SudokuSolver Forum

A forum for Sudoku enthusiasts to share puzzles, techniques and software
It is currently Fri Mar 29, 2024 10:04 am

All times are UTC




Post new topic Reply to topic  [ 5 posts ] 
Author Message
 Post subject: Semi-Symmetric Killer 1
PostPosted: Tue Sep 17, 2019 11:21 am 
Offline
Grand Master
Grand Master

Joined: Wed Apr 30, 2008 9:45 pm
Posts: 693
Location: Saudi Arabia
Semi-Symmetric Killer 1

I'm pushing the boundaries again, we have allowed symmetry as part of the solution method in the past - this is a slight extension.

I've come up with a new sudoku layout based on 10 givens which I will publish shortly. In creating an initial solution for simplicity I used symmetry however I made a mistake as in some cases the number was symmetrical with its partner and in some cases symmetrical with itself. As it solved I thought why not try with normal killer so here is the first one:

This is a simple killer however the solution is semi-symmetrical. There are four unknown number pairs and an odd number (which of course goes in r5c5).
Each number pair is semi-symmetrical, consider for example that the number pair is 4,7, so if r3c1 contains a 7 then r7c9 must contain a 4 or a 7.

So deduce the number pairs and solve the sudoku.

Both solvers actually solve this directly SS gives it 3.35 and JS uses 41 fishes.

Using the semi-symmetry I would guess it is at or below 1.0.

Comment welcome as usual.


On a style point as this is about symmetry so the cages should be symmetrical too, as below. On the other hand I can imagine opposing cages that partially overlap giving different information, so I will not rule these out.

Image


Top
 Profile  
Reply with quote  
PostPosted: Sat Sep 21, 2019 11:18 pm 
Offline
Grand Master
Grand Master

Joined: Tue Jun 16, 2009 9:31 pm
Posts: 280
Location: California, out of London
Very cool! I imagine not having symmetrical cages could make it much harder.
Here's how I started...
Semi-Symmeric Killer 1 WT:
1. Innies n12 = +45(6)
Can only be r1c56 = {89}, r3c123 = {789}, r3c4 = 4
-> r1c7 = 4
-> Innies n3 -> r3c9 = 3
Also r4c45 = {12}
Also r12c4 = +8(2)

2. 4 in n1 in r2c123
17(5)r1c3 = {123(47|56)}
4 cannot go in r2c1 since that would put whichever of {35} or {26} is in r1c12 also in r12c4
-> 4 in r2c23
-> 12(3)n1 = {156}
-> 27(5)r1c3 = {234} in n1 and {17} in r12c4

3. -> 1 in r3 in n3 in r3c78
-> 2 in n3 in 16(3)
-> 7 in n3 in r2c78
-> r12c4 = [71]
Also 3 in n2 in r2c56
-> r1c3 = 3 and r2c23 = {24}
-> 2 in n3 in r1c89
-> 2 in n2 in r3c56
Also r4c45 = [21]
-> 1 in c6 in n8 in r89c6
-> since 1 also in 8(3)n9 and since 1 in c3 only in n7 -> r7c3=1

4! Critical Step
11(4)n6 = {1235}
28(4)n4 = {(47|56)89}
Can that be {4789}?
That puts r789c3 = [165] (6 cannot go in r9c3 since innies n7 = r7c1,r9c3 = +12(2))
But that leaves no solution for 13(4)n7 since r2c23 = {24}
-> 28(4)n4 = {5689}

5! Breaking step
-> Since 11(4)n6 and 28(4)n4 must be semi-symmetrical -> {123} must pair with {689}
(It can't be that one of (123) in n6 pairs with the 5 in n4 since that would require the 5 in n6 pairing with a (123) in the 28(4)n4).
-> 4 must pair with one of (57) or be unpaired
-> 20(3)r6c5 must have two of (689) in r6c56 and one of (457) in r7c6
Only solution is [{69}5]
-> (12) pairs with (69) -> (38) is a pair
Also (45) is a pair
-> 7 is unpaired (and goes in r5c5).

6. If 3 were in 8(3)n9 one of (38) would be in 12(3)n1
Since that's not the case -> 8(3)n9 = {125}
-> (26) are a pair and (19) are a pair.

Straightforward from here


Top
 Profile  
Reply with quote  
PostPosted: Sun Sep 29, 2019 6:35 pm 
Offline
Grand Master
Grand Master

Joined: Wed Apr 30, 2008 9:45 pm
Posts: 693
Location: Saudi Arabia
Excellent solution

On symmetric cages: they help identify the pairs but give less on the overall solution.


Top
 Profile  
Reply with quote  
PostPosted: Thu Dec 05, 2019 12:05 am 
Offline
Grand Master
Grand Master

Joined: Wed Apr 23, 2008 6:04 pm
Posts: 1893
Location: Lethbridge, Alberta, Canada
Thanks HATMAN for an interesting puzzle. I think I'm now getting used to semi-symmetry having so far done a couple of the Triankle puzzles in the Other Variants forum; killer fans should also try them.

