Sudoku Players' Forum (clone)

I've created a new app for android ( called "Eye Candy Sudoku" that presents a set (150) of Sudoku puzzles I've generated which are aesthetically pleasing, i.e. due to the pattern of the clues. I am entertaining the idea of self publishing a hard copy book with similar puzzles.

I am interested in hearing the opinions of Sudoku enthusiasts about my endeavor. While the app costs $3.99 on google play, I'd consider proving a free version to individuals interested in reviewing it and providing feedback.

Example puzzles:
http://www.yesterdata.com/pub/puz1.jpg
http://www.yesterdata.com/pub/puz2.jpg

Thanks,
James

3FB504F3 wrote:I've created a new app for android ( called "Eye Candy Sudoku" that presents a set (150) of Sudoku puzzles I've generated which are aesthetically pleasing, i.e. due to the pattern of the clues.

I'm sure you're aware of the (very) long thread here, and also this. Also, you're perhaps unaware of the fact that the two 'puzzles' that you show are invalid. Here are valid versions:

Code: Select all

 . 9 3 . . 6 . . .
 . 2 1 . . 4 . . .
 . 4 5 . . 7 1 9 .
 . 5 . . . 8 . 1 .
 . 8 . . 9 . . 3 .
 . 3 . 4 . . . 6 .
 . 6 4 5 . . 3 2 .
 . . . 8 . . 6 5 .
 . . . 6 . . 4 7 .   easy

 . . . 4 6 1 3 . .
 . 7 . . . . . 8 .
 1 . . . . . . . 2
 6 . . . 3 . . . 5
 7 . . 1 . 6 . . 3
 8 . . . 7 . . . 1
 3 . . . . . . . 9
 . 2 . . . . . 1 .
 . . 9 8 2 7 . . .     hard

Regards,

Mike Metcalf

OK, I inadvertently posted "obfuscated" versions of my puzzles. I've overwritten the two given links with the real versions. Also, here are six more:

http://www.yesterdata.com/pub/a1.png

Again, just wondering if there is a market for publishing my large volume of "pretty" puzzles.

Another question: I've restricted all of my puzzles (I have over 25 million of them) to the "symmetric" type. Do Sudoku fans care? I could probably create some really cool shapes if I relaxed that restriction.

Thanks,
James

There are certainly people who care a great deal about symmetry, and also many who don't. Most, but not all, people are willing to forgo symmetry if the reason is obvious (as yours is). You will certainly get some complaints, but not so many that it should invalidate your idea.

$3.99 for an app that contains 150 puzzles is overpriced, especially on Android where free apps with unlimited puzzles are common. There is some extra value for your interesting patterns, but not nearly that much. Puzzle books have more leeway to charge reasonable prices for unique puzzles, but apps not so much. The Sudoku app market is especially problematic because there are several hundred Sudoku apps, most of which are junk, and little to help buyers distinguish the good from the bad. Thus it is almost impossible to get noticed.

All just my personal opinion.

Thanks for the comments. Yeah, I pulled the price out of thin air. I'm paying an android app review service, they'll likely confirm your point about the price.

Besides the attractiveness of the puzzles, I'd appreciate feedback about the usability of the app in terms of working on the puzzle.

If anyone is interested, I've made available a temporary free edition of the Eye Candy Android app:

http://www.yesterdata.com/pub/f.zip

Android will make you change your security to install from unknown sources; I haven't had time to figure that out. But you just open the ZIP and install the app. I personally use the "Show Possible" option for the harder puzzles only, and I'm curious how the controls are viewed by others.

Thanks,
James

3FB504F3 wrote:Another question: I've restricted all of my puzzles (I have over 25 million of them) to the "symmetric" type. Do Sudoku fans care? I could probably create some really cool shapes if I relaxed that restriction.

Symmetry is usually regarded as a matter of esthetics. Many published puzzles retain the 180-degree rotational symmetry used by crossword puzzles. Symmetry was required of patterns in the Patterns Game, so there is a lot of experience in generating puzzles with a given pattern. A higher level still is symmetry of givens, as in

Code: Select all

 . . . . . 1 . . 2
 . . 5 . 6 . 7 . .
 7 . . 3 . . . 8 .
 . . 8 . . . . 9 4
 . . . . 5 . . . .
 6 1 . . . . 2 . .
 . 2 . . . 7 . . 3
 . . 3 . 4 . 5 . .
 8 . . 9 . . . . .

where each value v at a location i,j is matched by a 10-v at location 10-i,10-j.

Another desirable but not essential attribute is that a puzzle be minimal (have no redundant clue), unlike

Code: Select all

 . . . . . 2 . . 6
 . . 7 . 4 . 8 . .
 1 . . 3 . . . 4 .
 . . 4 . . . . 5 3
 . . . . 8 . . . .
 3 6 . . . . 7 . .
 . 5 . . . 6 . . 1
 . . 8 . 5 . 2 . .
 2 . . 1 . . . . .

  3 8  is redundant
  7 2  is redundant

Regards,

Mike Metcalf

I checked and approximately 1 out of 400 of my puzzles are "minimal".

Here's a nice pattern that isn't trivial to solve.

