The other problem with a group karma system is that coders will obsess about it and spend their karma distribution time not writing code and not really getting

The other problem with a group karma system is that coders will obsess about it and spend their karma distribution time not writing code and not really getting better as a result of doling out positive karma. And then Sr. Architect X announces hiring of a replacement in the next 6 months and everyone spends 2 hours a day up-modding his/her old code, etc. Oh and what if I think that nested if statements are easy to follow so I go around dinging you for extracting to methods or using ? : form or something. Perhaps if you had a really bad-ass coder whose job was to apply karma to code from either lower-tier or less experienced for part of the day, that might work. It provides incentive, it comes from a source of understood quality and authority and it can be accompanied by constructive criticism or substitute code to study. A system like this would have to have a clearly defined goal of rewarding good programming according to some metric or standard. It's a good germ of an idea, but not a good idea as stated I think.