November 16, 2014

Post puzzles for others to solve here.

November 16, 2014

Postby ArkieTech » Sat Nov 15, 2014 10:59 pm

Code: Select all
 *-----------*
 |...|..1|.4.|
 |.5.|3..|91.|
 |1..|4..|7.2|
 |---+---+---|
 |...|...|.8.|
 |8.6|.7.|5.3|
 |.7.|...|...|
 |---+---+---|
 |5.1|..3|..9|
 |.63|..7|.2.|
 |.4.|1..|...|
 *-----------*


Play/Print this puzzle online
dan
User avatar
ArkieTech
 
Posts: 3355
Joined: 29 May 2006
Location: NW Arkansas USA

Re: November 16, 2014

Postby Leren » Sat Nov 15, 2014 11:22 pm

Code: Select all
*--------------------------------------------------------------*
| 6     239   29     | 7     59    1      | 38    4     58     |
| 47    5     47     | 3     28    28     | 9     1     6      |
| 1    d389  e89     | 4     569   569    | 7    c35    2      |
|--------------------+--------------------+--------------------|
| 234   29    2459   | 569   13    56     | 124   8     7      |
| 8     1     6      | 2     7     4      | 5     9     3      |
| 234   7     59     | 589   13    58     | 124   6     14     |
|--------------------+--------------------+--------------------|
| 5     28    1      | 68    2468  3      | 468   7     9      |
| 9     6     3      | 58    458   7      | 148   2     14     |
| 27    4     27-8   | 1     2689  2689   | 368  b35   a58     |
*--------------------------------------------------------------*

(8=5) r9c9 - r9c8 = (5-3) r3c8 = (3-8) r3c2 = (8) r3c3 => - 8 r9c3; stte

Leren
Leren
 
Posts: 5035
Joined: 03 June 2012

Re: November 16, 2014

Postby SteveG48 » Sun Nov 16, 2014 1:00 am

Code: Select all
 *-----------------------------------------------------------*
 | 6     239   29    | 7    c59    1     | 38    4     8-5   |
 | 47    5     47    | 3     28    28    | 9     1     6     |
 | 1     389   89    | 4     569   569   | 7     35    2     |
 *-------------------+-------------------+-------------------|
 | 234   29    2459  | 569   13    56    | 124   8     7     |
 | 8     1     6     | 2     7     4     | 5     9     3     |
 | 234   7     59    | 589   13    58    | 124   6     14    |
 *-------------------+-------------------+-------------------|
 | 5    b28    1     |b68    2468  3     | 468   7     9     |
 | 9     6     3     | 58    458   7     | 148   2     14    |
 |a27    4    a278   | 1    c2689  2689  | 368   35   a58    |
 *-----------------------------------------------------------*


(5=28*)r9c139 - (2=6)r7c24 - (6*28=5)r19c5 => -5 r1c9 ; stte

Or a variation on the theme:

(5=28)r9c139 - (28=69)r9c56 - (6=8)r7c4 - r7c2 = (8)r9c3 => -8 r9c9 ; stte
Steve
User avatar
SteveG48
2019 Supporter
 
Posts: 4232
Joined: 08 November 2013
Location: Orlando, Florida

Re: November 16, 2014

Postby bat999 » Sun Nov 16, 2014 2:22 am

And my noob version...
What happens when r3c8 or r9c8 is 5.
When r3c8 is 5.. r3c56 are 69 (ALS) and r3c3 is not 9, it's 8.
So r9c3 is not 8.
When r9c8 is 5, r9c9 is 8 so again r9c3 is not 8.
=> - 8 r9c3; stte

This is probably a longwinded version of Leren's solution. :lol:
8-)
bat999
2017 Supporter
 
Posts: 677
Joined: 15 September 2014
Location: UK

Re: November 16, 2014

Postby SteveG48 » Sun Nov 16, 2014 2:59 am

bat999 wrote:And my noob version...
What happens when r3c8 or r9c8 is 5.
When r3c8 is 5.. r3c56 are 69 (ALS) and r3c3 is not 9, it's 8.
So r9c3 is not 8.
When r9c8 is 5, r9c9 is 8 so again r9c3 is not 8.
=> - 8 r9c3; stte

This is probably a longwinded version of Leren's solution. :lol:


Not quite. This is Leren's:

(8=5) r9c9 - r9c8 = (5-3) r3c8 = (3-8) r3c2 = (8) r3c3 => - 8 r9c3; stte

And this is yours:

(8=5)r9c9 - r9c8 = r3c8 - (5=69)r3c56 - (9=8)r3c3 => -8 r9c3 ; stte

They're quite similar, but Leren uses r3c2 and you use the pair r3c56 for another nice solution. Both establish the strong link between 8's in r9c9 and r3c3. Keep at it. Try to express your solutions as AI chains.
Last edited by SteveG48 on Sun Nov 16, 2014 3:17 am, edited 1 time in total.
Steve
User avatar
SteveG48
2019 Supporter
 
Posts: 4232
Joined: 08 November 2013
Location: Orlando, Florida

Re: November 16, 2014

Postby bat999 » Sun Nov 16, 2014 3:11 am

SteveG48 wrote:.. but Leren uses r3c2...

I see, when r3c8 makes r3c2 the 3, only r3c3 can be an 8 in that box.
Last edited by bat999 on Sun Nov 16, 2014 12:35 pm, edited 1 time in total.
8-)
bat999
2017 Supporter
 
Posts: 677
Joined: 15 September 2014
Location: UK

Re: November 16, 2014

Postby SteveG48 » Sun Nov 16, 2014 3:18 am

bat999 wrote:
SteveG48 wrote:.. but Leren uses r2c2...

I think you mean r3c2.
I see, when r3c8 makes r3c2 the 3, only r3c3 can be an 8 in that box.


Right, thanks! I fixed the typo in my post.
Steve
User avatar
SteveG48
2019 Supporter
 
Posts: 4232
Joined: 08 November 2013
Location: Orlando, Florida

Re: November 16, 2014

Postby Leren » Sun Nov 16, 2014 5:40 am

Hi Bat,

I've noticed that you tend to look for Kraken constructions. In this case your wordy argument can be expressed as :

Code: Select all
Kraken Column 8 Digit 5:

5 r3c8 - (5=69)r3c56 - (9=8)r3c3 - 8 r9c3;

5 r9c8 - (5=8) r9c9              - 8 r9c3; => - 8 r9c3

There is nothing wrong with this logic, but when a Kraken construction has only 2 branches it can always be re-expressed more elegantly as a linear (non-branching) Alternating Inference Chain (AIC).

All you have to do is reverse the sign of the second leg and append it to the first leg as follows:

8 r9c3 - (8=5) r9c9 - r9c8 = r3c8 - (5=69)r3c56 - (9=8)r3c3 - 8 r9c3; => - 8 r9c3

In this form it is a discontinuous loop that assumes that 8 r9c3 is True and shows that it is False, so it must be False.

This AIC is still regarded as slightly inelegant because it starts and ends on Weak links involving the same single digit.

Aficionados of elegance tend to drop the first and last terms so that the AIC starts and ends on Strong links:

(8=5) r9c9 - r9c8 = r3c8 - (5=69)r3c56 - (9=8)r3c3; => - 8 r9c3

The logic of this form is slightly different - at least one of the 8's in the first and last terms must be True, so any 8 that can see both of the 8's in the first and last terms must be False.

Hope this helps, Leren

PS Have a look at the following Hodoku page on chains and loops.

http://hodoku.sourceforge.net/en/tech_chains.php#nl
Leren
 
Posts: 5035
Joined: 03 June 2012


Return to Puzzles