TL;DR: This is a nice question and has useful answers but it's an unbounded list/poll question with clarity issues. It should get a historical lock because it's still a useful reference despite not being a great fit for the Stack.
What's this question about, anyway?
There's a conflict between the title question and the body question.
There's a question in the title (do RPGs have to be fun?), and a distinctly different question in the body (how do RPGs reward players?) which doesn't even include the idea of games "needing" to be fun. It mentions 'fun' as a potential quality of rewards--yet the preamble talks about 'fun' and 'satisfaction' but not 'rewards,' so there's a logical leap being made which I'm not sure I follow. (I suspect the querent had something specific in mind they were trying to bring readers to Socratically; I see that a lot in early beta-site questions, but it's not really a thing the Stack supports beyond early betas.)
The title question might be answerable within the Stack but "answers should answer the question — that is to say, the body, not the title" and the body question is quite distinct from the title!
When I look at the body question, I see a little phrase that always raises a red flag for Stackability: "What are some?" Literally the querent's just asking for incomplete lists of examples of rewards people get for playing RPGs, and a little commentary on the role of fun in those lists. It's not even limited to system or anything like that which might make the list bounded.
There are many closely related questions which are imminently Stackable: clarification of the term "fun" in design theory (difficult to answer since RPG design is folk art but there's schema for it); how particular kinds of rewards create fun or other motives (this would be bounded); how did "fun" become a watchword in the design dialog (answers would be clearly rankable by accuracy); etc. And of course if the querent is struggling with something in their own design work, describing the specific problem would alleviate the matter dramatically. None of these questions were asked, but a lot of the better answers seem to be answering them.
As written this is an unbounded list/poll question that doesn't match title to body content.
What kind of answers is it getting?
Looking at the answers (a good way to gauge a question's success within the Stack Exchange paradigm!), I see a lot of awesomely experience-based answers, but they are all a little... off. Most of them are equally right, equally useful answers to the body-text question but they don't seem to actually be addressing its topic head-on, focused more on 'fun' and less on 'rewards.'
If I look at the answers as responses to the title-text question they make more sense! But they're trying valiantly to answer a question that's smashed into a pithy title without much regard for the body text. That's a problem.
Because of the way it's asked, every answer that provides an example of a reward is an equally valid response to the body-text question. On its own that's a problem of Stackability because that suggests it's a "chatty, open-ended question" where "every answer is equally valid".
But voting clearly shows the groupmind doesn't think they're all equally useful! Perhaps this is because it seems like most of the answers are focused on the title-text question instead? At any rate, I see strong indicators that we've run into a problem with the Stack's sorting algorithms here.
So what should we do?
There's room in the Stack for game theory questions, but we've learned from hard experience that on the Stack Exchange, being "on topic" is usually shorthand for "fits the formula" rather than anything about the actual subject matter.
At any rate, there's clear problems with the question's clarity and something needs to give. If it were a new question I'd ask the querent to re-phrase it to be clearer. But even if the querent returned to do so, we generally don't dramatically edit questions after so many answers have been given--we encourage new questions to be asked instead.
So let's give it a historical lock! It's a useful question, with useful answers, it's just not matching up with our current understanding of the Stack structure as a provider of rankable solutions to actual problems.
And if we want to continue giving answers related to these topics, we can ask new, better scoped questions about them.