As the site recently started, there are interesting questions which will eventually have a lot of answers, and some of them could be correctly answered before the beta goes open. I think that many questions with accepted answers could generate a lack of "competition" to make new answers and this can affect negatively to the site (without considering new questions). Obviously, this is only an opinion and read other viewpoints on this would be nice.
It's a very good question! I originally accepted the answer of Ihosvany Camps on my first question, and then shortly after, there was a second answer by Geoff Hutchison that was equally good.
Since it is the first week of our site, the traffic is going to be a bit unusual, as people will see the invitation email at different times of the day (or they are in different time zones).
I decided to un-accept the answer so that more people feel "encouraged" to think of an alternative answer. I still feel it is good to accept answers though, as it rewards the new users for their contribution.
Therefore on a case-by-case basis, whether or not to accept an answer is always up to the original asker, but what I will do is this:
Accept in cases where the answer is a "done deal". For example if I'm looking for something specific, like a particular paper or software, and I find it. In this case, there is not likely to be any benefit from "keeping the competition open" anyway.
Wait a bit in cases where more users (who might not yet have received the email or found out about the site!) might be able to contribute something. This makes it fair to those who come late.
I'd say don't worry too much about it. Right now, there are few questions and the (really) committed of us will have a general overview about most, if not all, of them. Answers will come as the experts will also arrive at the scene and most of the newer users will probably browse through the questions and add to them. It surely is good practise to accept answers, however, I have not had the experience where it would 'keep the competition open' or deter other users from adding an answer. It basically boils down to a matter of taste. In most of the cases, the OP, has to answer the question to themselves, whether or not the posted answer sufficiently answered the question/ solved the problem.
At the current point, we should not worry what will come after the private beta, but what happens right now. It is more important to focus on scope and tags, then (later) on accepts and accept rates. A question in terms of SE is answered, if it has at least one answer with a positive score. The rest is visuals.