Have you ever wondered if maybe one of us is an undercover Dev who wants to push us onto the right track?^^
Good to hear that there is one more again
Haha, as I was falling asleep last night, the thought entered my brain of "how long until accusations of foul play? cheating/dev/otherwise... *snore*".
Straight up, this was luck of the PRNG. I did not find it through cheating, through tool-assisted bits (like the verbose log stuff some CMDRs were using for RES post-1.3) & I don't work for Frontier (I wish, but no jobs for us 'Muricans), though I am a programmer. Which leads to this next bit...
We have the benefit of knowing Elite: Dangerous is a computer game, which means we're dealing with what people/computers can do (& not, say, string theory). Figuring out the true availability of the UA is something only Frontier would be able to know. We know there are (likely) only a couple possible sources for the total population of UAs available.
1. Local client PRNG - The likelihood of getting the right spawn is literally down to your clock, whatever magic seed(s) FDev is funneling in as well as any other sources of variables (rank, time in game, perhaps SS recently seen, etc.). This would result in a theoretically unlimited amount of UAs.
2. Networked PRNG - Similar to the local PRNG, but the occurrence is agreed upon across all instances (still theoretically unlimited UAs, but agreed upon between all clients). I find this unlikely, since every players' SS instances synced across all active players? That's a ton of traffic, even if the payloads are small. And for what gain?
2a. Population influenced PRNG - If you want guaranteed rarity, you'd likely be using a weighted percentage (e.g. we have 30k active players online right now, we want it to only appear 1:1000 tries, multiple/divide as desired to produce the actual spawn chance). This would result in low-count (but not hardcoded) availablity, which would be spiky in pattern. Lots of people online, low likelihood. Few people online, higher chance of hitting it.
3. Hardcoded limits - This would be the "only 2 available ever" theory. This one sucks (from a programmer's perspective) because in order to spawn the UA instance, you have to either run a painful query (across all commanders' cargo holds), be actively storing/updating total populace cargo counts (that poor database's I/O would be through the roof) or be caching those values upfront (and endure the invalidation/updating/unification pains). Not impossible, but not what I'd want to be doing for every SS ever.
I mention these because they may play into how I've been able to find the UAs. I'm hunt late US Central (I can rarely sign on before 10pm CDT/3am GMT) & I've been finding them at 3am my time (8am GMT). Broadly generalizing, only the US West players (& maybe some late-afternoon Aussie/Asia/Russia) would still be on at that time, so that's a theoretically low-traffic time (again, I have no inside knowledge). If there is some sort of population-based seeding going on, that might be a component of it.
That said, we'd need an FDev programmer (likely the one who actually implemented it or if other peer-reviewed it) to confirm/deny any of this, so it's all guesswork.