Yahoo Answers is pretty interesting and it works on a point based system. Basically, the more questions you answer, the more points you gain. You also get points for "voting" on the best answer to each question (not as many as answering though). If the most people vote for your answer as "best answer" then you get 10 points.
That may have been a confusing explanation, but I definitely like the system (which keeps users engaged). And Hacker News does this as well with its karma system, although it works in a different way. Some of my friends are hooked on to this one chatting website (www.zobe.com), and it has a karma system as well. Basically, all the sites i've seen that have a point based/karma system are quite "vicious" as you put it (they have very consistent users).
I also forgot to mention- I also began recently using Quora, and i've liked it so far! Reminds me of Yahoo Answers but it definitely seems newer and fancier.
I'm interested in hearing more about the project you're working on! Please elaborate on what you're trying to do :)
We see the problem as social media visibility is based on popularity - despite the fact that the popularity doesn't ensure quality. Afterall group think isn't necessarily a good thing...
Emerging social media and news platforms try to identify quality content but their quality metrics are based on popularity.
Thoughtly approaches the problem by accounting for users written contributions online. By quantifying users' personal data - from sites like Stackexchange and Quora we display an individual's involvement in a field. This allows you to objectively compare perspectives from different kinds of people. Think, "actions speak louder than words." Well, we finally for the first time in human history have enough data from disparate sources that we can rely on actionable data and not just resumes and mildly subjective about me sites.
I'd be happy to discuss it further and thanks for your interest. I hope you check us out. We are launching to iOS soon - the app is approved, now we are polishing the backend data sources.