SudokuSolver Forum http://www.rcbroughton.co.uk/sudoku/forum/ |
|
Christmas Killer 2017 http://www.rcbroughton.co.uk/sudoku/forum/viewtopic.php?f=3&t=1423 |
Page 1 of 2 |
Author: | Andrew [ Mon Dec 25, 2017 5:35 am ] |
Post subject: | Re: Christmas Killer 2017 |
Thanks HATMAN for a fun Christmas Killer. Here is my walkthrough: Rating Comment: Solution: |
Author: | enxio27 [ Mon Dec 25, 2017 4:37 pm ] |
Post subject: | Re: Christmas Killer 2017 |
1.1 on SS? I should be able to handle that. I will give it a try. Here's the Sumocue for anyone who wants it (let me know if I made any mistakes): Hidden Text: |
Author: | enxio27 [ Fri May 25, 2018 2:00 am ] |
Post subject: | Re: Christmas Killer 2017 |
I've started this puzzle again (for the third time). The first time, I misplaced it. The second time, I got part way through it, then got busy and set it aside. I'm trying again, but I will have to review killer-specific techniques. |
Author: | enxio27 [ Mon May 28, 2018 5:39 pm ] |
Post subject: | Re: Christmas Killer 2017 |
I'm having a tough time of it with this puzzle. Below is as far as I've gotten. Needless to say, I don't do many killers, but hopefully I haven't missed anything too simple or obvious. I understand the principle of innies and outies, but not sure how to apply them here (or even IF they can be applied here). I would prefer receiving instruction on applicable techniques rather than specific hints. Code: .-------------------------------.-------------------------------.-------------------------------.
| 45678 1234567 45678 | 12345678 89 56789 | 56789 12345678 12345678 | | 1234567 1234567 234567 | 123456789 1234567 123456789 | 3456789 123456789 123456789 | | 456789 1234567 456789 | 1234567 1234567 1234567 | 345678 12345678 12345678 | :-------------------------------+-------------------------------+-------------------------------: | 34567 34567 12 | 89 34567 89 | 12 34567 3456 | | 23456789 23456789 23456789 | 1234567 1234567 1234567 | 3456789 3456789 3456789 | | 4567 89 1234567 | 234567 234567 234567 | 12 89 456 | :-------------------------------+-------------------------------+-------------------------------: | 6789 123456789 23456789 | 123456789 89 123456789 | 3456789 123456789 6789 | | 1234567 1234567 234567 | 1234567 1234567 456789 | 456789 456789 456789 | | 123456789 123456789 23456789 | 123456 123456 123456 | 3456789 123456789 123456789 | '-------------------------------.-------------------------------.-------------------------------' |
Author: | Ed [ Tue May 29, 2018 7:04 am ] |
Post subject: | Re: Christmas Killer 2017 |
enxio27 wrote: I would prefer receiving instruction on applicable techniques rather than specific hints. Unfortunately, most of the everyday techniques we use here don't have nice explanations in separate posts (we have done that for some of the more advanced techniques). We write walk-throughs (WT) and hope others can learn through understanding them.Anyway, I can see a killer pair at that spot. Also, Andrews WT at your spot does find some other things implied cages (step 2f and 4) 4 cell hidden cage clean-up (step 3) combination block -> hidden killer pair (step 6) (this looks like the key step for this puzzle) If you want to go through Andrew's WT but can't understand something, please post! Good on you for giving this one a go. 1.10 is quite a high score for someone who hasn't done many killers. Do you use Sumocue? Why not SudokuSolver? Cheers Ed |
Author: | enxio27 [ Wed May 30, 2018 1:23 am ] |
Post subject: | Re: Christmas Killer 2017 |
Ed wrote: Unfortunately, most of the everyday techniques we use here don't have nice explanations in separate posts (we have done that for some of the more advanced techniques). We write walk-throughs (WT) and hope others can learn through understanding them. Ed, thanks for getting back to me on this. I'm familiar with the techniques that apply to plain vanilla Sudoku, but I'm still trying to get a handle on the killer-specific ones. I've been avoiding walk-throughs (as well as hints from SudokuSolver) in an attempt to understand the logic behind the techniques instead of just being given the answers. Maybe I do need to go through a few killer walk-throughs, though, in order to grasp the techniques. Quote: Anyway, I can see a killer pair at that spot. I'm still having difficulty understanding just what a killer pair is. Would you please explain? Quote: Also, Andrews WT at your spot does find some other things implied cages (step 2f and 4) 4 cell hidden cage clean-up (step 3) combination block -> hidden killer pair (step 