Andrew wrote:If the thread only had the one message and that was moved to another thread, I'm guessing the Spoiler Thread would be automatically deleted.
The Spoiler Thread will never be empty. The first post will explain what the thread is all about, how to use it, and so on. The second post is planned to be a template (incl. spoiler warning) that people can copy and paste into the appropriate referencing thread, overtyping as necessary. Only after these posts will the actual spoilers begin. This is because the thread can't be hidden, therefore the best option is to make a feature of it, providing useful information in the initial post when someone opens it directly.
The thread should probably be made sticky, both because of the useful advice it will contain in the initial post, and because otherwise it would be buried amongst the Assassin threads that reference it, frequently changing its position in the list when people post to it. This would probably not appear to be very elegant in practice.
The first two posts (introduction and template), would be followed by the spoiler "slots". I personally would probably choose to create them on demand, but never delete them, so that they can be re-used for other spoilers in the future without appearing in the "View posts since last visit" list for logged on users or causing "Topic Reply Notification" e-mail messages to be sent.
Andrew wrote:My WTs are therefore improved and, one could say, partially optimised
You have the balance just perfect I reckon Andrew. They are methodical (not a criticism), so human and have a good selection of creative moves - especially the hidden subsets and direct links.
Andrew wrote:Don't think I agree that the average of human WTs should be used. Surely the rating should be that of the lower rated of the human WTs?
Haven't gone any further with rating human WT's. If we do, we can talk about average/lowest rating WT's then.
Para wrote:I think Ed asked me if i found Assassin 60 easier than Assassin 66. Definitely not!!!
Thanks for guessing correctly what I was trying to say - very embarrassing for an English teacher . Your (optimized) walk-through made A60 look easy using a classic Para-finding-direct-links-indirectly!! Great walk-through.
I've found a way to get A60, A64V2, A57V2 etc to rate much higher than my earlier methods. But will wait for Richard to finish a bunch of changes to Sudoku Solver before going public. Might even be able to come out of solving retirement in the interim .
Andrew wrote:My WTs are therefore improved and, one could say, partially optimised
sudokuEd wrote:You have the balance just perfect I reckon Andrew. They are methodical (not a criticism), so human and have a good selection of creative moves - especially the hidden subsets and direct links.
Thanks Ed. I do my best. I think I know what you mean by "so human"; I'll take that as a compliment.
sudokuEd wrote:I've found a way to get A60, A64V2, A57V2 etc to rate much higher than my earlier methods. But will wait for Richard to finish a bunch of changes to Sudoku Solver before going public. Might even be able to come out of solving retirement in the interim .
CathyW wrote:I still do the Times puzzles every day - always in less time than they suggest! Deadlys on Fri/Sat vary but, like Gary, 30 mins is probably an average - I'll have to time myself! I would be surprised if many of the Times Deadlys rated above 0.5.
Just out of interest, the puzzles published on 13th and 14th September each took around 15 minutes (suggested time was 1hr 5mins!). Definitely no more than 0.5. There have been harder ones though I usually find there are quite a few 'givens' from 45 rules or cage sums.
The puzzle for 14th September is still on the Times site today http://entertainment.timesonline.co.uk/ ... 447695.ece
It may be gone tomorrow so let me know if you want an image via pm.
CathyW wrote:I still do the Times puzzles every day - always in less time than they suggest! Deadlys on Fri/Sat vary but, like Gary, 30 mins is probably an average - I'll have to time myself! I would be surprised if many of the Times Deadlys rated above 0.5.
Just out of interest, the puzzles published on 13th and 14th September each took around 15 minutes (suggested time was 1hr 5mins!). Definitely no more than 0.5. There have been harder ones though I usually find there are quite a few 'givens' from 45 rules or cage sums.
The puzzle for 14th September is still on the Times site today http://entertainment.timesonline.co.uk/ ... 447695.ece
It may be gone tomorrow so let me know if you want an image via pm.
These took me alittle longer than Cathy..25-30 mins.But certainly MUCH easier than anything on this site!!
In the past few weeks I've been working on Assassin 66V1.5 and Vortex Killer as "background" jobs.
Vortex Killer was definitely the easier of these two puzzles.
I would rate A66V1.5 as a solid 1.75; possibly a bit higher for my solving path with 1.75 for Para's slightly more direct route. Ed currently has Vortex Killer as 1.75 and A66V.15 as 1.5. Maybe he should consider swapping those ratings?
