Exotic patterns a resume

Advanced methods and approaches for solving Sudoku puzzles

Re: Exotic patterns a resume

Postby JC Van Hay » Fri Jul 26, 2013 5:34 pm

daj95376 wrote:I made a couple of corrections and changes to step #2.
I don't understand why the pointings on 7/8 should be noted LC(7/8B9)!? while the "correction(?)" following r8c1=9 was already stated in the previous post
daj95376 wrote:]Note: I already have r1c3<>7 after step #2.
see previous post where exclusions on 3 leads to UP50 (identical to your PM "after the dust clears ...").
daj95376 wrote:It would have been nice if you'd included the cover sets for #3..#5.
Is it what you want ...
#3. Jellyfish(7R3569) : [r3c12 & r5c3]=XWing(r34c89)-[r6c79 & r9c89]=XWing(r69c23) -> r3c12=r59c3 :=> -7r2c3
#4. Jellyfish(8R3478) : [r3c89 & r4c7]=XWing(r34c12)-r78c1=XWing(r78c67) -> r3c89=r48c7 :=> -8r12c7
#5. Swordfish(8C279) : r8c7=r4c7-[r4c2 & r5c9]=XWing(r39c29) -> r8c7=r9c29 :=> -8r9c8; UP81
daj95376 wrote:You miss the Eureka! forum ... don't you! _ :lol: _
Not at all _ :lol: _
JC Van Hay
 
Posts: 719
Joined: 22 May 2010

Re: Exotic patterns a resume

Postby daj95376 » Fri Jul 26, 2013 6:28 pm

JC Van Hay wrote:
daj95376 wrote:I made a couple of corrections and changes to step #2.

I don't understand why the pointings on 7/8 should be noted LC(7/8B9)!? while the "correction(?)" following r8c1=9 was already stated in the previous post

I'm working from the following grid at the beginning of step #2:

Code: Select all
 *-----------------------------------------------------------------------------*
 | 356789  34789   34789   | 3569    3789    35789   | 26789   246789  1       |
 | 156789  14789   4789    | 569     789     2       | 6789    3       4678    |
 | 36789   3789    2       | 369     4       1       | 5       6789    678     |
 |-------------------------+-------------------------+-------------------------|
 | 1389    1389    6       | 7       1239    4       | 2389    2589    2358    |
 | 13789   2       3789    | 1359    6       359     | 4       789     378     |
 | 4       379     5       | 8       239     39      | 23679   1       2367    |
 |-------------------------+-------------------------+-------------------------|
 | 378     5       3478    | 1234    1378    378     | 23678   24678   9       |
 | 3789    6       1       | 2349    3789    3789    | 2378    24578   234578  |
 | 2       34789   34789   | 349     5       6       | 1       478     3478    |
 *-----------------------------------------------------------------------------*

If r78c1=7, then r9c23<>7 and there's a Locked Candidate Type 2 in r9b9. Similarly for r78c1=8. You need the eliminations from these LC2s to justify your subsequent fish patterns for 7/8. That's why I made those corrections/changes.

JC Van Hay wrote:
daj95376 wrote:]Note: I already have r1c3<>7 after step #2.

see previous post where exclusions on 3 leads to UP50 (identical to your PM "after the dust clears ...").

If a candidate doesn't exist in a cell, then it should not be listed in the exclusions! You included r1c3<>7 in your original post for step #3.

JC Van Hay wrote:
daj95376 wrote:It would have been nice if you'd included the cover sets for #3..#5.

