Write up?

Everything about Sudoku that doesn't fit in one of the other sections

Write up?

Postby mollwollfumble » Wed Apr 09, 2014 11:00 pm

I'm new to this forum.

I've been tempted on many occasions to write up the techniques I use to solve Sudokus.

I use mental techniques: low-hanging fruit, arrows, and left-middle-right.
I use written techniques: casual dots, row end numbers, casual numbers, complete dots, and row guess.
Some techniques I no longer use except on rare occasions: X-wing, 5 spaces, complete numbers.

"Row guess" with "complete dots" sufficed to get me right through a Black Belt Sudoku book without difficulty.

Should I write up my techniques?

Are there any other write-ups of Sudoku solving techniques that distinguish between "casual dots" and "complete dots"?
mollwollfumble
 
Posts: 8
Joined: 09 April 2014

Re: Write up?

Postby JasonLion » Wed Apr 09, 2014 11:28 pm

You will get the most interest, and personal benefit, if you read up on the terminology we use around here before doing such a writeup. The odds are high that you are simply calling something we are already familiar with by another name.

There is a lot of value to learning and using common terminology. If we don't speak the same language we can't readily share and build on each others results.
User avatar
JasonLion
2017 Supporter
 
Posts: 642
Joined: 25 October 2007
Location: Silver Spring, MD, USA

Re: Write up?

Postby David P Bird » Thu Apr 10, 2014 12:01 am

Jason, that’s a well written reply giving sound advice.

Mollwollfumble, here’s a couple of links to get you started:
http://www.sudocue.net/guide.php. – describes methods in ascending order of complexity
http://sudopedia.enjoysudoku.com/. – a mirror Jason took of an old Wiki style site which is now no longer available
David P Bird
2010 Supporter
 
Posts: 1043
Joined: 16 September 2008
Location: Middle England

Re: Write up?

Postby mollwollfumble » Thu Apr 10, 2014 4:33 am

David P Bird wrote:Jason, that’s a well written reply giving sound advice.

Mollwollfumble, here’s a couple of links to get you started:
http://www.sudocue.net/guide.php. – describes methods in ascending order of complexity
http://sudopedia.enjoysudoku.com/. – a mirror Jason took of an old Wiki style site which is now no longer available


I've had a quick look through those. They look like a great introduction to some of the more advanced techniques that I haven't seen on the web before.

But I don't see anything equivalent (even under different names) on either site to any of the written methods of:
"casual dots, row end numbers, casual numbers, complete dots, and row guess" that I use constantly. Nor do they mention anywhere the memory help technique I've called "left-middle-right". So perhaps I'd better write this up.

http://www.sudocue.net/guide.php uses what I've called "complete numbers", what I've called "arrows", and an interesting colouring system that I haven't seen before.
mollwollfumble
 
Posts: 8
Joined: 09 April 2014

Re: Write up?

Postby mollwollfumble » Mon Apr 21, 2014 10:37 am

mollwollfumble wrote:I'm new to this forum.

I've been tempted on many occasions to write up the techniques I use to solve Sudokus.

I use mental techniques: low-hanging fruit, arrows, and left-middle-right.
I use written techniques: casual dots, row end numbers, casual numbers, complete dots, and row guess.
Some techniques I no longer use except on rare occasions: X-wing, 5 spaces, complete numbers.

"Row guess" with "complete dots" sufficed to get me right through a Black Belt Sudoku book without difficulty.

Should I write up my techniques?

Are there any other write-ups of Sudoku solving techniques that distinguish between "casual dots" and "complete dots"?


Now written up and on the web at http://freepages.misc.rootsweb.ancestry.com/~hallsofjamaica/sudoku%20solving%20dots.pdf

Where possible I've used the standard Sudoku terminology from the pages linked by David P Bird, but those pages seem to miss three fairly-well-known solving aids that have appeared elsewhere on the web and in Sudoku book introductions.
mollwollfumble
 
Posts: 8
Joined: 09 April 2014

Postby Pat » Tue Apr 22, 2014 7:54 am

mollwollfumble wrote:

    If there are more than 28 [given] digits in the puzzle
    then it’s likely to be very easy

really?
here's a little collection of 742 puzzles having 39 cells given,
tell me if any of them is "very easy"

here's an example — not a puzzle i would ever try —
Code: Select all

 . . . | . . . | . . 1
 . . 2 | . 1 3 | . 4 5
 1 4 . | . 2 5 | 3 6 .
-------+-------+------
 . 2 4 | . 7 8 | . 5 .
 5 7 . | 3 . 2 | . 8 4
 8 . . | . 5 . | . . .
-------+-------+------
 . 8 . | . . 7 | . 1 .
 2 . 7 | . 8 1 | 4 3 .
 4 1 . | 2 3 . | . . 8

User avatar
Pat
 
Posts: 4056
Joined: 18 July 2005

Re:

Postby 999_Springs » Tue Apr 22, 2014 2:18 pm

Pat wrote:
mollwollfumble wrote:

    If there are more than 28 [given] digits in the puzzle
    then it’s likely to be very easy

really?
here's a little collection of 742 puzzles having 39 cells given,
tell me if any of them is "very easy"

how about this one

Code: Select all
795|6..|8..
81.|.95|.76
.46|.87|.95
---+---+---
67.|8.9|.5.
92.|56.|78.
.58|.7.|6.9
---+---+---
56.|7..|9.8
.37|9.8|56.
.89|.56|..7


49 clues. SE rates it 10.4. thanks to dobrichev for the puzzle
999_Springs
 
Posts: 591
Joined: 27 January 2007
Location: In the toilet, flushing down springs, one by one.

