In less than a decade, Sudoku has gone from obscure logic puzzle to global phenomenon. Scientists have built up an entire sub-field of mathematics around this game, mostly to answer one question: what's the toughest possible puzzle that's actually solvable?

Admittedly, that isn't quite how mathematicians usually describe the question, but the end result is the same. Think of it like this - newspapers tend to provide 25 numbers in the 9x9 Sudoku grid as hints. By adding more numbers, the puzzle becomes easier to solve, while taking them away makes the task harder. The question then is what's the minimum amount of numbers a puzzle can have and still be a real, solvable puzzle.

For Sudoku, solvable means the puzzle only has one unique solution. So far, puzzle constructors have reported that they have been unable to create a solvable puzzle with fewer than 17 clues - with only 16 numbers provided, the puzzles always yield multiple solutions. That obviously suggests the lower limit is 17, but that can't simply be assumed in mathematics. A formal proof would be required, and the only way to do that would be to create a computer algorithm that could somehow check through all the countless possible combinations of Sudoku grids with 16 numbers given.