Is it what you want ...
#3. Jellyfish(7R3569) : [r3c12 & r5c3]=XWing(r34c89)-[r6c79 & r9c89]=XWing(r69c23) -> r3c12=r59c3 :=> -7r2c3
#4. Jellyfish(8R3478) : [r3c89 & r4c7]=XWing(r34c12)-r78c1=XWing(r78c67) -> r3c89=r48c7 :=> -8r12c7
#5. Swordfish(8C279) : r8c7=r4c7-[r4c2 & r5c9]=XWing(r39c29) -> r8c7=r9c29 :=> -8r9c8; UP81

OMG!!! If this is the way you see "fish", then any discussion here is essentially fruitless. I will say that r6c7<>7 is not a direct elimination from XWing(r34c89). That explains why I couldn't find a cover set for your base set.
daj95376
2014 Supporter
 
Posts: 2624
Joined: 15 May 2006

Re: Exotic patterns a resume

Postby JC Van Hay » Fri Jul 26, 2013 8:59 pm

Danny, Ok I used 78B8 instead of 78R9! Concerning -7r1c3, I think you didn't notice the difference between my 2 first posts.
Concerning "fish", I wrote some time ago that I don't use "fish" vocabulary anymore as it is too complex.
That's why I determine all the single digit exclusions through "manual templating" and I only look afterwards for their simplest interpretations.
However, it isn't always easy to write them down without ambiguity.
As you drove me into a corner concerning JF(7), here are some more details :
Code: Select all
+----------------------+------------+----------------------+
| 6789    4      3     | 69  78  5  | 26789  26789   1     |
| 5       1      89-7  | 69  78  2  | 6789   3       4     |
| 689(7)  89(7)  2     | 3   4   1  | 5      689(7)  68(7) |
+----------------------+------------+----------------------+
| 89      389    6     | 7   1   4  | 389    5       2     |
| 1       2      89(7) | 5   6   3  | 4      89(7)   8(7)  |
| 4       3(7)   5     | 8   2   9  | 36(7)  1       6(7)  |
+----------------------+------------+----------------------+
| 78      5      4     | 1   3   78 | 26     26      9     |
| 3       6      1     | 2   9   78 | 78     4       5     |
| 2       89(7)  89(7) | 4   5   6  | 1      8(7)    3     |
+----------------------+------------+----------------------+
Jellyfish(7R3569) :=> -7r2c3
Interpretations :
as a Forbidding Matrix
Code: Select all
7r2c3
7r5c3==7r5c89
       7r6c79=7r6c2
7r5c3===============7r5c8=7r5c9
7r3c12==============7r3c8=7r3c9
7r9c3=========7r9c2=7r9c8
or in English
r2c3=7(r9c3<>7)->r5c89=7=r6c2(r9c2<>7) and r35c89=7(r9c7<>7);R9 is devoid of 7 :=> -7r2c3
or in ~"Eureka notations"
[7r5c3=7r5c89-7r6c79=7r6c2-7r9c2 and 7r3c12.r5c3=XWing(7r35c89)-7r9c8]=7r9c3 -> 7r3c12=7r59c3 :=> -7r2c3
or as a Kraken 7R9 :
7r9c2-7r6c2=7r6c79-7r5c89=7r5c3-7r2c3
||
7r9c3-7r2c3
||
7r9c8-7r35c8=*FXWing[7r5c3=*7r5c9-7r3c9=*7r3c12]-7r2c3

Finally, from the columns of the Forbidding Matrix, the cover sets are 7[C3B1 B6 C2 C8 C9] !
JC Van Hay
 
Posts: 719
Joined: 22 May 2010

Re: Exotic patterns a resume

Postby champagne » Sat Jul 27, 2013 9:51 am

Hi JC,

I would agree with danny that your development for the point #2 is not so easy to decipher, but I followed your basic concept and could build my own set/cover in XSUDO.

I got the following


18 Truths = {789R3 59R5 78R9 78C7 78N1 123N4 789B4 9B8}
23 Links = {7r6 8r4 9r8 3c14 5c4 6c4 7c12 8c12 9c1248 5n36 7b379 8b379}
5 Eliminations --> r1345c1<>3, r5c3<>3

I only kept from your #2 the fact that 7;8;9 r7cc1 => 7;8;9 in r5c3. this is enough to establish 3r78c1 true, what I show using the link 3c1.

The use of a Kite (or sky scrapper) in the logic can not give a rank 0 logic, but your solution is a nice one.
In serate the kite is not part of the dynamic process nor the nested triplet, so your enter the field of nested chains.


