Quote from: x11joe on July 14, 2014, 05:49:37 AM

Quote from: socket on July 14, 2014, 05:23:41 AM

Quote from: x11joe on July 14, 2014, 05:17:01 AM

I thought about ratings, and there is some simple ratings involved, such as number of failed/successful transactions with trustless escrow. The problem with ratings is they can so easily be manipulated. People will just pay for an item to themselves to give themselves a good rating. What use will ratings be then?

I was thinking therefore of leaving the ratings up to the listees in their posts. One could link to a third party trusted site that the community agrees upon in their listing post and people could view the ratings there. If anyone has an idea for a p2p style rating system that can not be abused I'm open. I could not think of anything on the initial concept for OneMarket. The entire team is always open to ideas though.

Did you happen to see my suggestion about coming up with a formula that takes multiple factors into account? I was thinking some combination of listing activity, staking balance, and the wallets days destroyed? This way you get a complete picture of how a lister is operating. It's easy to game a single matrix of activity. But if you add in others it becomes harder to game.

Hi socket, yes I actually did think about this too. I had very long discussions all night about the idea of weighted system with theDagger. It got complicated rather quickly. Just like you said, I thought of using weight based on how many previous transactions that have been completed and money involved and age, all those factors could still be abused via automated accounts which could automate the process of making aged and reputable accounts to sell to fraudsters. Then their becomes a black-market selling highly rated accounts.

This is why I'm avoiding a rating system at the moment, even one based on weighted factors. In order for this to truly work, it needs to be based on the money passing hands in a way that the seller and buyer can truly be verified to be separate individuals that don't know each other. Which is tricky to prove. I've got some ideas for this I might propose later, but this is a tricky problem indeed.

I can see that, and it may be a good idea just to get a working trustless escrow model out, and it may be that OneMarket may find a niche for itself with the sales of items that lend themselves to Nash's game theory model.