Topic: Idea: Bail-based Web of Trust (Read 555 times)

It depends on how large the community is. If it still is small then 100 BTC could go a long way in controlling it. On the other hand if it is still small, then it is not that an attractive target for scammers yet.

Let's say the minimum bail for each account is 1 BTC and let's say the minimum number of votes to successfully vote out a scammer would be 100 'yes' votes more than any 'no' votes.

Higher amounts of bail for a single account would not give you more voting power. It is still only one vote per account regardless of bail size (as long as it's the minimum). Only if you split up the 100 BTC into 100 separate "sockpuppet" accounts, would you have enough voting power to do this. You would also have to create enough activity for the sockpuppet accounts so that people do not recognize them as such.

For example, the community is 1000 accounts.  If approximately 15% of the community participate, so 150 people vote 'yes' (for forfeiting the bail) then this person would need 51 BTC in 51 accounts to acheive this. If the community is twice the size thus 300 'yes' votes then this person would need 201 BTC in 201 accounts to stop it. So it can get very hard the larger the community is.

It seems unlikely to me that this would be an attractive strategy for a scammer.

Maybe it would be possible to test this in some small scale to see if it actually works.

Firstly , this means that the total amount of BTC I have would be directly proportional to the power a person has.

If I have 100 BTC, I could easily stop any scammer in their tracks, most probably with multiple accounts.
But if I have 1 BTC (or lesser) , I would have no power.

As far as I understand , this doesn't only guarantee that the person which has more BTC locked is likely to lose if he scams , it also gives the person more power.

So , if I invest 100 BTC , I could destroy all the wealth of anyone I chose.

I was thinking of a way to solve the problems of trust in online identities and trolling which Gavin mentioned in his recent speech. The problem is that people can build up an anonymous reputation online and then "cash it out" by betrayal. This is especially a problem for decentralized marketplaces.

The basic idea is that people must post a certain amount of bail (in digital currency) to join a marketplace or community which they risk losing if they misuse the system. So each online identity (which can be completely anonymous) in this system is connected to a certain amount of Bitcoin, for example, which the holder of the identity must pay and is locked up for a certain amount of time (the "bail").


For example, Bob creates an identity Satoshi in this system and pays 1 BTC into a time-locked transaction connected to this identity for one year. This means for one year Bob will not have access to this 1 BTC.

Then Bob interacts on the system. If he uses it to scam people, then people can flag him as scammer. If enough people flag him as scammer then a vote is triggered. If a sufficient number of people in the system agree with the accusation then Bob loses his bail. The funds are either destroyed or sent to charity.

All members of the system can see Bob's bail. People would be advised not to trust him in transactions higher than his 1 BTC. Therefore Bob can choose to make additional payments for a higher bail total.

Once the year is ending Bob can pay another bail to extend his trust in the system. If his year runs out without being renewed then his identity would become inactive on the system until the bail is renewed.

If Bob wants his money back he just waits until the year is over and it is automatically transferred to an account he specified when he created the identity.


Note that each identity in the system would have to pay a minimum bail with a minimum locking period to join. These limits would need to be sufficiently high to discourage spamming of accounts.

This could also be used against trolls and sockpuppet accounts. If a sufficient number of people agree in a community that an account is just trolling or is a sockpuppet then they can vote out those accounts and those accounts would lose their bail.

Also accounts that flag others as scammers, trolls or sockpuppets and so trigger many votes which do not succeed, would lose voting power in future votes. This is to discourage indiscriminate accusations.


What do you think of this idea?

Do you see any way to game this?

Do you have any suggestions for improvements?