I tried and solved my first (easy) Sudoku puzzle recently. It set me to wondering however: What is fewest number of initial entries into a 9 x 9 sudoku grid that permit such a puzzle to be solved? I suspect the number can be formally arrived at, but I don't know how. Any takers?

On the sudoku site the puzzles there show either 3 or 4 numbers
in each sub grid. So your only possiblity is 1 or 2 as smaller, not
much of a range. 1 you could eliminate as allowing many solutions
so the idea is to figure out if 2 in each grid could be set up
to have only one solution, I think not. I think with 2 in each grid you
have many more solutions so it looks to me like 3 and 4 are the
minimum already.

I had considered that, as well as 69 but the philosophical implications
stymied me from venturing a valid logically unimpeachable arguement
considered as a devils advocacy type of involvement with the
fundamental numerical formulations.

OK I doubt this will work. But there is a 17 number setup with a unique solution. No 16 numbers setups with unique solutions have been discovered but they haven't been shown not to exist.

well that blows my 3 and 4 thing. The puzzle I saw on the sudoku
site had 6 3's and 3 4's (30) It just seemed to allow more
solutions if you had less occupied squares.
So the key is the *position* in the squares.

Originally posted by sonhouse well that blows my 3 and 4 thing. The puzzle I saw on the sudoku
site had 6 3's and 3 4's (30) It just seemed to allow more
solutions if you had less occupied squares.
So the key is the *position* in the squares.

They've just started doing these in my local newspaper. The ones I've seen have 3 or 4 entries in SOME subsquares, but definitely not every subsquare.