List of similar seeds found so far: 19063, 19660 & 23 9287 & 17476 16 & 33 9756 & 20752 57 & 62 I've been looking for interesting planets that share similar terrain for relatively fair multi-planet starts. In doing so, I found something very peculiar: some seeds will always generate very similar terrain (not necessarily similar brushes) with minor differences, regardless of planet size. I don't know how the random generation takes place, so I couldn't say if this was the result of some strange self-organization, or if I just got extremely lucky with my seed selection. Anyway, I figured others would find this cool, and maybe someone here would like the seeds to some similar-but-not-identical planets.
now we need a cryptologist to figure out what seeds will give awesome planets (and a lot more information)
While I'd like multi-planets starts to be fair, I also want the planets to have some differences, to make choosing a starting planet a bit interesting. Normally I just look for seeds that have the same type of choke points, etc, usually with a totally different overall layout. In this case though, it seems you can get two similar planets with different brushes and somewhat different terrain that are probably balanced, without being completely identical. I'd usually want maps that are even more different than that, but it's an interesting middle-ground between exactly the same and completely different.
Was once again trying to get balanced planets for a system, and came across another pair of very similar seeds: 9287 & 17476. Since this actually appears to be a rather common occurrence, I figure I'll update the OP with a list if I happen to find any more, or if anyone else finds and suggest some. Still not sure why this happens though.
Hash collisions? With the hash referring to the planetary phenotype and collision being defined most widely?
From the games I've played metal distribution has been similar (not identical) for large groups, and quite different for smaller groups of metal spots. I've got a feeling it has to do with gradient hashing in the simplex noise used to generate the terrain height, but that's just wild speculation.
Decided to test how common this is by going through the first few seeds to see if I could find any that were extremely similar. I might have missed one or two because I was going one by one and trying to remember if I had seen a similar one as I went, but the first set I could find was 16 & 33, so out of 34 planets, at least two were nearly identical. EDIT: Wrote 23 instead of 33 accidentally. Ironically, it turns out that 23 is extremely similar to 19063 and 19660 as well, so chances are there are even more nearly identical planets than just 'twins'.
I found 3 or 4 that were somewhat similar; 44 is probably 80% similar, 66 is around 85% similar, just if I were to guess off the top of my head. Haven't found an identical one though.
I'll look at those as soon as I'm done messing around with build orders. If I find them to be satisfactory you will receive your cookie.
