In answer to a question Lonespark once asked, yes "world enough" does include money. In this case lots of money used in extremely reckless ways.
Consider everything that's set in, say, New York. The fact that this covers every Marvel game ever is just the start. Those games have to create their own slice of New York, almost always from scratch. That is... a pointlessly inefficient effort wasting way to do things.
I propose instead a pointlessly over ambitious money wasting way to do things. My game company would create the locations in which a games took place at a high level of detail and then pare it down for the needs of the game.
So, since New York was my example:
Want a game of Premium Rush (the answer is, 'yes', by the way) all that matters to you is the layout of the streets and traffic patterns. So you take that from the company's virtual New York. Want a Mirror's Edge style game but set in a real city? Then you take the rooftops. Want Daredevil (ok, so you'd never get the licencing rights) Hell's Kitchen is your playground. Almost everything in Marvel is in New York. Want a thing where you break into Wall Street banks, hack their computers and give the money to the poor? Well then you're going to be grabbing the Wall Street sections aren't you?
Escape From New York, the game, would have you take the virtual New York, and shove it into a state of disrepair, Deus Ex asks you to age it to 2052, assume severe economic downturn and horrible laws that have allowed entire sections to be walled off with the only access being checkpoints. Legendary posits the cars being lifted up and smashed back down, an earthquake, and a giant monster walking through buildings. All three could have a lot of time saved if they already had ordinary New York built, rather than starting from scratch, as their jumping off point.
What about Law and Order: The Game?
The thing is, the possibilities for re-using an environment that's created at a sufficient level of detail are pretty well endless. A social game, a mystery game, a paranormal game, (Ghostbusters: The Game), a racing game, a heist game, a superhero game, action, adventure, RPG, whatever.
The same if it's Moscow, the same if it's anywhere (including Fictional Place.)
But the idea isn't limited to doing way more work than necessary in hopes that it might make later games easier.
It's also in who you hire. Making games requires game designers, obviously. But it doesn't just have to be them. How many games could have benefited in the realism department from someone who understood that it's important to have fucking toilets somewhere?
If you paired off people who understand the constraints of good gameplay with someone who understood the constraints of real life you'd probably end up with a better balance between the two than you do now not because gameplay should be sacrificed for realism (though there are times and places when it would be worth it) but because game designers are not trained in how to design real floor plans, real buildings, real cities, and so forth.
In an ideal process a game set in a realistic-ish world would differ from realistic only when it was a conscious choice to do so (hopefully because it improved the game to do so in that way.)
Said process could begin with teams where one is a game designer and the other is a civil engineer/architect/interiordecorator/whatever but ideally the two would learn from each other via working together and thus both eventually be able to work on their own as developers.
If I wanted an Indiana Jones type game I'd want to hire people who had at least some background in archaeology and also people who had at least some background in whatever culture the magical mcguffin comes from. And the thing is, they could probably help for more than one game, perhaps giving ideas that game-company-without-them would never even consider.
I think there was going to be more stuff.