6) (this looks like the key step for this puzzle) If you want to go through Andrew's WT but can't understand something, please post! I guess I will resign myself to checking the WT each time I get stuck on this puzzle. Quote: Good on you for giving this one a go. 1.10 is quite a high score for someone who hasn't done many killers. I am apparently misunderstanding SS scores. I had thought that 1.10 was a pretty easy score. Evidently not. Quote: Do you use Sumocue? Why not SudokuSolver? I do have the puzzle set up in SudokuSolver (using the Sumocue code) and have solved this puzzle to the point you see above. However, I was trying to avoid getting hints, as giving me the answers doesn't really help me see the logic behind the techniques. |
Author: | Ed [ Wed May 30, 2018 8:16 am ] |
Post subject: | Re: Christmas Killer 2017 |
Here's one explanation of killer pairs http://sudopedia.enjoysudoku.com/Killer_Subset.html I think generally the newspaper killers get a score of .80 or .85. What we used to give a rating of 0.75 (ie, too easy for here, but harder than newspaper ones) is a score of .90 or .95. A score of 1.00 is around what Ruud did after the early ones. See here to get a sense viewtopic.php?p=706#p706 As you can see, the SS scores at the lower end don't match with the human ratings. This is because the algorithm is set at a high start since we (Richard and me) were trying to rate Assassins, not newspaper puzzles. BTW - you can JSudoku code which HATMAN put at the base of the puzzle pic in the first post. That works for SS. If you want to use SS for hints, I strongly recommend you go to the Solver menu, then Options, then Scoring radio button then Save radio button. That way you will get hints that match with the actual SSscores. The default settings are arranged according to some other criteria I never asked Richard about. They can be quite bizzare for killers: way too complicated. Also, his terminology is not standard and can be confusing. You are correct, that SS and our WTs don't give the underlying logic many times. I think Andrew's sticky, Advice for Newbies has some suggestions of puzzles to try first and some links to killer lingo, technique sites etc. viewtopic.php?f=3&t=466 As I said, please post if you can't understand Andrew's WT. Cheers Ed |
Author: | enxio27 [ Wed May 30, 2018 11:13 pm ] |
Post subject: | Re: Christmas Killer 2017 |
Ed wrote: Here's one explanation of killer pairs http://sudopedia.enjoysudoku.com/Killer_Subset.html Thanks! Between this and Andrew's explanation to me in email, I think I understand now. I'll have to practice spotting them, though. Quote: I think generally the newspaper killers get a score of .80 or .85. What we used to give a rating of 0.75 (ie, too easy for here, but harder than newspaper ones) is a score of .90 or .95. A score of 1.00 is around what Ruud did after the early ones. See here to get a sense viewtopic.php?p=706#p706 Yep, I definitely misunderstood the ratings for killers. I probably wouldn't have touched this one if I had understood the ratings correctly. I probably should start with newspaper-grade killers instead. Quote: BTW - you can JSudoku code which HATMAN put at the base of the puzzle pic in the first post. That works for SS. Yes, HATMAN added that to his post after I posted the Sumocue code I had exported from either JSudoku or SudokuSolver (I forget now which one). Quote: If you want to use SS for hints, I strongly recommend you go to the Solver menu, then Options, then Scoring radio button then Save radio button. That way you will get hints that match with the actual SSscores. I will try that, thanks! Quote: I think Andrew's sticky, Advice for Newbies has some suggestions of puzzles to try first and some links to killer lingo, technique sites etc. I did go through that sticky and the links that were mentioned, but was still having trouble. I'll have to look again for suggestions of puzzles that are more my speed. Quote: As I said, please post if you can't understand Andrew's WT. Will do, thanks! |
Author: | enxio27 [ Fri Jun 01, 2018 2:49 pm ] |
Post subject: | Re: Christmas Killer 2017 |
Going through Andrew's WT. I guess I'm not going to be able to solve this puzzle by myself. Quote: 2c. 17(5) cage at R1C2 = {12347/12356}, 1,2 locked for N2 Should that be N1, or am I misunderstanding something?Quote: 3a. R6C1 = {567} -> no 5,6,7 in R4C1 + R6C3 I have no clue how you are getting this. Please explain the logic.
|
Page 1 of 2 | All times are UTC |
Powered by phpBB® Forum Software © phpBB Group https://www.phpbb.com/ |