Author [EN] [ZH] [ES] [PT] [IT] [DE] [FR] [NL] [TR] [SR] [AR] [RU] [EN] [ZH] [ES] [PT] [IT] [DE] [FR] [NL] [TR] [SR] [AR] [RU] [EN] [ZH] [ES] [PT] [IT] [DE] [FR] [NL] [TR] [SR] [AR] [RU] Topic: Let's discuss reputation schemes.  (Read 596 times)

0 Members and 1 Guest are viewing this topic.

Offline toast

Let's discuss reputation schemes.
« on: February 07, 2014, 04:52:44 PM »

IMO a robust reputation/trust system will enable a new frontier of possible DACs. There are many proposed approaches to this problem. Let's discuss and collect ideas.

Idea #1
Each person has their own asset in a prediction market over contract fulfillment. When I make an agreement with someone that requires trust, I make them stake their rep on it by exposing it to a prediction market. It pretty much eliminates all counterparty risk if I can write my contract in a way that allows proof-of-breach and then hedge my contract by shorting that person's reputation.

Idea #2
PageRank-esque web-of-trust, with "initial rep" (in pagerank it's the raw text relevance) set by proof-of-burn. Weighting scheme ideally would make any strategy other than burning a few satoshi and then acquiring trust useless, and "reputation pumping" would be difficult for the same reason link-trading is difficult.
Do not use this post as information for making any important decisions. The only agreements I ever make are informal and non-binding. Take the same precautions as when dealing with a compromised account, scammer, sockpuppet, etc.

Offline Brent.Allsop

  • Sr. Member
  • ****
  • Posts: 242
    • View Profile
    • Canonizer.com
Re: Let's discuss reputation schemes.
« Reply #1 on: February 11, 2014, 04:25:28 AM »

Hi Toast,

Reputation is definitely a critically important function.  If you had a good reputation system, suddenly all the spam and scam in the world can easily be ignored.

And, basically, a 'reputation' system, in its most general sense,  is simply a concise and quantitative description of the experience everyone has had with a person, product or service.  And that is exactly what Canonizer.com is all about.

You point someone to your reputation survey topic, before you start the deal.  You can point out that if they are disappointed with your service, they can join a camp describing such.

Then, if they have a bad, or good experience, they can create a camp (or most likely, just join a camp someone else has already started, describing your particular experience) and so on.  No camp can be filtered by anyone, giving a true open and honest measure of experience that everyone can trust.

Imagine having one of these on your e-mail in-box.  You could set your personal canonized reputation score, below which no mail gets into your inbox.  If a particular one does make it through, and you don't like it, you give it a negative reputation, and nobody else will have the same problem.  And if something does make it through, you can trust that it will be worth your time.

Brent Allsop

Offline bytemaster

Re: Let's discuss reputation schemes.
« Reply #2 on: February 11, 2014, 04:29:46 AM »
IMO a robust reputation/trust system will enable a new frontier of possible DACs. There are many proposed approaches to this problem. Let's discuss and collect ideas.

Idea #1
Each person has their own asset in a prediction market over contract fulfillment. When I make an agreement with someone that requires trust, I make them stake their rep on it by exposing it to a prediction market. It pretty much eliminates all counterparty risk if I can write my contract in a way that allows proof-of-breach and then hedge my contract by shorting that person's reputation.

Idea #2
PageRank-esque web-of-trust, with "initial rep" (in pagerank it's the raw text relevance) set by proof-of-burn. Weighting scheme ideally would make any strategy other than burning a few satoshi and then acquiring trust useless, and "reputation pumping" would be difficult for the same reason link-trading is difficult.

Any reputation scheme must operate from the perspective of a user.  There is no such thing as a global reputation, it is all subjective. 

Shorting someones reputation is unlikely to work if their reputation is much greater than yours.  Any complaints you file against them are likely to hurt your reputation more than theirs initially.   Thus as a hedge, it only works if they develop a very bad reputation and that could take a lot of time.

Reputation must be calculated individually between every pair of users.  And there are different kinds of trust.  I have friends I trust, but I wouldn't trust them to make judgments about whom to trust.  Thus a web-of-trust style system could not operate on a single trust metric.

