This is a companion piece to my main article about the Gnosis ICO and why I won’t be participating.
The Gnosis Whitepaper describes 3 different types of oracles (and 1 hybrid). Let me go through each unique type:
- On-chain oracles- irrelevant for the vast majority of markets, not going to discuss further here.
- Centralized oracles. It appears that Gnosis plans on resolving most markets with a centralized oracle, which to some extent makes it pointless to run on a decentralized platform. While I don’t disagree that centralized oracles would be simpler and likely faster, the point is that a centralized oracle is a point of failure for the system. Whether by regulatory pressure, outright prosecution, or corruption/fraud, relying on a centralized oracle goes against the spirit and idea of what a “Dapp” should be. You take out that “D,” and you may as well just use Betfair, