Fork me on GitHub
<
2018-12-23
>
drowsy02:12:23

@misha yeah one of the important things is to have the set of what valids moves really are. Thats why the 3 dim solution really shines. It's easy to create a set of all the valid points of [x y tool] (including valid tool switches) and always filter you possible moves by them,

fellshard04:12:55

Ugh, finally finished day 15. That round-counting thing is a pain to weave in, I just ended up ignoring the edge case (last unit's turn finishes the match), and that worked fine for my input.

fellshard04:12:34

And actually used earmuffs for the time to weave in the elf attack values! It's kinda gross, but I don't need to extend or maintain this later, so...

norman08:12:17

I got a super late start today. Part 1 was trival. Part 2… seems like it’s too late for thatt

Average-user16:12:18

For day 23 part-2. I came up with something: In theory you can only check points that are in the surface of at least one sphere. The problem is that given a point (p) and a radius (r) the points that are in the surface of that sphere are in the order of r*r. and radius are quite large in the input

gklijs16:12:50

There is a theory the points with the most spheres coincide with one of the corners. But my data don't seem to support that..

Average-user17:12:59

not necessarily the corner, but with the surface of one. But yeah, I read some theories about corners being enough, but I don't think is correct

Average-user17:12:21

With my input the corners approach doesn't seem to work

gklijs17:12:39

With mine neither, but the surface is huge 😞

Average-user17:12:52

yes is too much. Maybe the edges?

Average-user17:12:03

There are still huge though

Average-user17:12:55

Im gonna try to find the best corner, and then walk to origin