I had been encouraged to keep going by Para's comment "Some fancy combination work".
We both used the same type of breakthrough in R7 with Para's solving path being more effective, possibly because he used R7C14569 together while I used R7C159 and R7C456 separately. That was clearly the key area which made it a narrow solving path that, together with the need to use steps together and "fancy combination work", made me assess it as a solid 1.75.
I missed two of his 45s, outies from R1 and innies for N78, which would have made it easier for me.
Here is my walkthrough for A66V1.5
1. R34C5 = {17/26/35}, no 4,8,9
2. R5C23 = {29/38/47/56}, no 1
3. R5C78 = {29/38/47/56}, no 1
4. R67C5 = {29/38/47/56}, no 1
5. 11(3) cage in N1 = {128/137/146/236/245}, no 9
6. 20(3) cage at R2C3 = {389/479/569/578}, no 1,2
7. 8(3) cage at R2C7 = 1{25/34}, CPE no 1 in R3C89
8. 23(3) cage in N3 = {689}, locked for N3
9. R5C456 = {389/479/569/578}, no 1,2
10. R6C234 = 1{25/34}, 1 locked for R6
10a. R6C678 = {279/369/378/468/567} (cannot be {459} which clashes with R6C234)
11. 11(3) cage at R7C3 = {128/137/146/236/245}, no 9
12. 22(3) cage in N9 = 9{58/67}, 9 locked for N9
12a. 9 in C7 locked in R456C7, locked for N6, clean-up: no 2 in R5C7
13. R34567C9 = {12358/12367/12457/13456}, 1 locked for C9
14. 45 rule on R1 1 innie R1C6 – 5 = 1 outie R2C4 -> R1C6 = {6789}, R2C4 = {1234}
[I missed 45 rule on R1 3 outies R2C456 = 9 which would have made the early stages a lot quicker.]
15. 45 rule on R9 1 innie R9C4 – 2 = 1 outie R8C6, no 1,2 in R9C4, no 8,9 in R8C6
16. R1C789 = {137/245}
17. R1C123 = {138/246} (cannot be {129/147/156/237/345} which clash with R1C789), no 5,7,9
18. 45 rule on R1 3 innies R1C456 = 22 = 9{58/67}
[Alternatively Killer quad 1,2,3,4 in R1C123 and R1C789, locked for R1 but the split 22(3) cage gives a bit more.]
18a. 9 in R1 locked in R1C456, locked for N2
19. 14(3) cage in N2 = {149/158/239/248/257/347/356} (cannot be {167} which clashes with R1C456)
20. 45 rule on R123 3 innies R3C159 = 19 = {289/379/469/478/568}, no 1, clean-up: no 7 in R4C5
21. 45 rule on R4 3 innies R4C159 = 10 = {127/136/145/235}, no 8,9
22. 45 rule on R5 2 innies R5C19 = 3 = {12}, locked for R5, clean-up: no 9 in R5C2378
22a. 9 in R5 locked in R5C456, locked for N5, clean-up: no 2 in R7C5
22b. R5C456 = 9{38/47/56}
23. 45 rule on N1 1 innie R3C1 – 2 = 1 outie R3C4, no 2,3,4 in R3C1, no 8 in R3C4
24. 45 rule on N1 3 innies R23C3 + R3C1 = 22 = 9{58/67}, no 3,4 in R23C3
25. Killer pair 6,8 in R1C123 and R23C3 + R3C1, locked for N1
25a. 11(3) cage in N1 = {137/245}
26. 45 rule on N3 1 innie R3C9 – 3 = 1 outie R3C6, no 3,5 in R3C6, no 2,3 in R3C9
27. 8(3) cage at R2C7 (step 7) = 1{25/34}
27a. No 2 in R3C6 because R23C7 = {15} would clash with R1C789, clean-up: no 5 in R3C9 (step 26)
28. 45 rule on N3 3 innies R23C7 + R3C9 = 11 = {137/245}
28a. 4 of {245} must be in R3C9 -> no 4 in R23C7
29. R3C159 (step 19) = {379/469/478} (cannot be {289/568} because R3C9 only contains 4,7), no 2,5, clean-up: no 3 in R3C4 (step 23), no 3,6 in R4C5
30. 45 rule on N2 3 innies R3C456 = 14 = {167/347} (cannot be {356} because R3C6 only contains 1,4) = 7{16/34}, no 5, 7 locked for R3 and N2 -> R3C9 = 4, R3C6 = 1, R3C45 = {67}, clean-up: no 6 in R1C6, no 2 in R2C4 (both step 14), no 2,5 in R1C789 (step 15), no 3 in R23C7 (step 7), no 5 in R4C5, no 3 in R9C4 (step 15)
30a. Naked pair {67} in R3C45, locked for R3 and N2
31. Naked triple {137} in R1C789, locked for R1, clean-up: no 8 in R1C123 (step 17)
