I think there are few issues at play here.
For one, the answer rate (the 82% you're quoting is the answer rate, not the acceptance rate; the percentage of questions that have at least one upvoted answer) will be a little low in the private beta, as lots of people ask questions on a variety of topics to probe what is on and off topic, and there is a somewhat more limited userbase who might not have the specific expertise in question. I suspect that will rise a bit as we close off-topic or bad questions, and as we go into the public beta and can widen our userbase a bit.
Now, the acceptance rate does seem to be a bit low, too. I've answered a bunch of questions, and none of those questions have had any answers accepted. Likewise, I've asked a few questions, and haven't accepted any answers on them. In some cases, we might need to just gently remind people to accept answers; on SO, I will frequently remind users that they should probably accept an answer that they found useful (I frequently see people post a comment saying "that's it, thanks!" without checking the checkmark). So, on questions that are clearly asking for a single answer, and in which there is at least one reasonable looking, upvoted answer, we should probably encourage users to accept those answers. Users will probably be reminded a bit more once we've had enough time for people's personal acceptance rates to start showing up; then people will see that they have acceptance rates of 0% or 10%, and be motivated to accept more answers.
One issue is that many questions on this SE are asking not for a single answer, but a list of answers. I've addressed these types of questions in another thread on meta; so far, it seems that people think these kinds of questions are OK, but not many people have chimed in yet. This type of question can't ever really have a single accepted answer; if there are 10 or 20 games which meet the criteria, then there's no reason to accept one over the other, unless you happen to have a very specific situation in mind (the very next game you buy, or the very next game you play with your niece).
A possible solution would be to make these sorts of questions community wiki, as I don't think community wiki questions are expected to necessarily have an accepted answer; but that's recently been disabled, except for moderators, and we don't have moderators yet.
tl;dr Go out and encourage people to accept more answers, and we need to figure out what to do with all of the questions for which a whole list of answers are a good response.
Here's a template for prodding people:
You might want to accept an answer for this question, if there's an answer that meets your needs. If not, I'd recommend asking for clarification in comments, or editing your question to be more specific, so people can edit their answers or write new ones that will better answer your question.
edit to add Oh! And another possible reason we're seeing this problem here is that for game recommendation questions, you can't just test out an answer to see if it works for you, as you can for a programming question on StackOverflow. Most of the time, to see if a game is good, you need to actually find some other players, or wait for your regular game night, convince them to play it, and then do so. It might be pretty hard to figure out in a timely manner if a game recommendation works for you.
I suspect, though, that many of the current recommendation threads are not because people are actually looking for a game to play in a particular situation that is likely to come up soon, but are intended as "example" questions to help start the community going, for which it's even harder to actually test and decide on which answer is best for the person asking the question.