SudoCue Users
A forum for users of the SudoCue programs and the services of SudoCue.Net

Author Message
Caida
Hooked

Joined: 03 Nov 2007
Posts: 38

 Posted: Sat Nov 10, 2007 8:07 pm    Post subject: Texas Jigsaw Killer 32 Hello, Here are my first seven steps for TJK #32. These ones are pretty easy - I need to do some digging before I can figure out the next move. edited to say: I found one more step without any real digging - hasn't led to anything - but I've added it below. Guessing this is supposed to be in little wee text as it is the latest TJK - let me know if I should make it bigger. edited to add: made it bigger As always, any comments/suggestion/corrections to my walkthrough is most appreciated. edited to add: made changes in red per Mike's post below Any additional steps would be really really really appreciated - I'm dying to see this one solved. Caida TJK #32 Walkthrough: Nonet Layout 1 1 2 2 2 2 3 3 3 1 1 1 2 2 5 3 3 3 1 1 2 2 2 5 6 3 3 1 1 4 4 5 5 6 6 3 4 4 4 4 5 6 6 6 6 7 4 4 5 5 6 6 9 9 7 7 4 5 8 8 8 9 9 7 7 7 5 8 8 9 9 9 7 7 7 8 8 8 8 9 9 Prelims: a. 9(3)r1c5 = {126/135/234} (no 7..9) b. 19(3)r1c8 = {289/379/469/478/568} (no 1) c. 8(3)r2c2 = {125/134} (no 6..9); no 1 in r2c3 or r3c1 (common peer) d. 10(3)r3c8 and r8c2 = {127/136/145/235} (no 8,9) e. 20(3)r2c1 and r6c3 and r8c1 = {389/479/569/578} (no 1,2) f. 30(4)r7c5 = {6789} (no 1..5) g. 6(3)r8c4 = {123} (no 4..9) 1. 30(4)r7c5 is fully within n8 1a. {6789} locked within n8 2. r6c7 <> 4,5 as this would eliminate all 4s and 5s from n8 3. LOL for 6789 3a. c789 only have 6789 in n369 3b. r56c6 no 6,7,8,9 4. r2c3 and r3c1 no 1, as this would eliminate all 1s from 8(3)r2c2 This step is no longer needed as it is in the prelims 5. Innies and Outtie n9: r79c7 less r6c9 = 6 5a. max r79c7 = 9(2), max r6c9 = 3 5b. min r6c9 = 1, min r79c7 = 7(2) 5c. r6c9 no 4..9 5d. r79c7 = 7(2)/8(2)/9(2) = {25/34/35/45} (no 1) 5e. LOL for 1 on c789 n369 5f. -> r56c6 no 1 6. 16(3)r5c8: min r5c89 = 13(2) (no 1..3) 7. Innies and Outtie r12: r2c12 less r3c5 = 1 7a. min r2c12 = 4; min r3c5 = 3 (no 1,2) 8. Innies r6789: r6c249 = 11(3) = {128/137/146/236/245} (no 9)Last edited by Caida on Thu Nov 15, 2007 7:55 pm; edited 1 time in total
gary w

Joined: 07 Sep 2007
Posts: 84
Location: south wales

 Posted: Tue Nov 13, 2007 11:28 pm    Post subject: jigsaw killer 32 I've solved this one now.I was able to make the first placement after only................... 1 step.OK I used T&E.The reason I persisted with this was two fold. 1. I hoped that it would give me a better understanding of the puzzle in preparation for,hopefully,a proper attempt at solving it. 2. To have a solution at hand for those trying to solve it..I found my reasoning at times to be rather tortuous,wasn't sure whether my logic was sound and didn't want to spend hours going down the wrong road. Well,all right,three fold..I just wanted to solve the damn thing!! The solution then is 714526839 548931627 631782945 296457381 389145276 827694153 165873492 452369718 973218564 Very useful in even the t&e solution were the 4 outies of the nonet whose top cell is r3c7.These 4 cells=17 and r7c6 and r6c9 are fairly restricted ..[12345} and {123}respectively.I hoped this would point to a rigorous solving path but I haven't found one yet. I hope someone will pick up Caida's initiative and "run with it"..my impression is that this will be a very pleasing puzzle to "get to know" Regards Gary
mhparker
Grandmaster

Joined: 20 Jan 2007
Posts: 345
Location: Germany

Posted: Thu Nov 15, 2007 2:30 pm    Post subject: Re: Texas Jigsaw Killer 32

Hi Caida,

 Caida wrote: edited to say: I found one more step without any real digging - hasn't led to anything - but I've added it below.

Many thanks for making a start on this one. BTW, it's standard practice for team (so-called "tag") solutions to append steps via making a new post, rather than via editing an existing message. Amongst other things, this stresses the team aspect, and generally adds to the excitement and/or interest.

