As far I researched meta, I cannot change my vote for a answer after 5 minutes, only when the answer is edited, despite how critic and brilliant the comments on this answer or another 1-2 days later given answer is. While this probably is no bigger problem on trilogy sites with so many experienced voters, on scientific betas with much less voters, much less experts and lower A/Q ratio it imho yields some very counter-productive disadvantages. Laymen will mostly just upvote a answer if it sounds right and logical to them, someone put much effort in a long nice edited post, and has a high rep and seems trustful. They might even upvote the better answer if they are completely unbiased by other upvotes. But as far I see on early betas fast and in some degree good answers get the most upvotes. Experts not related to programming dont have time to watch a site every hour like full-time programmers sitting 8 hours in front of their PC. Very often the same user give fast mediocre answers at first and gain much rep this way, obviously a win-win strategy if you like rep hunting. But in a beta phase trying to attract experts not familiar with SE and seeing mainly mediocre and fast answers succeeding and getting accepted will not really motivate experts to stay and improve quality of the site.
I'm not sure if the option to cancel your upvotes is the best improvment to change this phenomenon. You could also say:
Just wait, gather more users, create more competition, get a higher average vote/answer, higher A/Q ratio, the prob with this is, if the quality is not high on scientific betas from the very beginning, you will also in future not really attract more experts to a site with similar quality as digg or yahooanswers. So a solution would be that scientific betas need a much higher commitment number than e.g. culture proposals. The community dynamics and quality of Q&A between something like philosophy and christianity are simple uncomparable. And if the voting system is not wrong, than it doesnt work correctly as to less user. I dont see another explanation.
If you sort a scientific beta by high rep users, click on individual users, you will see very often (apart from moderators) many ask no or very very few questions but having 20-40 answers. Im not sure what I should think from such users, are they rep addicted, know just everything, frequent a Q&A site not having any questions and just enlighten everybody with their knowledge? A good scientist or interested guy always has more questions than answers, otherwise there is no progress. So can there something implemented like you can only answer 5 new questions after asking 1 question yourself If someone has no questions, I doubt he is able to rethink his answers considering comments/critics and actually think asking questions shows ignorance/imperfectness. There are currently scientific betas having 2 questions a day despite they are the most tricky fields in modern science.
Can there anything been done to improve community dynamics and expert attraction for scientific sites in the early beta phase (first 200 days), or will the Area 51 proposal process always work this way for the next months?