Or, perhaps you could define 'merit' as continously compounding interest with their average karma density (votes/items). Taking the interest forumula "Pe^(tr)"

Or, perhaps you could define 'merit' as continously compounding interest with their average karma density (votes/items). Taking the interest forumula "Pe^(tr)", and substituting might yield something like: P = 1, -1, it is the single up/down vote t = user life span in years r = karma/submissions (karma_density) for (user:item.voters) item.merit += (sign_of_vote)e^((user.life_span/12)(user.karma/user.submissions))