EDIT

I changed slightly the logic to stay within the multi floors and got that equivalent set/cover


24 Truths = {789R3 9R5 78R9 3789C5 78C7 78N1 128N5 678N6 789B4 9B8}
24 Links = {7r6 8r4 9r8 3c1 7c12 8c12 9c1248 5n3 3b58 7b3789 8b3789 9b5}
5 Eliminations --> r1345c1<>3, r5c3<>3

this is a global rank 0 logic, but with many overlapping sets and links
champagne
2017 Supporter
 
Posts: 7465
Joined: 02 August 2007
Location: France Brittany

Re: Exotic patterns a resume

Postby champagne » Sun Jul 28, 2013 2:14 pm

the second puzzle is very close to the first one, I tried the third one.

Here something very close to a double exocet.

The start is the following

........1.....2.....3.4.56......7.8...6.9.3...248..9....9....53.4...96..6...5....;371650;dob;12_12_03;160;1278

Code: Select all
A      B      C     |D      E     F    |G     H     I     
245789 56789  2578  |35679  3678  3568 |2478  23479 1     
145789 156789 1578  |135679 13678 2    |478   3479  4789 
12789  1789   3     |179    4     18   |5     6     2789 
---------------------------------------------------------
1359   1359   15    |123456 1236  7    |124   8     2456 
1578   1578   6     |1245   9     145  |3     1247  2457 
1357   2      4     |8      136   1356 |9     17    567   
---------------------------------------------------------
1278   178    9     |12467  12678 1468 |12478 5     3     
123578 4      12578 |1237   12378 9    |6     127   278   
6      1378   1278  |12347  5     1348 |12478 12479 24789


we have in r7c12 and r8c79 what looks like a double exocet.

No problem to establish that digits 2;7;8 can not share both AALS
If digit 1 share both, then r6c8=7 and we have a deadly pattern in box 4

So, this is identical to a double exocet

r8c3=5 r8c1=3 r7c7=4 ....

we go to the end without major difficulty
champagne
2017 Supporter
 
Posts: 7465
Joined: 02 August 2007
Location: France Brittany

Re: Exotic patterns a resume

Postby champagne » Thu Aug 29, 2013 4:26 pm

After one month of "vacation" I came back to the topic of the last posts.

I extracted puzzles having similarities with the last one and got a file of 2146 puzzles belonging to the data base of potential hardest having a good chance to show the "false double exocet pattern".

hidden below the top part of the file

All these puzzles have 2 conjugated bases but don't show an exocet pattern. The solution is in fact made of complementary digits, so the deal is to show easily that this must be.

The last three column give

. the potential of elimination in a multi floor analysis
. the corresponding digits
. the digits that can not be in both bases in a single floor analysis

in all cases, the potential of eliminations has been seen for the possible double exocet.
In that top list, as in the former example, 3 digits are self complying with the idea that a digit can not be in both bases.( this is not true for the entire list.)
So only one digit has to be considered with a wider scope (normally using only the digits of the multi floor.

I'll post the entire list on my google drive.

As with a double exocet, the puzzle can collapse immediately or still offer some residual resistance.