For the latest updates checkout my blog: http://bytemaster.bitshares.org
Anything said on these forums does not constitute an intent to create a legal obligation or contract between myself and anyone else.   These are merely my opinions and I reserve the right to change them at any time.

Offline bytemaster

Re: Let's discuss reputation schemes.
« Reply #3 on: February 11, 2014, 04:31:31 AM »
Ultimately the best reputation system is to post a surety bond with a trusted 3rd party whom you have already consented to resolve disputes.   You can withdraw your surety bond with 60 day notice to give time for last minute complaints.   

If everyone had a surety bond posted on themselves, then you could trust them as long as you trust the arbitration agency. 

It doesn't cost anything unless there is a dispute.

For the latest updates checkout my blog: http://bytemaster.bitshares.org
Anything said on these forums does not constitute an intent to create a legal obligation or contract between myself and anyone else.   These are merely my opinions and I reserve the right to change them at any time.

Offline bytemaster

Re: Let's discuss reputation schemes.
« Reply #4 on: February 11, 2014, 06:07:33 PM »
If each individual has their own 'stock' where the value of the stock is tied to how much someone would pay to have their reputation.   This would probably put pubic figures like Adam B. Levine very highly rated and anonymous every day people with relatively low ratings.

The challenge is that creating a market for every identity can become a lot of overhead to build.   Then again not everyone needs a full up reputation system.   Those whom want the trust of others would have to spend the time and effort to get others to invest in them personally. 

Assuming you had a market that actively traded on your reputation there are two ways this market could be tied to you for contract purposes:

1) It can be entirely independent of you, ie: setting up a market for the trust level in Obama.  I doubt he would own his identity, but others can speculate entirely independent of any action Obama may take.

2) Someone can sell you a collateralized OPTION on their identity where they agree to buy back their reputation stock for a specified price in the future.   

In the case of #2 they have made a bet on their own reputation payable to you.  If they take any action to hurt their reputation then they will lose that bet. 

The challenge is that reputation is often disconnected from trustworthiness, especially when the media and government wage a propaganda war against someone.   A perfectly good, well-intentioned, individual can be accused of being a scammer, and any girl can claim rape and thus destroy your reputation even in the complete absence of evidence.

So you have to ask yourself this... how much would you bet on your own reputation when someone given enough financial incentive (by betting against your reputation) can cause significant short term and even long-term damage through false accusations?   

Then suppose you are doing business with Apple which has a good reputation, you decide to hedge your position 'just in case' by shorting apple.   Apple then screws you on a contract, but everyone sides with Apple because they don't know you.  Are you really able to profit from your hedge enough to cover damages?   

I think that these markets can be a good indicator of the level of risk you take in dealing with someone, but not so good for hedging individual contracts unless they are very public and easily verified by everyone.   

I actually think that someone with a very high reputation would short their own reputation as insurance against attacks on their reputation.   Having a high reputation is very valuable and losing it would be very costly.   However, many people would feel that shorting your own reputation would be manipulating the market for a profit if you intentionally tank your reputation.

I think you should be allowed to short your identity to keep it in line because part of your identity is your reputation not to intentionally tank your ID.


 
For the latest updates checkout my blog: http://bytemaster.bitshares.org
Anything said on these forums does not constitute an intent to create a legal obligation or contract between myself and anyone else.   These are merely my opinions and I reserve the right to change them at any time.

Offline Geneko

  • Full Member
  • ***
  • Posts: 187
    • View Profile
Re: Let's discuss reputation schemes.
« Reply #5 on: February 11, 2014, 11:49:20 PM »
There is one book that could help resolve some reputation/trust issues. It is Steven Coveys -The Speed Of Trust.

http://www.myspeedoftrust.com/How-The-Speed-of-Trust-works/book

There he has established trust matrix with four trust zones. One of them is of something called Smart Trust. He says that the lowest risk and highest return occur in Smart Trust Zone 2 (judgment), where risk is wisely moderated and managed and you can carefully evaluate and consider issues.

So trust is basically about risk management. So how about it?

To be more specific, take a look at Colored Coin smart property or Bitcoin payment protocol.
First is providing ownership transfer of an asset the other value transfer in a way that manages most of potential fraud risks.   

