4 ms·
>I can definitely see where you are coming from since that mechanism of "most token wins", sounds very unintuitive and inherintly cheatable (and in a lot of pla
by Sangermaine 8y ago
>I can definitely see where you are coming from since that mechanism of "most token wins", sounds very unintuitive and inherintly cheatable (and in a lot of places has rightfully bad connotations like "pay to win" in video games).
It does indeed seem exactly that way. You said upthread:
>An additional actor X who knows about the ownership can add additional weight to the claim of Person A.
But how do you verify that X is someone "knows about the ownership"? What if X was someone A or B paid or incentivized to vouch for them so the consensus would be swayed? How do prevent bad actors from gaming the system in this way? If the asset in dispute is valuable enough it would certainly be worth it to a bad actor to pay a little to cheat and have the consensus fall on their side.
- hobofan 8y ago> What if X was someone A or B paid or incentivized to vouch for them so the consensus would be swayed? If B were to pay X a bribe of amount N, so X enforces the statement with amount N, B could also directly enforce the statement with amount N, which would lead to exactly the same result. If X would receive less than N, to add a weight of N, they would be gifting some money to B, and vice versa. So between rational untrusted parties, a simple bribe would not make sense. There are some rationality assumptions relating to 51% attacks similar to PoS oriented systems that have to be met, but there are some mechanisms in place that should reduce the practical requirement regarding rationality among participating parties significantly. > If the asset in dispute is valuable enough it would certainly be worth it to a bad actor to pay a little to cheat and have the consensus fall on their side. Good question, but I'm afraid that's one of the questions I can't answer right now without leaking too much information.
- a1369209993 8y ago> If B were to pay X a bribe of amount N, [...] B could also directly enforce the statement with amount N Okay, better question: how do you deal with the (pathologic, and therefore approximately guarranteed to come up) situation where A's life savings are ten dollars and a horse, B has 100 million dollars, and B claims (with amount ≥ 11$) that B owns A's horse.
- hobofan 8y agoAs long as the "100 million dollars" don't make up more than 50% of the tokens in the network, the incentives are supposed to play out in a way that all the other participants of the network contest the wrong statement made by B, and thus B isn't able to introduce information that is deemed wrong. So in general the situation should be preventable. There are some specific versions of that edge case that come to my mind that where that situation could indeed be unresolvable but in those cases that should be detectable and displayable to the users of the system. As I said in my previous comment, there are some rationality requirements regarding some of the token amount.
- admax88q 8y ago> If B were to pay X a bribe of amount N, so X enforces the statement with amount N, B could also directly enforce the statement with amount N You're assuming the bribe comes in your made up currency. B could bribe X with real money or out of band coerce/intimidate X with threats and violence.
- hobofan 8y ago> B could bribe X with real money So a connected market with goods that can be bridged similar to the stock market. > or out of band coerce/intimidate X with threats and violence I'm not pushing for a world without governments and law enforcement.