September 27th puzzle from the following:

http://www.djape.net/sudoku/wp/gallery.php?file=samunamupure/puzzles/

I can get the middle-middle (Ann B. Davis?) nonet, but hardly anything in any of the rest of them -- the singleton fences/cages aside. I don't write possible numbers in the squares, so maybe I'm missing something really easy...

ZD

[CathyW]

I've got the book but haven't yet had time to do any of the puzzles - there's enough with The Times and the extra Killer from djape.net site so can't comment on whether there's any replication. Would be disappointed if there is.

Agree with you about the 27 September killer from djape. I've managed all the others so far but gave up on this one and cheated by looking at the solution. Was none the wiser as to how it had been solved.

[Bigtone53]

We have probably strolled too far from The Times, but in relation to the 27/9 Djape, I am told that if having filled in the middle (and I am not clear how you can be so sure about the middle vertical 3 in this nonet), you consider all the possibilities for the 4 squares r4567 c8, you end up with a firm conclusion about a square in the bottom right nonet. It does not seem to help though and there are only so many hours in the day.

[Pi]

Does anyone know if killer puzzles are handmade or computer mad?

The difficulty levels seam more random than the standard puzzles which usualy take around the same amount of time for a puzzle grade

[PaulIQ164]

The old ones (I thought) were from Nikoli so were handmade. These new ones we seem to get now I don't know about - yesterday's was credited to AIIA corporation or something similarly vague, but today's has a human credit again so I have no idea.

[domt]

can anyone explain how to solve the above mentioned killer 27/9 methodically. i have looked at it and consider myself not bad at killers but this one has me baffled. is there a program out there for killers?

dom

[CathyW]

I've posted a message at www.djape.net on the Killer page to see if he can provide an explanation of the logical moves rather than just the solution.

[PaulIQ164]

(and I am not clear how you can be so sure about the middle vertical 3 in this nonet)

I got those 3 cells, but it was trial&error really (though I doubt the boundaries are as clear in Killers as the already fuzzy sudoku ones). The 5 in row 4 has to be in the 18-cage, and there's only two ways to make 18 in 5 cells using a 5 (5+1+2+4+6; 5+1+2+3+7). The latter leads to an impossibility when you try to fill box5, so it must be the first alternative, which lets you fill the box correctly.

[djape]

I see that there's been a lot of questions regarding the puzzle posted on Sep 27 on my site. I posted an explanation there, here is a summary:

This was only the second Killer puzzle generated by my software. The first version of the algorithm relied heavily on backtracking so it is possible (altho I think very unlikely) that the puzzle can be solved only by trial and error.

I haven't tried solving it by hand but there are advanced mathematical methods for solving such puzzles. When I get some time I'll try solving it and will let you know.

This, I suppose, answers another question - my Killers are computer generated. Now I can generate them with different options (choose which solving techniques can be used) but the first version was mostly trial/error.

Best regards to all,

DJApe
www.djape.net

[PaulIQ164]

Well, that answers that. Thanks very much.

[CathyW]

We have probably strolled too far from The Times, but in relation to the 27/9 Djape, I am told that if having filled in the middle ... you consider all the possibilities for the 4 squares r4567 c8, you end up with a firm conclusion about a square in the bottom right nonet. It does not seem to help though and there are only so many hours in the day.

It seems to me that the candidates for r7c8 can be 2,3,4,5,6,7,8 or 9!! Doesn't tell me anything about the bottom right corner box.

I got as far as filling in box 5 with other candidates for rows 4,5,6 as follows:

{123}{123}{23}{8}{6}{4}{9}{57}{57}
{468}{468}{468}{9}{7}{5}{23}{123}{123}
{57}{57}{9}{2}{3}{1}{468}{468}{468}

Apart from the 1s at r2c3 and r8c7, the only other clues I have is that the 21 cage (r8c6,r9c6,r9c7) must have a 7 in it.

Hopefully DJAPe will advise soon whether this is actually solvable by hand without trial and error. As mentioned previously, I did look at the solution but it didn't help to work out how the answer had been arrived at.
[/code]

[afjt]

Agree with you about the 27 September killer from djape. I've managed all the others so far but gave up on this one and cheated by looking at the solution. Was none the wiser as to how it had been solved.

Did you not find that 28th September was not unique ? The 2's and 3's in r2c78 and r9c78 are interchangeable.