Прошу администратора ОТМЕНИТЬ мою заявку.

Simple explanation here I think, unless I misunderstood the question. Two posts were pending approval as they were first time posts from people, and some of us were away for the weekend and slow to approve it.

Preliminary results have been published.
I ask Christian to check his Hitori.
Nikola and Duс, please check your answers to puzzle 12. The lines should not overlap.

Oh wow, thank you for a great contest Riad! The puzzles were hard, but they are always hard. And thank you to all the contestants with short names. 😎
I don't think anybody has got 100 points before. If it is useful for others, I will share some thoughts on how I approached them.

Opti-Sumdoku:
The first step was to find a grid with a high Samedoku score. I looked at patterns for each digit separately. It is not possible to place a digit in 9 cells so that it counts 18 times to the sum (when every digit can see another digit in both diagonal directions). The best possible is 16 times (see the 9s in my solution). But not all digits can be 16 - the best grid overall has {16, 15, 15, 14, 14, 14, 14, 14, 14} for the Samedoku. I think Carlos has the same grid as me.
The second step is to take away as many digits as possible. Check for uniqueness, try another digit. Eventually you will get down to ~22 clues.
The last step is to choose which digit goes in each part of the grid pattern. Calculate 16/15/14 minus the number of hints remaining (e.g. 9 gets the best place = 16-1, 8 gets the 14-0, 7 gets 15-2).
Please don't try to solve my classic sudoku! It is unique, but it needs a lot of Xwings, Ywings, chains... 😭

Name Hitori:
I decided to use all the 2-letter and 3 words, and many 4s, and I ignored all 5+ words.
It is easy to have too many blacks, or not enough blacks, so I started by trying to intersect many 4-words together.
I found a 4x4 with 6 4-words inside, I put that in the bottom-left corner. I found another 4x4 with 6 words inside. When I put that in the bottom-right, the common letters (e.g. ADEM-ENDO) helped fill 90% of those cells with black or white. I filled the top-right with 4-words too, sometimes choosing letters for Hitori, sometimes choosing for intersection. At the end I put the final 2- and 3-words in the gaps. I had two empty cells. One I used to finish the Hitori, and one was a bonus point.

Threads With Knots:
I wanted to start with long diagonal lines for the 1st, 3rd, 5th, 7th sections to get a high length. I put small numbers near the corners. The 8,5,3,7 clues define knots 1 to 7. The 2 clue is not needed, but it did give lots of empty squares on the long diagonal.
Then I switched to finding a unique solution with small clues. Thread length was less important now. Putting clues on the same parity/colour diagonals helped too. I found two medium numbers (9, 13) to finish.
SP1 and EKBM both had much better clues than me, but my long thread was important. It is possible to have length > 100, but then you need too many big clues to make it unique.

Riad, thanks for organizing a great contest!

And congratulations to kiwijam for a perfect win!!!

Thanks also to the LMI team for running the event.

I had not participated in this contest for a few years, but I really enjoyed it.



I especially like prob. 7 and 12.

I didn't have time to fix my solution for 12 because it was the last day of the contest when I realized that the length of lines was important.I regret that.

Anyway, I think prob.12 is a really good optimization setup!



I like optimization puzzles very much, and I am very happy to see the April Contest where I can enjoy them.

I'm planning to hold an optimization contest at a local site in Japan.

I hope that the fun of optimization will spread more and more.

Edited by SP1 2025-04-29 10:13 PM

Thanks a lot to everyone!

Congratulations to James, Takuma and Paul!

Success in optimization problems in 2025 earned more points than ever. James had a very impressive victory. Only two participants managed to score points in all optimizations. Takuma, great result! Interestingly written in the two posts above. I wish Takuma success in organizing the optimization contest in Japan.

There were questions like: Why ban word overlapping in Hitori? I think through the answer format so that it is convenient for the organizers. A lot of parentheses can prevent Deb from forming grids with solutions. For clarity, it is better to place the parentheses sequentially. Thanks to the LMI team, Deb and Prasanna, and a special thanks to Kota for the long-term support!

Thanks Riad and well done James! Are there solutions available to the puzzles? I had a tough time with some of them and was hoping to work through them with the solutions.