31a. Naked triple {246} in R1C123, locked for N1, clean-up: no 7 in R2C3 (step 24), no 5 in 11(3) cage (step 25a) -> R3C2 = 3, clean-up: no 8 in R5C3
31b. Naked triple {589} in R1C456, locked for N2
32. R3C7 = 2 (hidden single in R3), R2C7 = 5, clean-up: no 6 in R5C8
33. R3C3 = 5 (hidden single in R3) , clean-up: no 6 in R5C2
34. 45 rule on N7 1 innie R7C1 – 3 = 1 outie R7C4, no 1,2,3 in R7C1, no 7,8 in R7C4
35. 45 rule on N7 3 innies R7C1 + R78C3 = 14
36. 45 rule on N9 1 outie R7C6 – 2 = 1 innie R7C9, no 2,6 in R7C6, no 8 in R7C9
37. 45 rule on N9 3 innies R78C7 + R7C9 = 12 = {138/147/237/246/345} (cannot be {156} which clashes with the 22(3) cage)
37a. 2 of {246} must be in R7C9 -> no 6 in R7C9, clean-up: no 8 in R7C6 (step 36)
37b. THEN R78C7 + R7C9 cannot be {246} because that would make 14(3) cage at R7C6 4{46} -> no 6 in R78C7
37c. R78C7 + R7C9 = {138/147/237/345}
37d. R9C789 = {128/146/236/245} (cannot be {137} which clashes with R78C7 + R7C9), no 7
38. 45 rule on R9 3 innies R9C456 = 16 = {169/178/259/349/358/367/457} (cannot be {268} which clashes with R9C789)
39. 45 rule on N8 3 innies R7C456 = 16 = {169/178/259/349/358/367/457} (cannot be {268} because R7C6 only contains 3,4,5,7,9)
40. R34567C1 must have 1,2 in R5C1 = {24789/25689} (cannot be {15789} which clashes with R2C1), no 1,3, 8,9 locked for C1 -> R5C1 = 2, R5C9 = 1, clean-up: no 3 in R7C6 (step 36)
41. Killer pair 4,6 in R1C1 and R34567C1, locked for C1
42. 3 in C1 locked in R89C1, locked for N7
43. R6C234 (step 10) = 1{25/34}
43a. 2 only in R6C4 -> no 5 in R6C4
44. R4C159 (step 21) = {127/136/145/235}
44a. 7 of {127} must be in R4C1 -> no 7 in R4C9
44b. 6 of {136} must be in R4C1 -> no 6 in R4C9
45. R34567C9 (step 13) = {12457/13456} = 145{27/36}, no 8, 5 locked for C9
45a. 6 of {13456} must be in R6C9 -> no 3 in R6C9
46. Killer pair 3,7 in R1C9 and R34567C9, locked for C9
47. 11(3) cage at R7C3 = {128/137/146/245} (cannot be {236} which must have 3 in R7C4; that would make R7C1 + R78C3 6{26})
48. R7C1 + R78C3 (step 35) = {149/158/167/248} (cannot be {257}; that would make 11(3) cage at R7C3 {27}2)
49. 13(3) cage in N7 = {139/157/238/247/256/346} (cannot be {148} which clashes with R7C1 + R78C3)
50. R9C123 = {279/369/378/459/567} (cannot be {189/468} which clash with R7C1 + R78C3), no 1
51. 14(3) cage at R8C6 = 2{39}/2{57}/3{29}/3{47}/4[19]/ 4{37}/5[18]/5{36}/ 6[17]/6{35}/7[16]/7{25}/7{34} (cannot be 2{48}/3{56}/5{27} because R9C4 is 2 more than R8C6 which would make R9C456 4{48}/5{56}/7{27}, cannot be 4{28} because R9C456 = 6{28} clashes with R9C9), no 8 in R9C5
52. 45 rule on R789 3 innies R7C159 = 17 = {269/278/359/368/458/467}
52a. Using steps 34 and 36 the following relationships are not allowed
R7C1 2 more than R7C9 which would make R7C1 the same as R7C6
R7C1 3 more than R7C9 which would make R7C4 the same as R7C9
R7C1 5 more than R7C9 which would make R7C4 the same as R7C6
R7C1 3 more than R7C5 which would make R7C4 the same as R7C5
R7C5 2 more than R7C9 which would make R7C5 the same as R7C6
52b. R7C159 = [467/485/647/692/872/935] (cannot be [593/683/782/845/863/953/962] which are all inconsistent with step 52a), no 5,7 in R7C1, no 5 in R7C5, no 3 in R7C9, clean-up: no 2,4 in R7C4 (step 34), no 5 in R7C6 (step 36), no 6 in R6C5
52c. R7C456 = [169/187/349/394/574/637], using steps 34 and 36
[The relationships in step 52a were meant to be indented but neither Tab nor a set of leading spaces did that. Very strange!]