Hidden Text: Show
Code: Select all
........1.....2.....3.4.56......7.8...6.9.3...248..9....9....53.4...96..6...5....;371650;dob;12_12_03;160;1278  ;278
........1.....2.....3.4..5......3.6...46..5...6..5.7.4..5.3..7.2...6..4.89......5;418484;dob;12_12_03;158;1289  ;289
..............1..2..3.4..15......6....14...37.8..7...1..5..4.7..3..5.9.42..73...6;246934;dob;12_12_03;158;2689  ;269
........1.....2.....3.4..5.....6.5.3..54..6...6.7...4...4.7..3.1...5..6.89...4...;322025;dob;12_12_03;158;1289  ;189
..............1.23..2.345.....6.......3.27.5..8...37.....9.54....4.7...51....627.;759669;dob;12_12_19;156;1689  ;169
98.76.54.7...3...6.....5.2.4..3..69..37.....4....4....3...7...9..4...8.....1.....;1052456;GP;13_07;154;1258  ;258
........1.....2.....3.4.56......7.8...6.9.3...248..9....9....534....96..6...5....;694333;dob;12_12_19;154;1278  ;278
98.76.5..4....93.....2...6.7..3...45.39...7....4..7...3....54.....8.............1;217858;GP;12_11;153;1268  ;268
........1.......2...3.456...3...67..5..8.....7..93...6.47.......6..2....3...674..;372036;dob;12_12_03;153;1289  ;289
98.7..6..5..4...7.......3.27..5...6...9.8.........1...6.....4...97....5...59....6;149863;GP;12_11;153;1238  ;238
........1.....2.....3.4..5....6..7....8.3..4..15.8.9.6..94.8..353..9....8......9.;521159;dob;12_12_03;152;1267  ;167
........1.....2.....3.4..5...4.6..3..15.3.7.8.3.7....9..84.6..3.5..8....4......8.;521551;dob;12_12_03;152;1279  ;179
........1.....2.....3.4..5......67....5.3..8..247.9.3...9..35.834.....9.5...9....;519959;dob;12_12_03;152;1267  ;267
........1.....2.....3.4..5......3.6...64..5...5..6.4.7..4.3..7.2...56.4.89.......;421304;dob;12_12_03;148;1289  ;289
...........1..2.34.3..4.15........26.5......73...1.4....3.5.84..8.9.....54.......;724526;dob;12_12_19;148;2679  ;267
...........1..2.34.3..4.15........26.5......73...1.4....3.5.8...8.9.4...54.......;724515;dob;12_12_19;148;2679  ;267
........1.....2.....3.4..5......67....5.8..3..247.9.8...9..35.843.....9.5...9....;519958;dob;12_12_03;148;1267  ;267
........1.....2..3..4.3..5......4.6...63..5...5..6.3.7..3.4..7.2...56...89.......;712617;dob;12_12_19;148;1289  ;289
champagne
2017 Supporter
 
Posts: 7465
Joined: 02 August 2007
Location: France Brittany

Re: Exotic patterns a resume

Postby champagne » Mon Sep 02, 2013 4:33 pm

I made some adjustments in my code to test the entire file towards the possibility to establish the property.

I could do it for about 80% of the file using the following rules
. in priority a fish analysis (one floor)
. if the first test fails, a dynamic expansion including URs and {kites, sky-scrappers,empty rectangles}.

This is a very promising preliminary result. I'll introduce that pattern in the list of exotic patterns.

Normally the results can be improved splitting the pattern in four sub scenarios (when it applies).

Next step will be to see what is the residual difficulty for such puzzles after the property has been applied in the proper way.
champagne
2017 Supporter
 
Posts: 7465
Joined: 02 August 2007
Location: France Brittany

Re: Exotic patterns a resume

Postby David P Bird » Thu Sep 05, 2013 8:13 am

Hi Champagne

I'm afraid I can't follow your posts very well, but because we work in such different ways, perhaps it doesn't matter.

I've been analysing the first few examples in your list which share some common characteristics that produce some nice short cuts.

For any Almost Double JE with 4 base digits we have the following Potential Eliminations:
    1. The non base digits in the target cells
    2. The base digits in the cells seen by both pairs of base cells
    3. The base digits in the PE cells associated with the partial fish
When there are 4 target cells:
    4. The base digits in any cell seen by all 4 target cells
When the partial fish for all the base digits has a line parallel to the JE band that must hold 3 base digits
    5. The non-base digits in that line in the partial fish cells.

Now your opening examples have 4 target cells and 2 partial fish lines that must hold 3 base digits and when that happens there will be a cross-line that holds two target cells and two partial fish cells that are all confined to holding the base digits. Therefore the base digits must be false in the other cells in that cross line.

