SudokuSolver Forum http://www.rcbroughton.co.uk/sudoku/forum/ |
|
Bloomdoku, Torn-Bloomdoku & Ninjutsudoku http://www.rcbroughton.co.uk/sudoku/forum/viewtopic.php?f=13&t=846 |
Page 2 of 2 |
Author: | Andrew [ Sat Dec 04, 2010 4:53 am ] |
Post subject: | Re: Bloomdoku, Torn-Bloomdoku & Ninjutsudoku |
Thanks Simon for posting an interesting graphical walkthrough for Ninjutsudoku #1.: I found this puzzle a lot harder to solve: My walkthrough is fairly long but many steps come in groups of 2 or 4, because of the symmetry of the cage pattern, in the same way that some of Simon's steps apply in 2 or 4 different positions. Here is my walkthrough for Ninjutsudoku #1: When I first saw the puzzles in this thread I thought that this might be the easiest one, partly because I've only done a few non-consecutive puzzles and those not for a long time. Maybe I'll try one of those now and see how I get on with it. |
Author: | Andrew [ Mon Dec 06, 2010 7:22 am ] |
Post subject: | Re: Bloomdoku, Torn-Bloomdoku & Ninjutsudoku |
I assume we're not really meant to try NC Bloomdoku #1 so I didn't. Thanks Simon for NC Bloomdoku #2. A fun puzzle and much easier than Ninjutsudoku #1. It took me a while to spot my key steps 7 and 8. After that it was fairly straightforward. I suppose that step 21a is my hardest one: Here is my walkthrough for NC Bloomdoku #1. I've given NC eliminations and stated petal placements for those who, like me, don't work with software solvers in editor mode: |
Author: | Andrew [ Wed Dec 08, 2010 5:42 am ] |
Post subject: | Re: Bloomdoku, Torn-Bloomdoku & Ninjutsudoku |
I'm still struggling with Torn Bloomdoku #2 so I'll take the liberty of posting a short comment about Torn Bloomdoku #1 now. Simon wrote: Apperently, combining with the NC (Non Consecutive) rule, this variant has only a space of 2 solutions! So to make a valid puzzle, only 1 more constraint is required: R1C1<>1. That indicates that the other valid puzzle with this petal pattern has R1C1 = 1: |
Author: | simon_blow_snow [ Wed Dec 08, 2010 12:54 pm ] |
Post subject: | Re: Bloomdoku, Torn-Bloomdoku & Ninjutsudoku |
While NC Bloomdoku #2 is quite easily solvable for novices of NC solving, the same can't be said for NC Torn-Bloomdoku #2. In a certain old forum there were a few experts (h3lix, JC Bonsai, udosuk, etc) who developed a few advanced techniques on NC solving. Some of those include a move where you look at a certain group of cells (e.g. pair/triple/quad) and determine that they must include certain candidate(s), or else consecutiveness would occur. This is a special case of "NC locked candidates". I don't think it is trial-and-error or forcing chains but other people may disagree. To solve NC Torn-Bloomdoku #2 you will need to apply the aforementioned technique, not once or twice but many times. But I will still claim it is "human solvable", and I think some other (strong) solvers will agree. The #1 puzzles for these 2 series aren't really for human solvers, but to test a solving program etc. As for Andrew's observation: |
Author: | Andrew [ Thu Dec 09, 2010 7:31 am ] |
Post subject: | Re: Bloomdoku, Torn-Bloomdoku & Ninjutsudoku |
Thanks Simon for your comments about what is required to solve this puzzle without any forcing chains (I don't group those together with trial-and-error, that implies trying something to see if it works while forcing chains make one or more eliminations in cells other than at the start of the hypothesis and stop there; that seems to be a standard distinction between them on the killer forum). I used several forcing chains, mostly short ones but step 14, which was where I resumed today, is a bit longer (it looks longer than it really is because it includes NC eliminations and hidden single statements). Here is how I solved NC Torn-Bloomdoku #2: Maybe after a few more days, to allow other solvers the chance to use your hints, you can post a walkthrough showing how this puzzle should be solved using the non-consecutive techniques referred to in your post. I'm definitely a novice at NC solving so will be interested to see how they should be used. I may have managed: |