So the question is not why do I have to trust you but could it be done in trustless way.

« Last Edit: February 12, 2014, 03:02:54 AM by Geneko »

Offline luckybit

Re: Let's discuss reputation schemes.
« Reply #6 on: February 12, 2014, 05:17:07 PM »
IMO a robust reputation/trust system will enable a new frontier of possible DACs. There are many proposed approaches to this problem. Let's discuss and collect ideas.

Idea #1
Each person has their own asset in a prediction market over contract fulfillment. When I make an agreement with someone that requires trust, I make them stake their rep on it by exposing it to a prediction market. It pretty much eliminates all counterparty risk if I can write my contract in a way that allows proof-of-breach and then hedge my contract by shorting that person's reputation.

Idea #2
PageRank-esque web-of-trust, with "initial rep" (in pagerank it's the raw text relevance) set by proof-of-burn. Weighting scheme ideally would make any strategy other than burning a few satoshi and then acquiring trust useless, and "reputation pumping" would be difficult for the same reason link-trading is difficult.

Option 3, apply game theory. Make them put some Bitshares as collateral so that if there is a dispute they lose whatever is held in collateral. This way we can quantify trust in accordance to how much they hold in collateral.

This is basically like a security deposit. If things go just fine then they'll keep whatever is in collateral but if they try to scam or if they don't keep their end of the deal then we redeem for the Bitshares in collateral.

No need for over complicated prediction markets. If you want me to trust you then you should take either an equal amount of risk or greater risk. If I lose you lose with me and that is mutually assured destruction.
https://metaexchange.info | Bitcoin<->Altcoin exchange | Instant | Safe | Low spreads

Offline luckybit

Re: Let's discuss reputation schemes.
« Reply #7 on: February 12, 2014, 05:21:06 PM »
If each individual has their own 'stock' where the value of the stock is tied to how much someone would pay to have their reputation.   This would probably put pubic figures like Adam B. Levine very highly rated and anonymous every day people with relatively low ratings.

The challenge is that creating a market for every identity can become a lot of overhead to build.   Then again not everyone needs a full up reputation system.   Those whom want the trust of others would have to spend the time and effort to get others to invest in them personally. 

Assuming you had a market that actively traded on your reputation there are two ways this market could be tied to you for contract purposes:

1) It can be entirely independent of you, ie: setting up a market for the trust level in Obama.  I doubt he would own his identity, but others can speculate entirely independent of any action Obama may take.

2) Someone can sell you a collateralized OPTION on their identity where they agree to buy back their reputation stock for a specified price in the future.   

In the case of #2 they have made a bet on their own reputation payable to you.  If they take any action to hurt their reputation then they will lose that bet. 

The challenge is that reputation is often disconnected from trustworthiness, especially when the media and government wage a propaganda war against someone.   A perfectly good, well-intentioned, individual can be accused of being a scammer, and any girl can claim rape and thus destroy your reputation even in the complete absence of evidence.

So you have to ask yourself this... how much would you bet on your own reputation when someone given enough financial incentive (by betting against your reputation) can cause significant short term and even long-term damage through false accusations?   

Then suppose you are doing business with Apple which has a good reputation, you decide to hedge your position 'just in case' by shorting apple.   Apple then screws you on a contract, but everyone sides with Apple because they don't know you.  Are you really able to profit from your hedge enough to cover damages?   

I think that these markets can be a good indicator of the level of risk you take in dealing with someone, but not so good for hedging individual contracts unless they are very public and easily verified by everyone.   

I actually think that someone with a very high reputation would short their own reputation as insurance against attacks on their reputation.   Having a high reputation is very valuable and losing it would be very costly.   However, many people would feel that shorting your own reputation would be manipulating the market for a profit if you intentionally tank your reputation.

I think you should be allowed to short your identity to keep it in line because part of your identity is your reputation not to intentionally tank your ID.


 

Identity is important too. But once again if someone issues a currency and they promise to redeem it because it's backed by their goods and services they should put some Bitshares in collateral so that people can get their money back if they don't redeem or if there is a dispute. Then I would trust them.

https://metaexchange.info | Bitcoin<->Altcoin exchange | Instant | Safe | Low spreads

Offline fuzzy

Re: Let's discuss reputation schemes.
« Reply #8 on: February 12, 2014, 10:46:17 PM »
IMO a robust reputation/trust system will enable a new frontier of possible DACs. There are many proposed approaches to this problem. Let's discuss and collect ideas.