When all the partial fish cells occupy a 3x3 array (as in your opening examples) those eliminations will empty one of the cells leaving 8 cells to hold 8 digits and therefore the non-base digits in the two remaining partial fish cells can also be eliminated.

To illustrate using your 5th puzzle
..............1.23..2.345.....6.......3.27.5..8...37.....9.54....4.7...51....627.;759669;dob;12_12_19;156;1689 ;169
Code: Select all
  *-------------------------*-------------------------*-------------------------*
  | 3456789 1345679 156789  | 2578    5689    289     | 1689B   14689   146789  |
  | 456789  45679   56789   | 578     5689    <1>     | 689B    <2>     <3>     |
> | 6789    1679    <2>     | 78      <3>     <4>     | <5>     1689T   16789   |
  *-------------------------*-------------------------*-------------------------*
  | 24579   124579  1579    | <6>     14589   89      | 1389    13489   12489   |
> | 469     1469    <3>     | 148     <2>     <7>     | 1689T   <5>     14689T  |
  | 24569   <8>     1569    | 145     1459    <3>     | <7>     1469    12469   |
  *-------------------------*-------------------------*-------------------------*
  | 23678   2367    678     | <9>     18      <5>     | <4>     1368    168 B   |
> | 23689   2369    <4>     | 1238    <7>     28#     | 13689   13689T  <5>     |
  | <1>     359     589     | 348     48      <6>     | <2>     <7>     89  B   |
  *-------------------------*-------------------------*-------------------------*
    689     169               18               8!

The Almost Double JExocet is (1689)[a:r12c7,r5c9,r8c8],[b:r79c9,r3c8,r5c7]
The spoiler is (8)r8c6 which must be false for the pattern to comply with the partial fish requirements.
The partial fish for the 4 digits are confined to r358c124 with the candidates shown in the bottom margin.

For the DJE to be true the partial fish cells must hold 8 base digits and the target cells 4 base digits
The p fish cells r358c12 must therefore hold 6 base digits which will exclude all the non-base ones.
In row 5, c12 must contain 2 base digits and the targets in c79 must contain the other two.
Therefore r5c4 can't contain a base digit and 4 would have be true there for the DJE to be true
This requires r38c4 to hold the two remaining base digits in the p fish cells, ie (8)r3c4 & (1)r8c4.

Now (8)r8c4 = (8)r8c6 represents the cases when the DJE is true and false, and what we want to find is a contradiction caused when one or other of them is true. My analysis of the puzzles 1,2,4 & 5 (No 3 is a mistake I think) leads me to believe that generally this will involve the use of branched chains from the various PEs from the DJE pattern to the spoiler.

This becomes manageable for a human solver by marking all the PEs for the DJE and constructing chains using the alternating premise that either the DJE or the spoiler is true. This leads to a bunch of eliminations common to either case and eventually onto a contradiction. However this tactic is equivalent to using memory chains.

For example in box 7 r7c5,r8c45 <> 8 is easy to prove this way, but the different paths would be a bit of a pig to notate.

Perhaps some of this could be of use to you.

David
David P Bird
2010 Supporter
 
Posts: 1043
Joined: 16 September 2008
Location: Middle England

Re: Exotic patterns a resume

Postby champagne » Thu Sep 05, 2013 2:55 pm

David P Bird wrote:Hi Champagne

I'm afraid I can't follow your posts very well, but because we work in such different ways, perhaps it doesn't matter.

I've been analysing the first few examples in your list which share some common characteristics that produce some nice short cuts.

For any Almost Double JE with 4 base digits we have the following Potential Eliminations:
    1. The non base digits in the target cells
    2. The base digits in the cells seen by both pairs of base cells
    3. The base digits in the PE cells associated with the partial fish
.....


Hi David, I just react, coming back from sea side activities to the start of your post.

In that list, the property is that we have 2 AAHS sharing the same band and having complementary digits in the solution. I don't look for more, but you do, trying to apply partial exocet properties.

You are likely right and this has to be done to find a quickest path, but I wanted first to see what can be done using only the "complementary" property after it has been established.