Code: Select all

*-----------*
|7.6|...|..4|
|.29|...|..5|
|.4.|...|362|
|-----------|
|...|862|...|
|...|...|...|
|...|943|...|
|-----------|
|967|...|.3.|
|8..|...|54.|
|5..|...|1.6|
*-----------*

An interesting pattern. I have posted a puzzle with this pattern elswhere. Here are some more, in increasing order of difficulty:

Code: Select all

906000008072000003080000214000854000000000000000632000698000020700000830200000401   2
102000003034000005050000126000728000000000000000569000645000080700000210300000507   4   
102000003043000005060000274000537000000000000000146000278000060300000150400000308   5 
102000003045000006060000247000874000000000000000916000378000060200000850600000704   6   
102000003045000001060000752000826000000000000000495000527000030300000690800000205   8   

Here are those last 6 puzzles in line format. All solve with basics + 0 or 1 non-basic move :

7.6.....4.29.....5.4....362...862...............943...967....3.8.....54.5.....1.6 Basics + W Wing
9.6.....8.72.....3.8....214...854...............632...698....2.7.....83.2.....4.1 Basics
1.2.....3.34.....5.5....126...728...............569...645....8.7.....21.3.....5.7 Basics + XY chain length 4/ALS XZ Rule
1.2.....3.43.....5.6....274...537...............146...278....6.3.....15.4.....3.8 Basics + ALS XZ Rule
1.2.....3.45.....6.6....247...874...............916...378....6.2.....85.6.....7.4 Basics + W Wing
1.2.....3.45.....1.6....752...826...............495...527....3.3.....69.8.....2.5 Basics + 6 Cell DP Type 2 Uniqeuness

Leren

Can't seem to get a puzzle with the corners and center empty.

Code: Select all

*-----------*
|...|365|...|
|..5|.2.|...|
|7..|819|...|
|-----------|
|84.|...|.93|
|967|...|548|
|15.|...|.26|
|-----------|
|...|598|..7|
|...|.7.|9..|
|...|642|...|
*-----------*

000365000005020000700819000840000093967000548150000026000598007000070900000642000

A bit harder but still only 1 non-basic move :

Code: Select all

*--------------------------------------------------------------*
| 24    189  d18     | 3     6     5      |c248   7     1249   |
| 36    189   5      | 4     2     7      | 36    18    19     |
| 7     23    46     | 8     1     9      | 2346  35    245    |
|--------------------+--------------------+--------------------|
| 8     4     2      | 7     5     6      | 1     9     3      |
| 9     6     7      | 2     3     1      | 5     4     8      |
| 1     5     3      | 9     8     4      | 7     2     6      |
|--------------------+--------------------+--------------------|
| 2346 f12-3 e146    | 5     9     8      | 24   a36    7      |
| 2456  28    468    | 1     7     3      | 9     56    24     |
| 35    7     9      | 6     4     2      |b38    18    15     |
*--------------------------------------------------------------*

(3) r7c8 = (3-8) r9c7 = r1c7 - (8=1) r1c3 - r7c3 = (1) r7c2 => - 3 r7c2; stte

Here is the puzzle in . line format : ...365.....5.2....7..819...84.....93967...54815.....26...598..7....7.9.....642...

The 0 line format can sometimes cause problems in Excel. If you copy and paste, under certain conditions Excel treats the 0 line format as a large integer and you end up with something like 3.6500000502E+77

Leren

If you want difficult, this one isn't pretty but....

Code: Select all

*-----------*
|..2|.9.|.7.|
|5..|4..|...|
|...|27.|..8|
|-----------|
|.7.|...|3.5|
|9.5|...|7.4|
|3.8|...|.9.|
|-----------|
|8..|.64|...|
|...|..2|..1|
|.1.|.3.|8..|
*-----------*

..2.9..7.5..4........27...8.7....3.59.5...7.43.8....9.8...64........2..1.1..3.8..

The puzzle's not pretty and the solution is downright ugly - not solvable by human friendly methods as far as I can tell.

To illustrate the point I ran the puzzle through Hodoku. Here is just one of several forcing chain moves it uses.

Forcing Chain Contradiction in r4 => r5c4<>6 r5c4=6 r5c2<>6 r5c2=2 r4c1<>2 r5c4=6 r5c4<>3 r1c4=3 r1c9<>3 r2c9=3 r2c9<>9 r9c9=9 r9c46<>9 r8c4=9 r8c4<>8 r8c5=8 r8c5<>5 r6c5=5 r6c5<>4 r4c5=4 r4c5<>2r5c4=6 r5c4<>3 r1c4=3 r1c9<>3 r1c9=6 r6c9<>6 r6c9=2 r4c8<>2

That looks pretty ugly to me, but then beauty is in the eye of the beholder, so someone may like it

Leren

3FB504F3 wrote:Can't seem to get a puzzle with the corners and center empty.
[snip]
000365000005020000700819000840000093967000548150000026000598007000070900000642000

I think you mean a minimal puzzle with that pattern. There are too many givens in some boxes for that. By deleting four clues one can generate, e.g.:

Code: Select all

 . . . 3 . 6 . . .
 . . 1 . 4 . . . .
 3 . . 5 1 2 . . .
 1 7 . . . . . 8 4
 . 8 3 . . . 7 2 .
 6 2 . . . . . 3 9
 . . . 8 9 5 . . 7
 . . . . 6 . 9 . .
 . . . 4 . 7 . . .  Symmetry of givens, hard, minimal


 . . . 5 . 1 . . .
 . . 9 . 2 . . . .
 3 . . 8 7 9 . . .
 9 7 . . . . . 2 4
 . 5 3 . . . 7 6 .
 4 8 . . . . . 3 1
 . . . 1 3 2 . . 7
 . . . . 8 . 1 . .
 . . . 9 . 6 . . .   Hard, minimal


Regards,

Mike Metcalf

Leren wrote:The 0 line format can sometimes cause problems in Excel. If you copy and paste, under certain conditions Excel treats the 0 line format as a large integer and you end up with something like 3.6500000502E+77

This one is fairly easy to overcome - format the destination cell as text and use 'paste as text' in the paste special menu.

Far more annoying for me is how Excel automatically reformats a cell when it thinks the pasted data could be a date.

David