This answer has been discussed by @CRoss and myself, the moderation team, and reflects our thoughts on this issue.
I believe recommendations should be allowed for the time being. RPG systems, much more than most computer games or board games, are toolkits for creation and not simple products. They are also much more long lived and thus a recommendation for one isn't "out of date" in a year the way a computer game recommendation is.
However, there are potential severe quality problems with these questions that one can see affecting the existing game-recommendation questions on RPG.SE.
I propose the following criteria be strictly applied to game recommendation questions on RPG.SE.
Questions should be specific enough that there can be a single or best answer. We will not tolerate polls or lists or "community wiki" for game rec questions. "What are all the medieval fantasy RPGs" or "What's a fun game" are awful and should be immediately closed as "not a real question." A good question is of the format "I want to try a game that has these specific attributes and use it in this way, who has done this and what would you recommend and why?" Don't just ask about games, add context about your needs. This recent Wild West game question is pretty good, in that it is specific about the kind of game it wants, though it could stand to explain more about the asker's needs and predilections. You want "gritty combat?" Explain what you mean by that exactly. If your question is pulling a list and not a "best," it's a bad question.
Remember the goal of a SE isn't just to fart around, it's to help people with their actual gaming. Quizzing people for the sake of doing it or idle curiosity is likely to generate a poor question that is closed.
Game recommendation questions should be tagged [game-recommendation].
Answers MUST adhere to the "Back It Up!" principle set forth in the SO blog post Good Subjective, Bad Subjective. To wit:
- Something that happened to you personally
- Something you can back up with a reference
You may only answer one of these questions if you have actually done it or seen it done, or if you have a detailed reference from someone who has done it (with details like how and results). "Oh, I'm sure FATE would be great for your ultra gritty combat system needs, of course I've never done that or seen anyone that has, but it is SUCH A GREAT SYSTEM that I'm sure it will do it" should be viciously downvoted. If you have never even read the game, you shouldn't be considering answering - someone else can Google "wild west RPG" or put something down they've vaguely heard of. That isn't expert advice for experts. "I know someone put out a Western supplement for game X" is not an appropriate answer. "We played a Wild West game, used system X, and here's how that turned out" is the best kind of answer.
Similarly, nobody wants to hear that you don't like that kind of gaming. If someone asks for gritty combat and you say "oh people don't like gritty combat" - move along. You can make a parenthetical warning about pitfalls as part of an otherwise legitimate answer if you want.
Answers should be voted up the more specific expertise they demonstrate - someone who played that kind of game for a while should be heavily upvoted, someone who has a game like that and has read it somewhat, someone who just knows about the existence of a game should be downvoted, and someone recommending their pet system with no grounds should be nuked.
If we get a huge load of game recommendation questions and the quality of them stays poor, however, we may change our take on this.
Also note the related question How to deal with questions that just don't understand the scope of the RPG landscape? - we get some game-rec (adventure-rec, etc.) questions that try to be "flexible" (e.g. "anything I can conceivably adapt to my goals") and as a result are too broad. When asking your question you should focus it down on the exact thing you want to accomplish - don't worry, people will suggest a bunch of things that don't hit your criteria 100% as it is.