dailysudoku.com

[smith55js]

1.5i2 2.6d3 3.6b6 4.6f4 5.6h9 6.7g6
7.8i1 8.8f5 9.8c9 10.1g1 11.4h4 12.4g3
13.2f1 14.2h2 15.4b9

16.1b2 {3,9} in b1,a3
17.1a9 {3,9} in b8,c7

18.1e8 19.9d9 20.5g9 21.3f9 22.4f6 23.4e2
24.7c2 25.4c1 26.3d2 27.7d1 28.7e4
29.2a7 {3,9} in row 7

30.7a8 31.5f8 32.7f7
33.2g8 {3,9} in h8,i7

34.9g5 35.3e5 36.9e6 37.1c5 38.9c4 39.5a4
40.3a6 41.9a3 42.3b1 43.9h1 44.3i3 45.2d5
46.1d4 47.5d6 48.2i4 49.1i6 50.3c7 51.9b8
52.9i7 53.3h8

634 752 198
817 349 625
925 681 473

589 176 342
471 238 956
362 594 781

253 467 819
796 815 234
148 923 567

[zaks]

This is my version:
we begin with evident moves:
01 6h1 02 6d7 03 6b4 04 6f6
05 7g4 06 8c1 07 8f5 08 8i9 (so much so good...)

now remove 2 as candidate from cells c8,c8; this creates pair c7,c8=> 3,7
09 4e8 10 2f9 11 4g7 (hidden single in row 7)
12 4f4 (hidden single in column f) 13 4b1 (hidden single in box 7)
14 4c9 (hidden single in box 1) 15 7c8 (hidden single in box 1)
16 3d7 17 7d8 18 1b8 19 2h8 20 5i8
21 1g9 22 4h6 23 7e6 24 1a1 25 9g5
26 1i4 27 2i6 28 5g1 29 2g2 30 3f1
31 9d1 32 1e2 33 7f3 34 5f2 35 5d4
36 1d6 37 2d5 38 3e5 39 9e4 40 1c5
41 9c6 42 5a6 43 3a4 44 9a7 45 3b9
46 9h9 47 3i7 48 9i3 49 3h2 50 3c3
51 2a3 52 7a2 53 9b2 x!

or i missed smth or this sdk isn't "very hard",

enjoy, zaks

[smith55js]

No, I really don't feel this one should have been labeled "Very Hard". It might be Hard, or maybe even Medium. There were only four moves that required any actual thought processes aside from looking up and down rows and columns or in groups for numbers that stood alone. Are these puzzles being classified by difficulty by a computer or by someone actually doing the puzzles?

[David Bryant]

[smith55js]

Ok, so the difficulty is graded by a computer algorithm. It's apparent now that a change should be made. After doing different puzzles, some that are 'Very Hard' are in fact, Medium or Hard. Some 'Hard' puzzles should be 'Very Hard', etc...

From what I've gleened from the posts about the difficulty of the puzzles and how they are graded, the rating is based on the number of 'difficult' steps required.

Can I make a suggestion? Change it to count not the number of 'difficult' steps, but rather the number of types of 'difficult' steps. If one 'Very Hard' puzzle, like today's, contains 4 or 5 choke points, but they are all the same type of problem, then wouldn't that decrease the difficulty of the puzzle? If I solve a problem once, I should be able to easily recognize the same problem when I encounter it again. But if I encounter a totally new type of problem, I have to think even harder to figure it out. I can't resort to the logic I applied earlier as it won't apply. Thus, making the puzzle more difficult.

[alanr555]

[PoppaPoppa]

I agree with Alan Rayner- I used the mandatory pairs technique and finished it in a little over 30 minutes. This puzzle gave me less pause than some "mediums" I have pulled my hair out over!

Chris