 Caida wrote: Guessing this is supposed to be in little wee text as it is the latest TJK - let me know if I should make it bigger.

It may be the latest TJK (although the "About That Number 4" (ATN4?)should have been labelled a TJK really), but it's been around for a long, long time. Not only that, but it's an extreme puzzle, which Ruud clearly intended as a team exercise. [Aside: Ruud posted it shortly after Para's variants for TJK31 had received an unusally high degree of interest and team activity, which Ruud no doubt hoped would continue with the TJK32 if he could make it sufficiently difficult.] For these reasons, normal text size is fine in this case.

 Caida wrote: As always, any comments/suggestion/corrections to my walkthrough is most appreciated.

I have two points, both of which relate to understanding your steps:

 Quote: c. 8(3)r2c2 = {125/134} (no 6..9)

Because 1 is locked in this cage, and r2c3 is a common peer of all 3 cells, we can eliminate the 1 from r2c3 here. Maybe you could edit that in to your original post?

 Quote: 7a. min r2c12 = 3; min r3c5 = 2 (no 1)

Actually, min. of r2c12 is 4, because the 20(3) cage at r2c1 (which is not listed in the Preliminaries for some reason) cannot contain either of {12}. Thus, by the same logic, r3c5 can't contain a 2 either. If you could update this sub-step accordingly, and add a "Prelim" for 20(3)r2c1, that would be great. For now, I'll carry on assuming that these extra eliminations have been made.

 Caida wrote: Any additional steps would be really really really appreciated

How about one easy one (plus an obvious follow-up) first?:

9. LoL c6789n3689: 4 innies r1234c6 = 4 outies r789c5+r9c4
9a. {45} unavailable in outies
9b. -> no 4,5 in innies

10. 9(3)n25 = {126/135/234}
10a. 4 only available in r1c5
10b. -> no 3 in r1c5

We have thus now (after step 10b) reached the following grid state:

 Code: .-----------------------------------.-----------.-----------------------.-----------.-----------------------. | 123456789   123456789   123456789 | 123456789 | 12456       1236      | 123456789 | 23456789    23456789  | :-----------.-----------.           |           '-----------.           |           '-----------.           | | 3456789   | 12345     | 23456789  | 123456789   123456789 | 1236      | 123456789   123456789 | 23456789  | |           |           '-----------+-----------.           :-----------'-----------.-----------'-----------: | 3456789   | 12345       12345     | 123456789 | 3456789   | 1236789     123456789 | 1234567     1234567   | :-----------+-----------.-----------'           '-----------+-----------.           :-----------.           | | 123456789 | 3456789   | 123456789   123456789   123456789 | 1236789   | 123456789 | 123456789 | 1234567   | |           '-----------+-----------------------.-----------+-----------'-----------+-----------'-----------: | 123456789   123456789 | 123456789   123456789 | 123456789 | 2345        123456789 | 456789      456789    | :-----------.-----------+-----------.-----------+-----------'-----------------------+-----------.           | | 123456789 | 12345678  | 3456789   | 12345678  | 123456789   2345        1236789   | 123456789 | 123       | |           '-----------:           '-----------+-----------.           .-----------'-----------+-----------: | 123456789   123456789 | 3456789     3456789   | 6789      | 12345     | 2345        123456789 | 123456789 | :-----------.-----------'-----------.-----------:           '-----------+-----------.           |           | | 3456789   | 1234567     1234567   | 123       | 6789        6789      | 123456789 | 123456789 | 123456789 | |           '-----------.           |           '-----------.           |           '-----------'-----------: | 3456789     3456789   | 1234567   | 123         123       | 6789      | 2345        123456789   123456789 | '-----------------------'-----------'-----------------------'-----------'-----------------------------------'

From the above grid position, there are several other smaller moves that scrape off a few candidates without making any real impression on the puzzle, but (inspired by Gary) I would prefer to take the following approach instead:

11. LoL r789n789: 3 innies r7c34+r8c4 = 3 outies r6c189
11a. {12} in innies only available in r8c4
11b. -> if r6c9 is either of {12} it must contain the same digit as r8c4

12. r6c9 = 1 -> r79c7 = 7 (step 5) = {25/34} -> r9c45 <> {23} -> r8c4 <> 1 (permutations 6(3))
12a. but this contradicts step 11b
12b. -> no 1 in r6c9

13. r6c9 = 2 -> r79c7 = 8 (step 5) = {35} -> r9c45 <> {3..} -> r8c4 = 3
13a. but this contradicts step 11b
13b. -> no 2 in r6c9

14. Naked single r6c9 = 3

I'll leave the follow-up for now in case someone else has an even better idea, but there's something I like about the above breakthrough combination. True, it may be a hypothetical, but even hypotheticals can be elegant or ugly, and (imho) there's something quite attractive about this one. It's quite compact and the chains don't rely on any modifications to the grid like "tryfurcation" does.