My solving path was a bit longer as I didn't spot wellbeback's very powerful step 1. After I'd reached the results of that, we followed fairly similar paths but saw the 'critical step' differently.

Here is my walkthrough for Semi-Symmetric Killer 1:
“This is a simple killer however the solution is semi-symmetrical. There are four unknown number pairs and an odd number (which of course goes in r5c5).
Each number pair is semi-symmetrical, consider for example that the number pair is 4,7, so if R3C1 contains a 7 then R7C9 must contain a 4 or a 7.
So deduce the number pairs and solve the sudoku.”

Prelims

a) 21(3) cage at R1C5 = {489/579/678}, no 1,2,3
b) 7(3) cage at R3C4 = {124}
c) 19(3) cage at R5C9 = {289/379/469/478/568}, no 1
d) 20(3) cage at R6C5 = {389/479/569/578}, no 1,2
e) 20(3) cage at R8C1 = {389/479/569/578}, no 1,2
f) 8(3) cage at R8C9 = {125/134}
g) 28(4) cage at R4C3 = {4789/5689}, no 1,2,3
h) 11(4) cage at R4C7 = {1235}
i) 13(4) cage at R7C2 = {1237/1246/1345}, no 8,9
j) 17(5) cage at R1C3 = {12347/12356}, no 8,9

Steps resulting from Prelims
1a. 8(3) cage at R8C9 = {125/134}, 1 locked for N9
1b. 28(4) cage at R4C3 = {4789/5689}, 8,9 locked for N4
1c. 11(4) cage at R4C7 = {1235}, locked for N6
1d. 13(4) cage at R7C2 = {1237/1246/1345}, 1 locked for N7
1e. 7(3) cage at R3C4 = {124}, CPE no 1,2,4 in R56C4

2. 45 rule on N3 2 innies R1C7 + R3C9 = 7 = [43/52/61]
2a. 21(3) cage at R1C5 = {489/579/678}, R1C7 = {456} -> no 4,5,6 in R1C56

3. 45 rule on N7 2 innies R7C1 + R9C3 = 12 = {39/48/57}, no 1,2,6
3a. 20(3) cage at R8C1 = {389/569/578} (cannot be {479} which clashes with R7C1 + R9C3), no 4
3b. 13(4) cage at R7C2 = {1237/1246} (cannot be {1345} which clashes with R7C1 + R9C3), no 5
3c. 45 rule on R89 4 outies R7C2345 = 13 = {1237/1246/1345}, no 8,9

4a. 45 rule on R12 4 outies R3C5678 = 14 = {1256/1346/2345} (cannot be {1238} which clashes with R3C9, cannot be {1247} which clashes with R3C4), no 7,8,9
4b. R3C123 = {789} (hidden triple in R3), locked for N1
4c. 12(3) cage at R1C1 = {156/246/345}
4d. 6 in R3 only in R3C5678 = {1256/1346}, 1 locked for R3, clean-up: no 6 in R1C7 (step 2)
4e. 21(3) cage at R1C5 = {489/579}, 9 locked for R1 and N2
4f. 7(3) cage at R3C4 = {124}, 1 locked for R4 and N5
4g. Min R3C1 = 7 -> max R45C1 = 7, no 7 in R4C1, no 6,7 in R5C1

5. 45 rule on N1 using R3C123 = {789} = 24, 2 outies R12C4 = 8 = {17/26/35}, no 4
5a. 45 rule on N2 using R12C4 = 8, 1 outie R1C7 = 1 innie R3C4 -> R1C7 = 4, R3C4 = 4, R3C6 = 3 (step 2), R4C45 = {12}, 2 locked for R4 and N5
5b. R1C7 = 4 -> R1C56 = 17 = {89}, 8 locked for R1 and N2
[wellbeback got some of these results more directly from 45 rule on N12 6 innies R1C56 + R3C1234 = 45.]

6. Hidden killer pair 8,9 in R2C78 and R2C9 for N2, 22(4) cage at R2C7 cannot contain both of 8,9 (because R3C78 cannot total 5) -> R2C78 must contain one of 8,9 and R2C9 = {89}
6a. 19(3) cage at R5C9 = {469/478/568} (cannot be {289} which clashes with R2C9), no 2
6b. Killer pair 8,9 in R2C9 and 19(3) cage, locked for C9