Postby Pat » Fri Apr 25, 2014 6:32 am

mollwollfumble wrote:

    If there are more than 28 [given] digits in the puzzle
    then it’s likely to be very easy

right. "likely" means 70% ?
and "very easy" -- i shan't try guessing
User avatar
Pat
 
Posts: 4056
Joined: 18 July 2005

Re: Write up?

Postby enxio27 » Sat Apr 26, 2014 5:23 pm

You are confusing solving AIDS with solving TECHNIQUES. Pencil marks, dots, etc. are solving AIDS, but they are NOT techniques, or strategies, to solve a puzzle. Quite frankly, your writeup doesn't make a whole lot of sense to anyone who is reasonably experienced with solving sudoku. I think you need to familiarize yourself with known techniques and fully comprehend them before you are in any position to add your own.
User avatar
enxio27
 
Posts: 532
Joined: 13 November 2007

Re: Write up?

Postby mollwollfumble » Thu Sep 11, 2014 10:38 pm

After posting http://freepages.misc.rootsweb.ancestry ... 20dots.pdf I was too chicken to look at the forum, too worried about what sudoku experts might say. Have only just got back to it now.

enxio27 wrote:You are confusing solving AIDS with solving TECHNIQUES. Pencil marks, dots, etc. are solving AIDS, but they are NOT techniques, or strategies, to solve a puzzle. Quite frankly, your writeup doesn't make a whole lot of sense to anyone who is reasonably experienced with solving sudoku. I think you need to familiarize yourself with known techniques and fully comprehend them before you are in any position to add your own.


OK, "aids" rather than "techniques" or "strategies" if it makes you happy.

999_Springs wrote:how about this one

Code: Select all
795|6..|8..
81.|.95|.76
.46|.87|.95
---+---+---
67.|8.9|.5.
92.|56.|78.
.58|.7.|6.9
---+---+---
56.|7..|9.8
.37|9.8|56.
.89|.56|..7

49 clues. SE rates it 10.4. thanks to dobrichev for the puzzle


The first thing I did was simplify the look of it, replace all 5,6,7,8,9 with X and, because only four numbers remain unfilled use the dot method placement for four unknowns rather than for 9 unknowns.

You're right about two things. First, the "line guess" method I described doesn't suffice on its own. Second, it's a remarkably difficult Sudoku for so little fill-in.

I solved it in about 1 hour, which is about twice as long as I normally need to solve an average Black Belt Sudoku (ie. Level 3 and above).

There are 24 possibilities for Row 1. Try all 24. 14 are quickly eliminated leaving 10. For each of the 10 in turn try all possibilities for Row 8. That's all it took.

Pat wrote:
mollwollfumble wrote:

    If there are more than 28 [given] digits in the puzzle
    then it’s likely to be very easy

really?
here's a little collection of 742 puzzles having 39 cells given,
tell me if any of them is "very easy"

here's an example — not a puzzle i would ever try —
Code: Select all

 . . . | . . . | . . 1
 . . 2 | . 1 3 | . 4 5
 1 4 . | . 2 5 | 3 6 .
-------+-------+------
 . 2 4 | . 7 8 | . 5 .
 5 7 . | 3 . 2 | . 8 4
 8 . . | . 5 . | . . .
-------+-------+------
 . 8 . | . . 7 | . 1 .
 2 . 7 | . 8 1 | 4 3 .
 4 1 . | 2 3 . | . . 8



I timed the solution to this one (I don't normally time Sudokus). 1 hour and 8 minutes.

The first thing I did was to fill all the blank cells using the method of complete dots I posted earlier. That took 8 minutes. Then I wasted time listing all possibilities for Row 7 (too many possibilities) and Row 9 (too few possibilities) before settling on Row 4 with Row 9 as a backup. I tried all eight options for Row 4 before finding the solution.

I've noticed in the past that whenever anyone posts what they claim is a difficult Sudoku, somebody else comes along and says "no, I found it easy". Well, I didn't find those two easy, but I didn't find them very difficult either. Perhaps there aren't any difficult Sudokus.
mollwollfumble
 
Posts: 8
Joined: 09 April 2014

Re: Write up?