Hope that helps.
_________________
Cheers,
Mike
gary w

Joined: 07 Sep 2007
Posts: 84
Location: south wales

 Posted: Fri Nov 16, 2007 12:23 am    Post subject: 32 Lovely,Mike.A tiny addition.. 15. So r79c7={45} 16. r56c6={45} 17. So r5c89={67} Regards Gary
mhparker
Grandmaster

Joined: 20 Jan 2007
Posts: 345
Location: Germany

Posted: Fri Nov 16, 2007 1:40 pm    Post subject: Re: 32

 gary w wrote: Lovely,Mike.A tiny addition..

Thanks, Gary.

For those attempting to follow this WT in the future, here are Gary's steps in more detailed form:

15. r79c7 = 9 (step 5) = {45}, locked for r7 and n8

16. Hidden pair in c6 at r56c6 = {45}, locked for n6
(Note: also derivable via LoL c789n369)

17. Split 13(3) at r5c89 = {67} (last combo), locked for r5 and n6

Carrying on...

Edit: Fixed minor typo in step 18 (thanks, Caida!)

18. split 8(2) (innies r6789, step 8) at r6c24 = {17/26} (no 4,5,8)

19. 14(3)n56 = {239} (last combo)

20. 17(4)n568 must have one of {45} due to r6c6
20a. -> {1268/1367} blocked
20b. can't have both of {45} (i.e., can't be {1457/2456}), because {456/457} only available in r6c56.
20c. -> no 4,5 in r6c5
20d. additionally, {2357} combo blocked by r6c24 (step 18)
20e. -> possible combos are {1259/1349/1358/2348} (no 6,7)

21. LoL r6789n6789: 7 innies r6c2345+r7c34+r8c4 = 7 outies r3c7+r4c78+r5c6789
21a. outies contain both of {67} (step 17)
21b. -> innies must contain both of {67}
21c. r6c24 (step 18) contains 1 of {67}
21d. only other place for {67} in innies is 20(3)n45,
21e. which must therefore contain 1 of {67}
21f. -> {389} combo blocked
21f. -> 20(3)n45 = {479/569/578} (no 3)

22. LoL r789n789 (step 11): r8c4 = 3
22a. r9c45 = {12}, locked for r9 and n8

23. Naked single (NS) at r7c6 = 3

24. 9(3)n25 (step 10) = {126} (last combo)
24a. -> no 1,2,6 in r2c45 (CPE)

25. 3 in 14(3)n56 (step 19) now locked in r34c7 for c7 and n6

26. Outies n689: r3c6+r6c5 = 11(2) = {29} (last combo), locked for n5
26a. no 9 in r3c5 (CPE)

27. -> split 14(3) at r6c567 = [248/941]
27a. -> r6c6 = 4, r6c7 = {18}

28. NS at r5c6 = 5

29. 17(3)n5 = [647/746] (only possible permutations)
29a. -> r5c5 = 4
29b. r4c6+r6c4 = {67}, locked for n5
29c. no 6,7 in r4c4 (CPE)

30. NS at r2c6 = 1
30a. r1c56 = {26}, locked for r1 and n2

31. Hidden single (HS) in c4 at r6c4 = 6
31a. -> r4c6 = 7 (cage sum), r6c2 = 2 (step 18)
31b. -> split 10(2) at r5c34 = {19}, locked for r5 and n4

32. r6c57 (step 29) = [91]

33. NS at r3c6 = 2
33a. -> r1c56 = [26]

34. r9c45 = [21]

35. HS in r5 at r5c7 = 2
35a. -> r4c8 = 8 (15(3) cage sum)

36. 16(3)n3 = {268/367}
36a. -> r2c7 = 6, r1c7 = {78}, r2c8 = {23}

Now all singles and cage sums to end.

And it only took us 4 months to solve it...
_________________
Cheers,
Mike

Last edited by mhparker on Thu Nov 29, 2007 10:56 pm; edited 1 time in total
gary w

Joined: 07 Sep 2007
Posts: 84
Location: south wales

 Posted: Sat Nov 17, 2007 12:33 am    Post subject: jigsaw 32 Brilliant Mike.Being new to these type of sudokus I completely missed your move 21.I did learn quite a bit about the structure of this puzzle "solving it" by T&E and following your wt I see that this was the crucial breakthrough.How do you spot these LoLs? And even the naked singles don't always leap out at you. I also found the combos for the 17(4) cage to be pivotal and,as I said,the outies of the nonet - top cell r3c7 - 17/4 placed severe restrictions on possible permutations. So I loved the brevity and clarity of your path through this puzzle.You've put one over on Ruud !!! and it didn't turn out to be another no.18.I'm sure Ruud wouldn't quibble over the placement of your initial 3 at r6c9...or would he?? Regards Gary
mhparker
Grandmaster

Joined: 20 Jan 2007
Posts: 345
Location: Germany

Posted: Sat Nov 17, 2007 8:36 pm    Post subject: Re: jigsaw 32

Thanks also to both you and Caida, not only for contributing some of the moves, but also for not being intimidated by this puzzle and for providing the initial spark.

 gary w wrote: How do you spot these LoLs?

