Much as the entire early wave of fantasy RPGs made some design choices purely because "That's how D&D did it", most space-travel focused science fiction games offered a random star system generation system which was at least vaguely "scientific", because Traveller did, and many decided to trump Traveller by offering "realistic" probabilities for habitable worlds (based on, you know, all the habitable worlds which had been found in 1981), which meant a DM could wear out his dice hand before generating a random system which contained anything other than barren rocks.
I think the worst offender in this regard has probably got to be Spacemaster, which has chart after chart that seldom produced an inhabited system even after hours of rolling. (No, I do not speak from experience. Even I have my limits.) Traveller never made that mistake, and when you rolled up systems for the game you were guaranteed to get inhabited systems. You know, places for adventure? Another game of that era that had the sense to put the adventure first was FASA's Star Trek rpg. They took it for granted that you would be exploring Class M planets, so when you got to a new system the question wasn't if there were habitable planets present, but how many were there. They even condensed the whole generation system into one convenient page.
I'm all in favor of having realistc science in my games, but not when it interferes with the fun. That's especially true of space opera games which are rather unrealistic to begin with. Skewing the system generation system to produce exotic locals for adventure seems to me to be the best direction to go.