In that limited process, the only elimination is in cells seeing the 2 bases, (your point 2). This is usually enough to have one or 2 cells assigned.
champagne
2017 Supporter
 
Posts: 7465
Joined: 02 August 2007
Location: France Brittany

Re: Exotic patterns a resume

Postby champagne » Thu Sep 05, 2013 6:43 pm

David P Bird wrote: My analysis of the puzzles 1,2,4 & 5 (No 3 is a mistake I think)


As you see puzzle 3 as a mistake, I use it as example on my side.

Code: Select all
A       B       C      |D      E     F       |G    H     I   
1456789 1245679 246789 |235689 2689  2356789 |3478 4689  389
456789  45679   46789  |35689  689   1       |3478 4689  2   
6789    2679    3      |2689   4     26789   |78   1     5   
------------------------------------------------------------
34579   24579   2479   |123589 1289  23589   |6    24589 89 
569     2569    1      |4      2689  25689   |258  3     7   
34569   8       2469   |23569  7     23569   |245  2459  1   
------------------------------------------------------------
1689    169     5      |12689  12689 4       |1238 7     38 
1678    3       678    |1268   5     268     |9    28    4   
2       149     489    |7      3     89      |158  58    6   
                                     vvv


assume (it is true) that r12c5 and r89c6 must have different digits in the solution.

Then

r7c5=1
r3c6=7
r1c6=3|5
r4c4=1

and you can continue.

The question is now can you establish the property.

This will come in the next post, but it is very easy for digits 2;6;9 and slightly tougher for digit 8.
champagne
2017 Supporter
 
Posts: 7465
Joined: 02 August 2007
Location: France Brittany

Re: Exotic patterns a resume

Postby champagne » Thu Sep 05, 2013 8:15 pm

Code: Select all
floor "2" after 2r12c5+2r89c6

... .2. ...
... ... ..g
.2. ... ...

.22 2.. .2.
.2. ... 2..
..2 2.. 22.

... ... 2..
... ..2 ...
g.. ... ...

2r3c2 => XW c34 => conflict c7


Code: Select all
floor "6" after 6r12c5+6r89c6

666 .6. .6.
666 .6. .6.
66. ... ...

... ... g..
66. ... ...
6.6 6.. ...

66. ... ...
... ..6 ...
... ... ..g deadly pattern rows 357



Code: Select all
floor "9" after 9r12c5+9r89c6

999 .9. .99
999 .9. .9.
99. ... ...

999 9.. .99
99. ... ...
9.9 9.. .9.

99. ... ...
... ... g..
... ..9 ... same deadly pattern


Code: Select all
after 8r12c5 + 8r89c6 => 8r5c7;9r4c9;7r3c7
A       B       C      |D      E     F       |G    H     I   
1456789 1245679 246789 |235689 2689  235679  |347  4689  389
456789  45679   46789  |35689  689   1       |347  4689  2   
6789    2679    3      |2689   4     2679    |7    1     5   
------------------------------------------------------------
34579   24579   2479   |123589 129   2359    |6    2459  9 
569     2569    1      |4      269   2569    |8    3     7   
34569   8       2469   |23569  7     23569   |245  2459  1   
------------------------------------------------------------
1689    169     5      |1269   1269  4       |1238 7     38 
1678    3       678    |126    5     268     |9    28    4   
2       149     489    |7      3     89      |158  58    6   

here several path to conclude as

7r3c7 => 34 r12c7 => <3>r1c9
9r4c9 => <9>r1c9
Jellyfish 8c3568 => <8>r1c9 conflict
champagne
2017 Supporter
 
Posts: 7465
Joined: 02 August 2007
Location: France Brittany

Re: Exotic patterns a resume

Postby David P Bird » Thu Sep 05, 2013 8:38 pm

Champagne wrote:assume (it is true) that r12c5 and r89c6 must have different digits in the solution.

Then

r7c5=1
r3c6=7
r1c6=3|5
r4c4=1

and you can continue.

The question is now can you establish the property.

