auroux - 2022-04-21 10:47 PM

That depends on the specific optimizer, but basically I first try to build a solution, any solution. Then to think about how happy I am with that solution in the context of the quantity to be optimized, whether my solution could be improved by small changes, or whether a redesign from scratch, building the solution a different way, would likely give a much better answer. After a few attempts, usually, one has a sense of whether one has produced something good (possibly close to optimal) or if it feels that there are some guiding principles that aren't understood yet. And of course, often, there are BIG surprises when discovering what other people have managed to do with the same optimizer problem. I don't want to comment too much while the contest is going on, plus I'd risk making a fool of myself if I made hasty pronouncements and it turns out my solutions are really really far from the optimal ones, but after initial skepticism I became quite fond of this year's optimizers. I think they are extremely well designed, in the sense that one can profitably approach them by hand and get a sense of how they work; and, if I dare say so, I don't think computer brute force search would be particularly more helpful on these than careful thinking and formulating intuition.

kiwijam - 2022-04-22 3:32 PM

Hi Swagatam, Riad's example has players at those same locations too. They are allowed to stand where they can see their own goal. But when you count the number of goal paths, you can only include paths that end at the opposite goal from the start goalkeeper #1.

Riad Khanmagomedov - 2022-04-21 3:36 PM


Hello Christian,
If I knew the answers to all the questions, of course I would have suggested it. I don't program and participate in all tournaments without resorting to this process. Let's leave this question to the discretion of each puzzler.

Carlos - 2022-04-23 7:08 AM

Woah it's nice to see all the puzzlers' solutions for the optimizers!
About the equation, the first three are not valid: "_PLUS_65821=-9403x-7" and "65821=-9403x-7" use two consecutive operators, which is not allowed (based on an answer on this thread). Maybe you will cap these solutions to the optimal one (with score 658210), since this issue was not clarified in the Puzzle Booklet and was fixed only the last day of competition.
Huge thanks for all these beautiful puzzles, some classic ones were really hard and very enjoyable to solve!

Carlos - 2022-04-23 11:08 AM

Woah it's nice to see all the puzzlers' solutions for the optimizers!
About the equation, the first three are not valid: "_PLUS_65821=-9403x-7" and "65821=-9403x-7" use two consecutive operators, which is not allowed (based on an answer on this thread). Maybe you will cap these solutions to the optimal one (with score 658210), since this issue was not clarified in the Puzzle Booklet and was fixed only the last day of competition.
Huge thanks for all these beautiful puzzles, some classic ones were really hard and very enjoyable to solve!

Nizz - 2022-04-23 2:24 PM


Yes, THANK YOU RIAD for some really great puzzles. I love the format, much more satisfying than 60/90 minute frenzy with the printer
On first seeing them, I thought I would hate the optimizations, but in the end I loved them all. It turns out the only thing I really hate is Mechanical Sudoku. The number equation was the least fun of the optimization - mostly because I resolved to manual brute force,. I am happy to see that I managed "local optimums" for Record Score (best solution with symmetrical paths), and Opti-Road (best solution with G=g (or V=v).
As for the top three solutions for "From 0 To 9":
Since these solvers were aware of the possibility to use negation/unary minus, which was explicitly endorsed by Riad on the forum, I would think it fair to allow them the top solution utilizing it:
-65821=-9403x7

Riad Khanmagomedov - 2022-04-23 11:20 AM


Thank you Nis,
In your solution 12. Record Score there were forbidden passes, which reduced your result.
The result -65821 would be minimal, but it was necessary to strive for the maximum.

Christian - 2022-04-23 8:40 PM

Hello,

Sorry, but it seems my answer for problem 12 is not counted.
Sincerely,
Christian

Nizz - 2022-04-23 8:51 PM


Indeed, and indeed.
I also see that I had errors in 2 of the regular puzzles. Bummer.
It does seem like some scores are missing for puzzle 10 (including mine) - not sure if this is also due to invalidity of the answers? Seems unlikely, since "moeve" submitted what looks like the same solution as me, modulo reflection, and scored 8 points for it
Anyway: Thank you, once again.