The New Sudoku Players' Forum

[billsimp]

A simple question from a new sudoku player. My daily paper has six puzzles each week rated from one to six stars. When I do the six star, I get to the point where I have loops, in that there seems to be no solution without taking a probability guess, which I don't want to do. Is my assessment correct or is there always a defined,logical move in one square that I am missing ?

[champagne]

A simple question from a new sudoku player. My daily paper has six puzzles each week rated from one to six stars. When I do the six star, I get to the point where I have loops, in that there seems to be no solution without taking a probability guess, which I don't want to do. Is my assessment correct or is there always a defined,logical move in one square that I am missing ?

it would be easier to comment on an example. Can you give one

champagne

[Luke]

billsimp, if you're "stateside" it sounds like your daily paper publishes the Knight Features puzzle distributed by Universal Press Syndicate. They're the only ones I've seen with the one to six star format.

The six star is almost always what is known as a "One Trick Pony." There's one advanced technique that will solve the puzzle. As champagne suggested, if you post one of these, someone will point out the way home.

[billsimp]

You are correct. I am stateside and the Sudoku I use is the Knight Features version in the daily paper. I tossed Sunday's puzzle out of frustration. I will keep the next one and try to figure out how to send you as such over this forum. Question? Do you want me to fill as many squares as possible before I send it to you?...Thanks for your help as my question is being followed by several of my frustrated friends.

[billsimp]

Me again with one more point. The one trick pony you spoke of......is this simply a move based on a probability(or a guess as we call it) or is there a defined logical move based on elimination of all numerical possibillities?

[Luke]

No problem, just post next week's puzzle if you get stuck, or maybe one of the folks around here can find this week's and put it up.

I did the puzzle of which you speak, but I don't recall the "one trick". No guessing is required to solve any published puzzle these days, and most assuredly not the weekly six star. That said, this puzzle is a great way to pick up on many of the next level techniques that you're looking for.

Click here for some tips on how to post your puzzle. Best is both the original puzzle and the point at which you got held up.

[ronk]

The one trick pony you spoke of......is this simply a move based on a probability(or a guess as we call it) or is there a defined logical move based on elimination of all numerical possibillities?

Reading the short One Trick Pony thread might help.

[Luke]

There's also Ruud's One-Trick Pony archive.

[Luke]

Here's the Knight Features "6 star" from the week after the post by billsimp.

Code: Select all

7291..8.......7.3..3...9...5..2.439.4.......2.925.6..4...9...7..8.6.......7..8469
 
 *-----------------------------------------------------------*
 | 7     2     9     | 1     356   35    | 8     4     56    |
 | 16    4     156   | 8     2     7     | 9     3     156   |
 | 168   3     158   | 4     56    9     | 126   12    7     |
 |-------------------+-------------------+-------------------|
 | 5     7     168   | 2     18    4     | 3     9     16    |
 | 4     16    1368  | 7     9     13    | 156   158   2     |
 | 138   9     2     | 5     138   6     | 7     18    4     |
 |-------------------+-------------------+-------------------|
 | 136   156   136   | 9     4     125   | 125   7     8     |
 | 9     8     4     | 6     7     125   | 125   125   3     |
 | 2     15    7     | 3     15    8     | 4     6     9     |
 *-----------------------------------------------------------*


I've got a "One-Trick Pony" solution that immediately reduces the puzzle to singles (not a chain.) I'll put off posting it in case anyone else wants to "saddle up."

[storm_norm]

Code: Select all

.------------------.------------------.------------------.
| 7     2     9    | 1     356   35   | 8     4     56   |
| 16    4     156  | 8     2     7    | 9     3     156  |
| 168   3     158  | 4     56    9    | 126   12    7    |
:------------------+------------------+------------------:
| 5     7     168  | 2    #18    4    | 3     9     16   |
| 4    *16    1368 | 7     9    #13   | 156   158   2    |
| 138   9     2    | 5    #138   6    | 7     18    4    |
:------------------+------------------+------------------:
| 136 *15-6   136  | 9     4     125  | 125   7     8    |
| 9     8     4    | 6     7     125  | 125   125   3    |
| 2    *15    7    | 3    *15    8    | 4     6     9    |
'------------------'------------------'------------------'

(6=1)r5c2 - (1)r5c6 = (1)r46c5 - (1=5)r9c5 - (5)r9c2 = (5)r7c2; r7c2 <> 6.

singles.

[hobiwan]

Code: Select all

.-----------------.--------------.---------------.
| 7     2    9    | 1   356  35  | 8    4    56  |
| 16    4    156  | 8   2    7   | 9    3    156 |
| 168   3    158  | 4   56   9   | 126  12   7   |
:-----------------+--------------+---------------:
| 5     7    168  | 2  *18   4   | 3    9    16  |
| 4    -16   1368 | 7   9   *13  | 156  158  2   |
| 138   9    2    | 5  *138  6   | 7    18   4   |
:-----------------+--------------+---------------:
| 136   156  136  | 9   4    125 | 125  7    8   |
| 9     8    4    | 6   7    125 | 125  125  3   |
| 2    *15   7    | 3  *15   8   | 4    6    9   |
'-----------------'--------------'---------------'

Empty Rectangle: 1 in b5 (r9c25) => r5c2<>1
Singles

[Luke]

ER is what I saw, also.

[udosuk]

Code: Select all

+-------------------+-------------------+-------------------+
| 7     2     9     | 1     356  *35    | 8     4    *56    |
| 16    4     156   | 8     2     7     | 9     3     156   |
| 168   3     158   | 4     56    9     | 126   12    7     |
+-------------------+-------------------+-------------------+
| 5     7     168   | 2    -18    4     | 3     9    *16    |
| 4     16    1368  | 7     9    *13    |-156  -158   2     |
| 138   9     2     | 5     138   6     | 7     18    4     |
+-------------------+-------------------+-------------------+
| 136   156   136   | 9     4     125   | 125   7     8     |
| 9     8     4     | 6     7     125   | 125   125   3     |
| 2     15    7     | 3     15    8     | 4     6     9     |
+-------------------+-------------------+-------------------+

r1c69 can't be [55]
=> r4c9+r5c6 can't be [63]
=> r4c5+r5c78, seeing r4c9+r5c6, can't have 1

(Alternatively people might see this as an ALS-xz or xy-chain.)

And then a simple turbot fish, not unlike the one you used to eliminate 6 from r3c3, solves it.