• Vehicles • Fashion • Recipes • Blogs • Hunt • Travels • Sport • Fun • Handmade • IT • Education
Mini-Games
x

x
zakruti.com » Knowledge, science, education » Numberphile
17 and Sudoku Clues - Numberphile

17 and Sudoku Clues - Numberphile

FBTwitterReddit

video description

Rating: 4.0; Vote: 1
17 and Sudoku Clues Marcos: I'm not sure about the conclusion that you need at least one clue placed where two pairs could be swapped to get valid puzzles. The fact is, per definition that solutions are unique, hence the chance of having two pairs that could be swapped is often a valuable indirect clue to discard candidates in some other cells, hence it is not true that you need one of the four digits shown as a clue, when it is enough to have some (perhaps chained) indirect clue about either any of the four numbers, or about the mere existence of two pair of numbers that could be swapped which invalidates an ongoing hypothesis. And then, the same goes for chains of pairs ending in a row, column (and perhaps a section) in which two candidates could be swapped at each end of the chain (let's say you have a pair A, B at the bottom column, then the same pair A, B diagonally in the first and the forth column, -chained- with the pair A, B in the last row at the first and fourth column intersection, such that A and B couldn't be repeated a third time within the first column and the last row, and pondering that these chains may have more than one pair of intermediate -nodes-, e. g the end and the final pairs may be on two columns at different rows or in two rows at different columns. So, I believe it is far more complex as to reach to such a premise that the clues should be explicit (you may even have two sets of this kind in which there's a shared candidate, and the -chained- effect of choosing one configuration in a set leads to a non univoque distribution in the second set, hence that's the clue telling you which configuration from the first set is invalid.
Date: 2022-04-08

Comments and reviews: 9


James, I wonder:
Is there a logical, algorithmic approach that always suffices to solve a sudoku puzzle? Or does a sudoku puzzle exist that is sufficiently random (in some sense lacking discernible patterns) that one must guess a solution and work the puzzle from there to see if the guess is valid?
I know a set of techniques for solving sudoku puzzles, but on some difficult puzzles, I end up guessing one or more squares and then have to work out the results to see if I guessed correctly or if I have to backtrack.
Are there more techniques I could learn, or is guessing sometimes inevitably required?
On reflection, this is related to writing a computer program to solve sudoku puzzles, and whether it could be completely algorithmic or would sometimes have to proceed by trial and error.

reply

Hi, I have an interesting question: this proof shows that you won't find a puzzle that is uniquely solved with fewer than 17 clues, alright. But there might be some of the 6. 7 thousand million million million distinct grids out there (or some of the 5 billion whatever grid families, for simplicity) requiring more than 17 clues to be uniquely solved. Now, how do we get to know which is the highest minimum number of clues that you will ever get from any of the possible grids/families? Does somebody know the answer, or any research on the subject? I find this is a valid and interesting problem too.
Thank You!

reply

For me the main problem I have is a computerized Sudoku puzzle always has hints that are if you place the wrong number it will tell you and that makes the game basically unplayable. Because there's no actually logic getting it out you just poke until it's done and I hate that. But a paper copy of a Sudoku I like
reply

I have two corollary questions:
(a) in a simpler 2X2 (instead of 3X3) sudoku, what is the minimum clues for a base-4 2X2 sudoku, ie, in a simple grid of 1, 2, 3, 4
(b) in a far more complex 4X4 sudoku, what is the mimum clues for a base 16, i. e. 1-9, a, b, c, d, e, f, g sudoku?

reply

The thing about enjoying Sudoku is to remember that its just logic and reasoning and elimination, no math in solving. The puzzles would work exactly the same with A-B-C-D-E-F-G-H-I, or any 9 other unique symbols. If you are repelled away because -ugh, math-, don't be: )
reply

Pedantic point: this doesn't prove that a smaller number than 16 clues could give rise to a unique sudoku grid - although it would seem on the face of it not to be possible.
But then, that's why we rely on mathematics rather than guy hunches.

reply

In the end of this video, James Grime gives his opinion about if he personally likes Sudoku game and he answers he dont because -it is a game of pure logic-. Question: Do you think the work of Gary McGuire to discover 17 was pure logic?
reply

Sudoku was invented by an american in late 70's actually so it does not come from Japan after all. It was first widely published in Japan magazines and that's where misunderstanding about it's origins may be coming from.
reply

If you are into Sudoku puzzles, then this is the books for you. It really gives you a challenge and it is so much fun. The more puzzles you solve the more you want to solve more puzzles. I love books for dr. khalid alzamili
reply
Add a review, comment






Other channel videos