Postby mollwollfumble » Mon Sep 15, 2014 5:05 am

mollwollfumble wrote:
Code: Select all
795|6..|8..
81.|.95|.76
.46|.87|.95
---+---+---
67.|8.9|.5.
92.|56.|78.
.58|.7.|6.9
---+---+---
56.|7..|9.8
.37|9.8|56.
.89|.56|..7

49 clues.
Solved in about 1 hour

Code: Select all
 . . . | . . . | . . 1
 . . 2 | . 1 3 | . 4 5
 1 4 . | . 2 5 | 3 6 .
-------+-------+------
 . 2 4 | . 7 8 | . 5 .
 5 7 . | 3 . 2 | . 8 4
 8 . . | . 5 . | . . .
-------+-------+------
 . 8 . | . . 7 | . 1 .
 2 . 7 | . 8 1 | 4 3 .
 4 1 . | 2 3 . | . . 8

Solved in 1 hr
Perhaps there aren't any difficult Sudokus.


With fewer clues, "difficult" Sudokus get much more difficult. This one has 23 clues.
Code: Select all
 5 . . | 4 . . | 8 . .
 . . . | . 9 . | . 1 .
 . , 2 | . . 1 | . . 5
-------+-------+------
 6 . . | 3 . . | 4 . .
 . 5 . | . 7 . | . . .
 . . 4 | . . . | . . 8
-------+-------+------
 3 . . | 6 . . | 7 . .
 . 6 . | . . . | . 8 .
 . . 8 | . . 2 | . . 1


10 and a half hours for this last one. I didn't say I was fast.

The solution method I used is exactly the same for all three puzzles. Line guess with a second line as backup. It's not really fair to this technique to dismiss it as "brute force". It's better thought of as an example of "divide and conquer", which in computer-speak would be called "parallel processing" and in scientific terms is a type of "analysis", ie. the process of dividing a single difficult problem into a number of easier ones.

For the last puzzle above I first listed all 30 possibilities for row 7. All but five of these possibilities are quickly eliminated. For each of these 5 list all the possibilities for Row 9, (12, 20 or 8 possibilities) and work though each in turn until finding the solution.
mollwollfumble
 
Posts: 8
Joined: 09 April 2014

Re: Write up?

Postby mollwollfumble » Sat Oct 18, 2014 10:58 am

mollwollfumble wrote:
mollwollfumble wrote:I'm new to this forum.

I've been tempted on many occasions to write up the techniques I use to solve Sudokus.

I use mental techniques: low-hanging fruit, arrows, and left-middle-right.
I use written techniques: casual dots, row end numbers, casual numbers, complete dots, and row guess.
Some techniques I no longer use except on rare occasions: X-wing, 5 spaces, complete numbers.

"Row guess" with "complete dots" sufficed to get me right through a Black Belt Sudoku book without difficulty.

Should I write up my techniques?

Are there any other write-ups of Sudoku solving techniques that distinguish between "casual dots" and "complete dots"?


Now written up and on the web at http://freepages.misc.rootsweb.ancestry.com/~hallsofjamaica/sudoku%20solving%20dots.pdf

Where possible I've used the standard Sudoku terminology from the pages linked by David P Bird, but those pages seem to miss three fairly-well-known solving aids that have appeared elsewhere on the web and in Sudoku book introductions.


Now updated write-up. On the web at http://freepages.misc.rootsweb.ancestry.com/~hallsofjamaica/sudoku%20solving%20dots3.pdf

The new write-up includes some changes on considering your feedback, addition of more examples of gridding and casual dots. Addition of "two-line alternatives" for the hardest Sudokus, and methods suitable for the Sudoku variants "arrow sequences sudoku" and "killer sudoku".
mollwollfumble
 
Posts: 8
Joined: 09 April 2014

Re: Write up?

Postby eleven » Wed Oct 22, 2014 7:18 pm

If i understand it right, your approach to solve hard puzzles is very similar to one of 2 favorites i had, when i thought about how to solve hardest sudokus manually on the lonely island, some years ago.
The other was to try the 2 or maximum 3 digits in a cell, and devide and conquer.

However i think, that in champagnes list of 1.5 mio hardest puzzles (there, in ph_1409.zip) there are enough, which will take you a lot longer to solve them.
I had no time to calculate, which of the two methods would be better on average, but for sure there will be puzzles, which are extremely hard for the one or the other.
eleven
 
Posts: 3082
Joined: 10 February 2008

Re: Write up?

Postby ixsetf » Wed Oct 29, 2014 4:39 am

I think part of the reason you are finding sudoku puzzles so easy is that the "row guess" you do. Most people who play sudoku try to avoid guessing based techniques as they tend to limit the potential difficulty of a puzzle. If you want to challenge yourself I would recommend trying to complete these puzzles without your row guess or other guessing techniques.
ixsetf
 
Posts: 50
Joined: 11 May 2014

Re: Write up?

Postby tarek » Wed Oct 29, 2014 10:54 am

Designing a difficult singles semi-backtracking puzzle for manual solvers who like to guess:

Small sized M3 backdoor puzzle. That has small sized M2 puzzles. That has Small sized M1 Puzzles
Morph the puzzle so that the the 3 backdoors are centrally located within the puzzle (or at the end if targeting left to right backtrackers only)
Morph the puzzle so that the the 3 backdoors have middle values (or high values if targeting small to big backtrackers only)

If it is pure backtracking that is employed then adopting more "worst case scenario" morphs of the puzzle will achieve also a good result

Tarek
User avatar
tarek
 
Posts: 3762
Joined: 05 January 2006

Next

Return to General