The New Sudoku Players' Forum

[afjt]

All you need to do is change some numbers in the solution and re-calculate the cage-sums.

Surely verifying uniqueness of the solution is the hardest bit when creating a puzzle ? If we change some of the numbers, are we still guaranteed uniqueness ?

[PaulIQ164]

If all you did was to get a Killer puzzle, change some of the numbers (say, change all the 2s to 14s, all the 6s to 10s and all the 8s to 17s,) there's no guarantee you'd get a valid Face Ripper (as this type has been dubbed) puzzle, is there? Because Killers are constructed so that they can ve solved with the knowledge that the numbers used are 1-9, so the equivalent Face Ripper might not be solveable.

I think the puzzle was constructed by getting a finished (regular) sudoky grid, changing some of the digits used, and then putting cages in so that it was solveable.

[Animator]

True, you need to do a bit more then just replacing all numbers... (To guarantee uniqness that is. It should always be 'solvable'.)

You need to make sure that the list of numbers is the only possible combination.

(For example, if you have 18 in your grid then you need to make sure the person can select the right combination. 2 + 16 for example and not 4 + 14. Basically, doing what I did in my previous post and seeing that you end up with 1 combination.)

[Mel-o-rama]

Oh, I thought I would add: when I solved this puzzle, I first looked for single cell values (e.g. doing a '65' on C9). That provided 4 or 5 values. Then I did the N2/N5 thing to get the rest.

[AndrewTaylor]

Was it hard to create?

You know, I've no idea. I've never made any other sudoku puzzle by hand.

It was a bit of a pain, but I don't think I encountered any problems that were unique to the "Face Ripper". I don't know if I went about it the most sensible way, but it didn't seem worth coming up with a system to make one puzzle. The first one I came up with had four solutions, but that was fairly easily fixed (I didn't have to start over).