This will come in the next post, but it is very easy for digits 2;6;9 and slightly tougher for digit 8.

I considered puzzle 3 to be a mistake because it has a Double JExocet – no almost about it!

Code: Select all
  *----------------------*----------------------*----------------------*
  | 1457   1457   246789 | 235689 2689 A 357    | 3478   4689   389    |
  | 457    457    46789  | 35689  689  A <1>    | 3478   4689   <2>    |
> | 69     269    <3>    | 2689 b <4>    7      | 8      <1>    <5>    |
  *----------------------*----------------------*----------------------*
  | 3457   457    2479   | 135    1289   23589  | <6>    24589  89     |
> | 69     269    <1>    | <4>    2689 b 2689 a | 5      <3>    <7>    |
  | 345    <8>    2469   | 35     <7>    23569  | 245    2459   <1>    |
  *----------------------*----------------------*----------------------*
> | 689    69     <5>    | 2689 a 1      <4>    | 2      <7>    38     |
  | 17     <3>    678    | 1268   <5>    268  B | <9>    28     <4>    |
  | <2>    14     489    | <7>    <3>    89   B | 158    58     <6>    |
  *----------------------*----------------------*----------------------*
    689    269                                    28

(2689)DoubleJExocet:[a:r12c5,r5c6,r7c4],[b:r89c6,r3c4,r5c5]

All the PEs I listed in my previous post are therefore good and this is the reduced grid after they are made.

The two targets cells for A see base set B, and vice versa, so the two JEs must contain 4 different digits. This is true for all double JExocets that have 4 target cells, so there is no need to dive into examining cases.

You will see that I haven't eliminated (1)r4c5 because that depends on r6c5 having no base digits as candidates, which won't always be true.
I also haven't made any follow-on eliminations. But from here the puzzle reduces with simple methods to a (357)BUG pattern.

[Edit] Oops! I've just noticed that there is a spoiler (8)r7c9 that makes it an Almost Double JE. That means that a net is required to prove it's false before all those PEs can be made.
David P Bird
2010 Supporter
 
Posts: 1043
Joined: 16 September 2008
Location: Middle England

Re: Exotic patterns a resume

Postby Leren » Fri Sep 06, 2013 1:30 am

Hi Champagne and David. I've noticed your recent posts on Almost Double Exocets (ADEs) and there seems to be a difference in your understanding of how this move is supposed to work.

I've coded this move based on Champagne's post of 29 December 2012, so I''ll explain how I think this move is supposed to work as I understand Champagne envisioned it.

I'll use puzzle 3 as an example, but my comments should be applicable to any Almost Double Exocet move.

Code: Select all
*--------------------------------------------------------------------------------*
| 1456789 1245679 246789   | 235689 B2689    357-2689 | 3478    4689    389      |
| 456789  45679   46789    | 35689  B689     1        | 3478    4689    2        |
| 6789    2679    3        |t2689    4       7-2689   | 78      1       5        |<<
|--------------------------+--------------------------+--------------------------|
| 34579   24579   2479     | 123589  1289    23589    | 6       24589   89       |
| 569     2569    1        | 4      t2689   T25689    | 258     3       7        |<<
| 34569   8       2469     | 23569   7       23569    | 245     2459    1        |
|--------------------------+--------------------------+--------------------------|
| 1689    169     5        |T12689   1-2689  4        | 1238    7       38       |<<
| 1678    3       678      | 1268    5      b268      | 9       28      4        |
| 2       149     489      | 7       3      b89       | 158     58      6        |
*--------------------------------------------------------------------------------*

The PM shows puzzle 3 status prior to the ADE move. We have an Almost double Exocet (1289) r1c5 r2c5 r5c6 r7c4 / r8c6 r9c6 r3c4 r5c5.
The cross rows are 3, 5 & 7 and the Spoiler digit is 8 which appears 3 times in Row 7 (outside of the Exocets band).

Now it's only necessary to prove that the Spoiler digit 8 can't appear in both of r12c5 or r89c6 for the ADE move to work.

