The criteria by which a question is deemed off topic should be changed such that these types of questions, provided they are specific enough and the implication is "How does X do Y, so that I may do something simliar in my game, which I may or may not be currently developing," are considered on-topic.
I have two primary reasons:
First: These questions are not trivia and are absolutely relevant for developers seeking out knowledge on how to develop their game.
StackExchange is a great thing because you can google for a specific question and get a specific answer. But sometimes your question wording doesn't quite work out, and you really meant something entirely different which may come up based on search terms in another question or answer framed in the context of asking how a specific game works.
Second: This site has a problem in that it overlaps significantly with StackOverflow.
It was pointed out in chat recently that a question on StackOverflow that would clearly fit extremely well in GameDev.StackExchange was found, but moderators on StackOverflow refused to migrate it. These "how did game X accomplish specific effect Y" would help differentiate GDSE. And if we encourage a culture of focusing on games that had a unique effect they can be a really valuable part of this site.
But isn't that hard to moderate?
Not any harder than filtering out all the redundant questions we currently get anyway that aren't in violation of this rule. and it would actually add explicit value to this community over StackOverflow, as these kinds of questions would definitely not fit on SO but that is not clearly the case for GDSE.
Of course, we could end up with a bunch of "How did Call of Duty 1 implement Physics? How did Call of Duty 2 implement Physics?" and such. But that problem comes with any question asker who thinks their particular problem is special. The community standard is that the effect Y in game X should be unique enough to be exclusively associated with game X as a GDSE question. Everyone who asks, answers, or reviews one of these questions should have an obligation to search for preexisting questions that already provide an answer out of the context of the game, closing the question if it is asked and is found to be redundant.
But proprietary knowledge, breaking contracts!
Who gives a shit?
Don't answer if you can't legally do so. Askers won't care about how it was actually implemented. They want how it was probably implemented. Out of all arguments against allowing these, this is the weakest one.
What about effects that are just painfully out of date?
It's not entirely bad to have trivia like that on here, in my opinion, but I understand why we'd want to keep it out.
One reason to have it is that old approaches can often inspire modern solutions both from a technical and design perspective. "Game Development" is supposed to include design, right? This information is valuable for design, or possibly people experimenting with weird platforms. And if you're experimenting on a weird platform, nobody's going to have a real answer for you, so it's much more useful to ask about older games that were on older platforms with similar but different peculiarities.