Vitalik made a few proposals for things that test the efficiency of market. Of course we would love to do those markets on Gnosis.
- Launch a prediction market on "what will the output of this smart contract be on Mar 30, 2016?". The contract will be simply and verifiably coded to output 5. Launch a separate smart contract where some specific set of users is rewarded if the average market price exceeds 5; the more it exceeds 5 the higher their reward. This encourages them to try to manipulate the market upwards. See how well the market manages to keep the price close to 5.
This is very easy to do with Gnosis. We can create events and plug in any resolver contract and of course it is not hard to create a contract that returns 5. The second market is slightly harder because we need the average price of market one and this is not directly doable. However - we can modify our market contract so every trade on it will also update this average number.
My very concrete proposal of doing it is this: Issue x (10,000?) shares and sell them in a auction. This x shares will have the value of x ETH if the market would comply fail: the forecast of market nr. one would be 10 for a whole week (x blocks). They have the value 0 if the forecast is constantly 5. Everything in between is payed out proportional.
Now there are two ways to look at it: one - this is a prisoners dilemna/ tragedy of the commons game where if all participants would cooperate they could all share the 10,000 ETH. I don't expect that there is any cooperative solution to this game. However - Vitalik had a different perspective in mind: can one (evil) actor manipulate a price in a free/open market. I think this setup can be a very good measurement of the state of Ethereum and the quality of the marketplace.
It will boil down to: how cheap are transaction and can someone censor transactions. So one obvious approach the attacker could do, if the stakes are high enough, would be to flood the network with high paying transactions or find other ways to bribe miners not to include transactions that should be included otherwise.
My assumption is that today you can manipulate market nr. one at least a little bit if the incentive is high enough. One thing that would make the market much more stable if it would not be to integrate LMSR with an order book. If people could out up big orders at around 5 this would increase costs of manipulating tremendously.
One more market experiment I would like to propose is wether or not a lot of small shareholders can stand up against the interest of a single big shareholder.
So the market would be like this: you distribute 10,000 tokens widely and each of them will have a value of one unit if decision A is made. If decision B is made a single person/ other token would get 5,000 units. To make it harder the 10,000 can not be moved until the decision of A or B is made. This decision is made depending on what site pays more.
The other token that pays 5k in decision B can be auctioned of beforehand.
So bottom line: especially if the Foundation to fund this markets we can set them up really soon and would love to do so!