Why is that ? Well, here's the key observation:

Suppose 8 can't appear in r12c5. Then a (269) Jexocet exists in r1c5 r2c5 r5c6 r7c4.

On the other hand suppose 8 can't appear in r89c6. Then a (269) Jexocet exists in r8c6 r9c6 r3c4 r5c5.

So one of two (269) Jexocets exists but we don't know which one it is. No problem !

Suppose the (269) r1c5 r2c5 r5c6 r7c4 Jexocet exists and it's Base cells eventually hold 26. In that case its Target cells will also hold 26.
But the Base cells of the other potential Exocet see these Target cells so they must eventually hold 89.

You can repeat this argument for either potential (269) Jexocet and any combination of digits 269 and you always come to the same conclusion: r12c5, r89c6 must eventually hold (1289).

Therefore you can eliminate (1289) from any cells that can see these 4 cells. As far as I can see these are the only direct eliminations you can make from an ADE move. However in practice
these eliminations usually solve from 1 -3 cells, so there are a lot of follow-on eliminations using basics only.

I've tested Champagne's recent list of ADE puzzles and I find ADE moves in all of them. In some cases the ADE move + basics completely solves the puzzle. In other cases, once you've cleared the board
via the ADE move + basics, further Jexocets are unearthed, leading to a compete solution. Puzzle 3 was the only puzzle that required a move other than the ADE move + basics + other Exocet moves to solve.

Leren
Leren
 
Posts: 5123
Joined: 03 June 2012

Re: Exotic patterns a resume

Postby champagne » Fri Sep 06, 2013 6:46 am

Leren wrote:
I've coded this move based on Champagne's post of 29 December 2012, so I''ll explain how I think this move is supposed to work as I understand Champagne envisioned it.

I'll use puzzle 3 as an example, but my comments should be applicable to any Almost Double Exocet move.


The PM shows puzzle 3 status prior to the ADE move. We have an Almost double Exocet (1289) r1c5 r2c5 r5c6 r7c4 / r8c6 r9c6 r3c4 r5c5.
The cross rows are 3, 5 & 7 and the Spoiler digit is 8 which appears 3 times in Row 7 (outside of the Exocets band).


Leren


First of all, I loaded the entire file in my google drive
with the file name ph_band_comp_AAHS.zip


Hi Leren,

Good that you coded the study of that pattern.

I feel, reading your post, that we are repeating the fruitful old discussion on Exocets and JExocets.



Leren wrote:

Suppose 8 can't appear in r12c5. Then a (269) Jexocet exists in r1c5 r2c5 r5c6 r7c4.

On the other hand suppose 8 can't appear in r89c6. Then a (269) Jexocet exists in r8c6 r9c6 r3c4 r5c5.

Leren



I feel uncomfortable with your proof.
what if '8' is in both bases. :?:
For me, the only thing to do is to prove that this can't be. In your writing, I don't see such a proof.





In the entire file, you'll find a wider variety of situations although all these puzzles have a minimum potential of elimination in a multi floor approach.

I intend to enlarge the scope later selecting all puzzles having complementary AAHS without any limitation.
champagne
2017 Supporter
 
Posts: 7465
Joined: 02 August 2007
Location: France Brittany

Re: Exotic patterns a resume

Postby champagne » Fri Sep 06, 2013 8:53 am

David P Bird wrote:I considered puzzle 3 to be a mistake because it has a Double JExocet – no almost about it!
....
[Edit] Oops! I've just noticed that there is a spoiler (8)r7c9 that makes it an Almost Double JE. That means that a net is required to prove it's false before all those PEs can be made.


Hi David,

Happily I missed your post. I could have waisted time to look for a non existing Bug ;)

None of the puzzles in the file has a JE pattern that my code could have found. I started the search on the file of puzzles having no existing "exotic" pattern.
champagne
2017 Supporter
 
Posts: 7465
Joined: 02 August 2007
Location: France Brittany

PreviousNext

Return to Advanced solving techniques