What you are pointing out is why I feel that Reputation really can't be determined by the opinions of anonymous others. Our assignment of reputation to an account is based on our personal values, and these vary from user to user, as you point out.
Assigning 1 vote to 1 account is actually assigning 1 vote to one bot, not to one person. Some Hive users have in excess of 10k accounts. Absent some application of criteria to sort which accounts are the only accounts some user uses isn't ever going to work, because it's not how users use accounts. Unless users can be convinced to claim a certain account is their 'human' account, sorting which of their accounts are bots and which are human isn't going to work.
Good luck figuring out how to achieve your goals. Sorting that dilemma is beyond my abilities, and I will spend my efforts on things I think I have a chance of achieving. However, it may well be possible to assign a reputation score to accounts based on their history and how you feel about that. It won't be sorting people, because some people run multiple accounts, but they tend to run each of those accounts in a certain way, and that may enable you to assign a reputation score to those accounts that is meaningful to you.
If I had a one vote per account ranking configured then it would eliminate accounts doing an inhuman amount of work or other indicators of automation, like instant votes and such. Since that would be my preference, it might not be worthwhile for somebody to set up to game it. Nothing is perfect, but we don't have to make it easy for people to fool us.
Dan from Blocktrades is working on a system to allow people to manage a network of trust which will also contain evidence of people's knowledge too - not a social credit system, but kind of an equivalent to that for your own uses. It allows you to track who you know, what you trust about them and why you trust that about them too. It should be possible to integrate it into a variety of Hive tools, but we don't have a huge amount of info on it atm.