4 ms·
My local newspaper publishes a puzzle called "The Challenger" (where you guess what numbers in a 4 x 4 grid will add up to the totals at the end of each row and
by MilnerRoute 4y ago
My local newspaper publishes a puzzle called "The Challenger" (where you guess what numbers in a 4 x 4 grid will add up to the totals at the end of each row and column). It frustrated me because there's no one right answer -- you were just supposed to start guessing numbers until you stumbled onto a sequence that worked.
So I wrote a nice elegant program that crunched through every possibility (using 12 nested loops), eventually spitting out solution after solution after solution. The first time it produced a valid solution, I was thrilled -- and from then on I felt like I'd already somehow solved all the puzzles in advance.
- lysergia 4y agoWould be cool if you win money for solving it!
- Anunayj 4y agoHa! I absolutely love solving stuff like this, automating it many times is more fun for me than actually solving a problem. But I'm always dissapointed/bored when I have automated it, suffering from success.
- JoshCole 4y agoRegarding that puzzle type... There is a rule that lets you transform between valid solutions, right? What happens when you use that rule to help impose additional constraints onto the problem such that you are solving for a particular solution rather than a solution? I feel like that should sometimes let you choose pathologically transformed variants. For example - lets say you were stuck with eight and nine as your valid answers in a bottom left corner, but the constraints don't forbid nine. We ought to be able to pick nine in this situation under the logic that there will be a transformation rule which makes nine a valid solution. Such a choice means we are now solving for a particular solution - not a solution - because eight might have worked just as well. In other words can you use the "do you know a related problem technique" to make your problem more specific?