53. 14(3) cage at R7C6 = {149/347}, no 8
53a. R78C7 contains 1/3 -> R9C789 (step 37d) must contain 1/3 = {128/146/236}, no 5
53b. R78C7 + R7C9 (step 37) = {147/237/345}
54. Looking at the interactions between R7C159, R7C456, the 11(3) cage at R7C3 and the 14(3) cage at R7C6
54a. R7C159 = [467], R7C456 = [169] clashes with R7C7 = {14}
54b. R7C159 = [485], R7C456 = [187], R78C7 = [34], R78C3 = [28]
54c. R7C159 = [647], R7C456 = [349], R78C7 = [14], R78C3 = [26]
54d. R7C159 = [692], R7C456 = [394], R78C7 = [73], R78C3 = [17]
54e. R7C159 = [872], R7C456 = [574], R78C7 = [37] clashes with R7C3 = {24}
54f. R7C159 = [935], R7C456 = [637], R78C7 = [43], R78C3 = [14]
54g. 4 locked in R7C1567, locked for R7
54h. 7 locked in R7C679, locked for R7, clean-up: no 4 in R6C5
54i. Other eliminations from steps 54a to 54f: no 8 in R7C1, no 6,8 in R7C3, no 5 in R7C4, no 6 in R7C5, no 1,2 in R8C3, no 1,7 in R8C7, clean-up: no 5 in R6C5
54j. R7C159 = [485/647/692/935], R7C456 = [187/349/394/637]
55. 22(3) cage in N9 (step 12) = 9{58/67}
55a. 7 of {679} must be in R8C8 -> no 6 in R8C8
56. R7C1 + R78C3 (step 48) = {149/167/248}
56a. 6 of {167} must be in R7C1 -> no 6 in R8C3
[At this stage I made a flawed step, which happened to lead to the correct solution. After a week away from this puzzle, I looked at this position again and found the following.]
57. 45 rule on R6 3 innies R6C159 = 19 = {289/379/478/568} (cannot be {469} because no 4,6,9 in R6C5)
57a. 2 of {289} must be in R6C9 -> no 2 in R6C5
57b. 9 of {379} and 4 of {478} must be in R6C1 -> no 7 in R6C1
57c. 5,6 of {568} must be in R6C19 and 4 of {478} must be in R6C1 -> no 8 in R6C1
57d. 3 of {379} must be in R6C5, 7 of {478} must be in R6C9 -> no 7 in R6C5
57e. Summary R6C1 = {4569}, R6C5 = {38}, R6C9 = {2567}, clean-up: no 4,9 in R7C5
58. R3C1 = 8 (hidden single in C1), R2C3 = 9, R3C4 = 6, R3C5 = 7, R3C8 = 9, R4C5 = 1, clean-up: no 4 in R8C6 (step 15)
59. R8C9 = 9 (hidden single in C9), R78C8 = {58/67}
59a. Killer pair 6,8 in R2C8 and R78C8, locked for C8, clean-up: no 3 in R5C7
60. Naked pair {38} in R67C5, locked for C5
61. 11(3) cage at R7C3 (step 47) = {128/137}, no 4
61a. 1 locked in R7C34, locked for R7, clean-up: no 9 in R7C6 (step 53), no 7 in R7C9 (step 36)
62. 14(3) cage at R7C6 = {347}
62a. 3 locked in R78C7, locked for C7 and N9
63. R9C789 (step 53a) = 1{28/46}
63a. 6 of {146} must be in R9C9 -> no 6 in R9C7