It's just practice. The best way of learning LoL that I can recommend is to look at the solution, maybe even printing it out on paper like I used to do [Note: don't use SumoCue to do this though, because it doesn't include the nonets in a printout of the solution ] That way, because you have the digits staring you in the face, you have immediate feedback as to whether you've made a mistake in identifying the matching innies and outies.

 gary w wrote: So I loved the brevity and clarity of your path through this puzzle.

Me. too! If someone were to ask me what type of moves give me the most satisfaction, I would reply: "shortcuts". Other moves, like killer triples and so on, are of course satisfying, too. But nothing (for me) matches that "Eureka!" feeling of finding a shortcut and then afterwards watching the latest version of JSudoku struggling through via the tortuous route, or even giving up completely (as it does with this puzzle). BTW, you're pretty damn good at finding these types of moves yourself!

Having said that, I'm not sure that the word shortcut is appropriate in this case, because it's totally uncertain at the moment as to whether a long-winded, traditional approach is even available for this puzzle! Even Sudoku Solver, which (as expected, coming from Richard) has a combination/permutation engine that is second-to-none, gets completely stuck on this one.

The only problem with these more complex moves is that many newcomers tend to use them much too early, when much simpler moves are still available.

 gary w wrote: I'm sure Ruud wouldn't quibble over the placement of your initial 3 at r6c9...or would he??

The only thing worth quibbling about would be an error in my logic. Unlike some other Sudoku forums, the main focus of the sudocue.net forums is actually doing the puzzles, not identifying individual moves that can readily be implemented as new solving techniques in software. Of course, that happens here as well, but it's only a by-product, not the main objective. Therefore, anything that works to solve a puzzle is automatically valid. The only reason that some moves may sometimes appear to be "rejected" is that there is a better approach available, although even the interpretation of the word "better" is itself very subjective.

The real question for Ruud is: When can we expect TJK33? To be quite honest, I'm fed up of seeing TJK32 staring at me for the last 4 months (!!) whenever I click on the saw image in the picture bar. This puzzle has been holding us up for far too long now, and I'm tired and just want to finally move on.

All the best.
_________________
Cheers,
Mike
gary w

Joined: 07 Sep 2007
Posts: 84
Location: south wales

 Posted: Sat Nov 17, 2007 11:34 pm    Post subject: jigsaw 32 Hi Mike, Agree with everything you say..and I'll take your advice re the LoLs.Without a solution I'm too afraid of mistaking a logical step and going down a blind alley. Best regards Gary
Caida
Hooked

Joined: 03 Nov 2007
Posts: 38

 Posted: Thu Nov 29, 2007 7:05 pm    Post subject: Yay! Mike and Gary, Thank you so much for completing this puzzle!! I'm so glad it is over I'm struggling with TJK33 - but relieved to see that Para has posted a walkthrough already - so no need to wait 4 months to be put out of my misery on that one! Cheers, Caida Gary, To see LOLs I actually have my excel spreadsheet set up so that I can look at only one number at a time. This means that I have a grid that shows me where all the 1s have been placed, can still be placed, and can not be placed. (with separate grids for the 2s and 3s etc) I find this helps me to see LOLs - because I can more easily see where a number appears in only 2 nonets for 2 rows (and therefore cannot be in any other row for those particular nonets. Not sure if I have explained this adequately - but being able to see LOLs really helps me with solving the jigsaws.
 Display posts from previous: All Posts1 Day7 Days2 Weeks1 Month3 Months6 Months1 Year Oldest FirstNewest First
 All times are GMT Page 1 of 1

 Jump to: Select a forum SudoCue - the Website----------------Daily Sudoku Nightmare & ArchiveClueless SpecialsClueless ExplosionsWeekly AssassinsTexas Jigsaw KillersSudoku LiteX-FilesDaily WindokuDaily Jigsaw SudokuSolving Guide & GlossarySamurai ContestGeneral Website Comments Sudoku - the Community----------------Help Me! I'm stuck!Solving Techniques & TipsWebsitesSoftwarePuzzlesPublicationsOff-Topic SudoCue - the Software----------------SupportWishlistCommentsReleases
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 vote in polls in this forum