7. 17(5) cage at R1C3 = {12347/12356}, R12C4 = 8 (step 5)
7a. 12(3) cage at R1C1 (step 4c) = {156} (cannot be {246/345} = {26}4/{35}4 which clash with R1C3 + R2C23 + R12C4 = {135}{26}/{126}{35}), locked for N1
[I spent a lot of time trying to find a way to express this step as a forcing chain without contradictions.]
7b. R1C3 + R2C23 = {234} -> R12C4 = {17}, locked for C4 and N2 -> R4C45 = [21]
7c. 1 in R3 only in R3C78, locked for N3
7d. 22(4) cage at R2C7 contains 1 = {1579/1678}, no 2, 7 locked for N3
7e. R1C4 = 7 (hidden single in R1) -> R2C4 = 1
7f. R1C3 = 3 (hidden single in R1), clean-up: no 9 in R7C1 (step 3)
7g. Naked pair {24} in R2C23, 2 locked for R2
7h. R2C789 = {789} (hidden triple in R2)

8. 28(4) cage at R4C3 = {4789/5689}
8a. Consider combinations for 13(4) cage at R7C2 (step 3b) = {1237/1246}
13(4) cage = {1237} => 6 in C3 only in 28(4) cage
or 13(4) cage = {1246}, 4 locked for N7 => 4 in N4 only in R456C1
-> 28(4) cage = {5689}, locked for N4

9. 28(4) cage at R4C3 = {5689} and 11(4) cage at R4C7 = {1235} correspond to each other -> 1,2,3 must correspond with 6,8,9
9a. R4C45 = [21] corresponds with R6C56 -> R6C56 = {689}
9b. 20(3) cage at R6C5 = {69}5/{89}3 -> R7C6 = {35}
9c. R7C6 = {35} corresponds with R3C4 = 4 but 3 corresponds with one of 6,8,9 -> R7C6 = 5, R6C56 = {69}, locked for R6 and N5
9d. 4 corresponds with 5, 1,2 correspond with 6,9 -> 3 corresponds with 8 leaving 7 as the remaining number which must be in the central cell -> R5C5 = 7
9e. 19(3) cage at R5C9 (step 6a) = {469/478}, 4 locked for C9
9f. 8(3) cage at R8C9 = {125} (only remaining combination), locked for N9
9g. 1 in C9 only in R89C9, locked for N9
9h. 4 in N5 only in R45C6, locked for C6
9i. 7 in C6 only in R89C6, locked for 26(5) cage at R8C6
9j. 7 in N9 only in R7C789, locked for R7

[Now a lot more use of corresponding cells.]
10. 12(3) cage at R1C1 = {156} corresponds with 8(3) cage at R8C9 = {125} -> 2 corresponds with 6 and 1 with 9
10a. R2C23 = {24} correspond with R8C78, no 2,5 in R8C78 -> R8C78 = [64]
10b. R89C6 = {17} (hidden pair in C6), R9C7 = 8 (cage sum)
10c. Naked triple {379} in R7C789, 3 locked for R7 -> R7C4 = 6
10d. R7C1 + R9C3 (step 3) = [84] (only remaining permutation), R2C23 = [42], R7C23 = [21], R8C23 = [37], R89C6 = [17], R46C2 = [71]
10e. R7C45 = [64] -> R8C45 = 10 = [82]
10f. R3C1 = 7 (hidden single in N1) -> R45C1 = 7 = {34}, locked for C1
10g. R6C1 = 2 corresponds to R4C9, no 2 in R4C9 -> R4C9 = 6
10h. 19(3) cage at R5C9 (step 9e) = {478} = {48}7, 8 locked for C9 and N6
10i. Naked pair {35} in R6C47, 5 locked for R6, R6C3 = 8 -> R56C9 = [84]
10j. R6C9 = 4 corresponds to R4C1, no 5 in R4C1 -> R4C1 = 4
10k. R8C9 = 5, R12C9 = [29] -> R1C8 = 5 (cage sum)

and the rest is naked singles, without using corresponding pairs.

Solution:
1 6 3 7 8 9 4 5 2
5 4 2 1 6 3 7 8 9
7 8 9 4 5 2 1 6 3
4 7 5 2 1 8 3 9 6
3 9 6 5 7 4 2 1 8
2 1 8 3 9 6 5 7 4
8 2 1 6 4 5 9 3 7
9 3 7 8 2 1 6 4 5
6 5 4 9 3 7 8 2 1

HATMAN wrote Using the semi-symmetry I would guess it is at or below 1.0:
I would put it somewhat higher. I used a short forcing chain; also step 7a feels like it's in the 1.25 range.


Top
 Profile  
Reply with quote  
PostPosted: Thu Dec 05, 2019 8:16 pm 
Offline
Grand Master
Grand Master

Joined: Wed Apr 30, 2008 9:45 pm
Posts: 693
Location: Saudi Arabia
I also missed that step.


Top
 Profile  
Reply with quote  
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 5 posts ] 

All times are UTC


Who is online

Users browsing this forum: No registered users and 7 guests


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot post attachments in this forum

Search for:
Jump to:  
Powered by phpBB® Forum Software © phpBB Group