Author: | Ed [ Sun Dec 12, 2010 12:47 am ] |
Post subject: | Re: Bloomdoku, Torn-Bloomdoku & Ninjutsudoku |
simon_blow_snow wrote: NC Torn-Bloomdoku #2.... a few advanced techniques on NC solving. Some of those include a move where you look at a certain group of cells (e.g. pair/triple/quad) and determine that they must include certain candidate(s), or else consecutiveness would occur. This is a special case of "NC locked candidates" I used closer to the inverse, an extreme sort of "NC blocked candidates". Same as Andrew's steps 6 & 10. Then found some more instead of Andrew's step 14, like a squeeze move. Very enjoyable puzzle. Thanks Simon!NC Torn-Bloomdoku #2 ALT step 14: Cheers Ed |
Author: | simon_blow_snow [ Tue Dec 14, 2010 11:56 am ] |
Post subject: | Re: Bloomdoku, Torn-Bloomdoku & Ninjutsudoku |
After retrying NC Torn-Bloomdoku #2 and eliminating the unnecessary steps, I found this unbelievably short solving path! 4-step crack-in for NC Torn-Bloomdoku #2: How about this, a one-step-to-break-it-all puzzle (did Ruud call these one-trick-ponies or something)? So I want to see what Andrew & Ed think: 1. Are steps 1-3 forcing-chain-free? (I think they are.) 2. Should step 4 (first line) be considered as a forcing chain? (I don't think so.) |
Author: | Ed [ Wed Dec 22, 2010 11:40 pm ] |
Post subject: | Re: Bloomdoku, Torn-Bloomdoku & Ninjutsudoku |
About Torn-Bloomdoku 2 simon_blow_snow wrote: I found this unbelievably short solving path! Really good solution Simon! Don't think it's particularly short but have no hesitation in adding my . It's quite similiar to Andrew's with my alternative step though a couple of your steps are simpler. I've described them here.simon_blow_snow wrote: a one-step-to-break-it-all puzzle (did Ruud call these one-trick-ponies or something)? Sure doesn't feel like a O-T-P to me. Lots of NC tricks in there (see link above). You must be very experienced at NC puzzles if all the other NC tricks felt routine. Interesting that we still think about Ruud nearly 3 years after he disappeared! You've presumably been lurking for quite a while!! As to your questions about forcing chains...I can't comment since I don't even know what the technical definition of one is...I just know I like chains when I've found one!! Thanks again for a really great puzzle. Ed |
Author: | Andrew [ Sat Jan 08, 2011 12:45 am ] |
Post subject: | Re: Bloomdoku, Torn-Bloomdoku & Ninjutsudoku |
I've been busy with other things so have only just gone through Simon's and Ed's latest posts in this thread and the linked posts in the techniques forum. Nice Crack-In by Simon. It shows what a NC expert can achieve, compared with my walkthrough from a NC novice. It included several neat steps; as Ed commented the final lines of steps 2 and 3 were both killer-like, well they were killer pairs even though this isn't a killer sudoku. Simon wrote: Are steps 1-3 forcing-chain-free? (I think they are.) I'd say almost forcing-chain-free. The neat second line of step 2 "R5C56 must include {2} of R5 --> R5C47<>[13] --> R5C4<>{1}" is probably a short forcing chain.Ed wrote: As to your questions about forcing chains...I can't comment since I don't even know what the technical definition of one is...I just know I like chains when I've found one!! I haven't seen a definition either. I agree that it feels good once one has found a forcing chain. I've got nothing against forcing chains; they are preferable to contradiction moves although I find the latter easier to spot.Simon wrote: Should step 4 (first line) be considered as a forcing chain? (I don't think so.) It's permutation analysis, using NC properties, so not a forcing chain.
|
Page 2 of 2 | All times are UTC |
Powered by phpBB® Forum Software © phpBB Group https://www.phpbb.com/ |