Why Sudoku’s clue count has a hard floor
A Sudoku clue does more than fill a square: it rules out possibilities elsewhere. Each digit must appear once in every row, column, and 3-by-3 box, so a starting number can help constrain several overlapping groups at once. A puzzle has a unique solution only when those constraints leave exactly one completed grid—not merely when a solver can find a grid that fits.
Why 16 clues cannot be enough
With too few starting digits, the rules leave room for more than one completion. The surprising part is that this is true even when the clues are placed in the most favorable arrangement: exhaustive computer analysis showed that no standard 9-by-9 Sudoku with 16 or fewer clues has exactly one solution. Some such grids might have a completion, but they cannot have just one.
Seventeen clues, by contrast, can suffice. The locations and values matter: carefully chosen clues can work together to eliminate alternatives across the grid, while a larger set of poorly placed clues may still leave a puzzle ambiguous. The number of clues alone does not tell you how hard a puzzle will be; it tells you only how much information the puzzle starts with.
Few clues do not automatically mean hard
A 17-clue puzzle is an extreme case of sparse starting information, not a guarantee of a long or difficult solve. Difficulty depends on how the remaining possibilities interact and what deductions a solver must make. Still, the 17-clue boundary gives Sudoku a remarkable mathematical limit: remove even one clue from any uniquely solvable standard puzzle at that minimum, and at least one competing completion must become possible.
