The New Sudoku Players' Forum

[smurfdaddy]

Since I am new to the forum, these questions might be stupid but how else to learn (sorry if the lingo is not always correct and I do appreciate tips abput that). 1. When a puzzle is made does the puzzle creator know in advance how hard it will be and what techniques will be required to solve. 2. Does the number of clues (gimmies) affect the complexity. 3. Is the puzzle designer designing the puzzle to lead me through it a certain way, assuming I will fill in all the clear cut answers from 1 to 9 first then move onto the squares, rows, columns with the most clues. 4. If I think "outside the box" so to speak and attack the puzzle a different way than described above, would I be getting quicker results (ie am I attacking it in an efficient manner). Sorry to sound anal but enquiring minds want to know. Thanks to all who assisted me with me first posting and told me about chains and coluring and protocol.

[motris]

1. It depends on how the puzzle is constructed - I would argue that most puzzles currently computer-generated (the majority of the sudoku you will find) are merely screened to ensure they are solvable with particular sets of techniques to set the difficulty rating. Truly talented writers of hand-made sudoku seem, at least to me, though to be able to design in a reasonable solving route and set of techniques necessary to the solution. Pappocom has often claimed that it took him so long to develop his software because he wanted his computer-generated puzzles to feel similar to the hand-made ones from Japan and while they are close, I would still say I can tell most good hand-made puzzles apart from computer-generated ones.

2. There are many many examples to show that 17-18 clue sudoku can be easy and 30+ clue sudoku can be very hard. The number of givens does not at all indicate the actual difficulty. That being said, on average over all published puzzles, the more givens you get, the easier the puzzle is. Certainly, the "very easy" sudoku that are out there start with over 30 givens to aid the solver.

3. (See 1)

4. (See 1) - generally, any accessible route you have to solving the puzzle is a good one. If you care about just solving the puzzle using logical moves (and no T&E), then do whatever techniques you enjoy most - only a rare hand-made puzzle will tend to "force" an obvious route. If you care about solving the puzzle fast, you will need to decide from experience how best you can accomplish this. Myself, I tend to attack a puzzle with a few slice-and-dice type passes first and then look for points of weakness to tease out hidden pairs/triplets/etc. Of course, if any row/column/box has only a few cells left to fill, I may forgoe the slice-and-dice and just finish those groups as soon as possible.

Thomas Snyder

[smurfdaddy]

interesting, are there hand made puzzles accessible on the internet that one could try to see the difference.

[tso]

The Guardian's are hand made.

Nikoli's are all hand made, but few are on the web.

Handmade puzzles generally top out at a difficulty level roughly equivalent to Pappocom's Hard -- many newspaper puzzles get no harder as well, regardless of the labels.

You'll have to decide for yourself if there's a difference. I've been solving them since the mid-80s, so though I think there *is* a difference I cannot possibly be objective.

[fermat]

interesting, are there hand made puzzles accessible on the internet that one could try to see the difference.

Puzzles at http://www.sudoku.frihost.net/forum/viewforum.php?f=12 are hand made with computer assistance. That is not quite what you asked. Deano, at that site, is/was a "hands on" creator but found this software to help, really help.

Edited to add: I would say that the difference is that hand made puzzles usually are solvable for a while, then hit a harder requirement. Computer puzzles may have the first step the hardest. (Tons of markup).