Idea #1
Each person has their own asset in a prediction market over contract fulfillment. When I make an agreement with someone that requires trust, I make them stake their rep on it by exposing it to a prediction market. It pretty much eliminates all counterparty risk if I can write my contract in a way that allows proof-of-breach and then hedge my contract by shorting that person's reputation.

Idea #2
PageRank-esque web-of-trust, with "initial rep" (in pagerank it's the raw text relevance) set by proof-of-burn. Weighting scheme ideally would make any strategy other than burning a few satoshi and then acquiring trust useless, and "reputation pumping" would be difficult for the same reason link-trading is difficult.

Option 3, apply game theory. Make them put some Bitshares as collateral so that if there is a dispute they lose whatever is held in collateral. This way we can quantify trust in accordance to how much they hold in collateral.

This is basically like a security deposit. If things go just fine then they'll keep whatever is in collateral but if they try to scam or if they don't keep their end of the deal then we redeem for the Bitshares in collateral.

No need for over complicated prediction markets. If you want me to trust you then you should take either an equal amount of risk or greater risk. If I lose you lose with me and that is mutually assured destruction.

This was precisely the reason I ended up feeling so good about Invictus.  They took on more risk than any of us in their no-premine launch of PTS...
WhaleShares==DKP; BitShares is our Community! 
ShareBits and WhaleShares = Love :D

Offline luckybit

Re: Let's discuss reputation schemes.
« Reply #9 on: February 13, 2014, 02:17:11 PM »
IMO a robust reputation/trust system will enable a new frontier of possible DACs. There are many proposed approaches to this problem. Let's discuss and collect ideas.

Idea #1
Each person has their own asset in a prediction market over contract fulfillment. When I make an agreement with someone that requires trust, I make them stake their rep on it by exposing it to a prediction market. It pretty much eliminates all counterparty risk if I can write my contract in a way that allows proof-of-breach and then hedge my contract by shorting that person's reputation.

Idea #2
PageRank-esque web-of-trust, with "initial rep" (in pagerank it's the raw text relevance) set by proof-of-burn. Weighting scheme ideally would make any strategy other than burning a few satoshi and then acquiring trust useless, and "reputation pumping" would be difficult for the same reason link-trading is difficult.

Option 3, apply game theory. Make them put some Bitshares as collateral so that if there is a dispute they lose whatever is held in collateral. This way we can quantify trust in accordance to how much they hold in collateral.

This is basically like a security deposit. If things go just fine then they'll keep whatever is in collateral but if they try to scam or if they don't keep their end of the deal then we redeem for the Bitshares in collateral.

No need for over complicated prediction markets. If you want me to trust you then you should take either an equal amount of risk or greater risk. If I lose you lose with me and that is mutually assured destruction.

This was precisely the reason I ended up feeling so good about Invictus.  They took on more risk than any of us in their no-premine launch of PTS...

Invictus isn't the issuer I was talking about though. Invictus is a group of people we know are a real company with their real identities out in public. They are putting themselves on the line financially and putting their real life reputations at stake so we don't have to worry about them as much.

We have to worry about the shady sort of Bitcoin businesses we usually end up interacting with which turn out to be complete scams.
https://metaexchange.info | Bitcoin<->Altcoin exchange | Instant | Safe | Low spreads

Offline fuzzy

Re: Let's discuss reputation schemes.
« Reply #10 on: February 13, 2014, 03:20:19 PM »
and also keep our eyes open to the ones that pop up in this ecosystem as it grows... :o

Reputation and BitShares as collateral would be interesting as well. But then again, we have a great many unknowns in this space...especially with government and, more importantly, the consensus of the general public about "BitCoin". 

How do you guys suppose navigating these two landscapes will look?
« Last Edit: February 13, 2014, 03:40:11 PM by fuznuts »
WhaleShares==DKP; BitShares is our Community! 
ShareBits and WhaleShares = Love :D

 

Google+