-
Notifications
You must be signed in to change notification settings - Fork 451
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Whitewashing-proof reputation system #4015
Comments
yes, I recommended a solution here; |
Alright, I'll close this. But it seems like a quite fundamental issue. Someone who has only leeched for a week should not get preference over a new user, and in an anonymous system it is inherently impossible to have a lasting reputation. |
I'd leave this issue open for now. |
Allright then. Let me layout the incentives as I understand them:
It seems that the game theoretic conclusion then becomes that a negative balance should be treated in the same way as a balance of 0, which means that new users are penalized as much as leech-only network nodes. It is possible to favor users with a positive balance, but since network capacity is limited, this is equivalent to a penalty for others. This issue is due to the possibility of a negative balance. Usually a monetary system only has debt with the understanding that it should be paid back, which requires accountability. |
Very good question. The attack you are referring to is called the whitewashing attack since it allows users to 'whitewash' their accumulated reputation and re-join the system under a new identity. This is indeed an open challenge. One approach to fix this is by requiring validated identities (i.e. by linking some real-world personal data like a bank account). Our lab is working on a decentralized identity system (see Tribler/py-ipv8#328) where users are verified by a governmental stamp (attestation) but we doubt that it is a good idea to integrate this into Tribler. Since more and more users are joining the Tribler network (as indicated by the growth of our blockchain and results from the network crawler), we have become a bit conservative and focus on stability first. Also, @Captain-Coder is working on the problem to detect free-riding behavior (see also #2571). Before we decide on a policy, we should first investigate available TrustChain records and see how a different policy effects usability. The solution proposed by @Dmole:
I think this addresses the issue and we will keep this in mind when doing more work on the whitewashing attack. Presenting this kind of information to the user is non-trivial. |
Thank you for this.
It addresses the issue in the sense that it dis-incentivizes whitewashing. It doesn't fix the issue that new users are on the bottom of the reputation ladder, together with or even below the worst freeriders. |
ThomasdenH, "Together but not below"; Anyway It's not a pressing issue for now as points are not worth anything yet (AFAIK) . The ~48 issues tagged as bug or worse should take priority. I do hope that before points become worth anything this issue is addressed and the points are reset for all users. |
Yes, we will definitely address this issue at one point, but not for the 7.2 release. As indicated by the amount of issues left for this milestone, we have more than enough work to do. I will postpone this issue. |
v7.8.0 "Showed token balance again" without fixing this. |
It looks like the question of OP has been answered years ago and the remaining discussion is about the token component that was also removed years ago. I'll close this issue now. |
I have a question about the Turstchain balance. Isn't it trivial to reset any negative balance to 0 by deleting local Tribler data? Doesn't that mean that bad actors employing this technique can only be punished as severely as new users?
The text was updated successfully, but these errors were encountered: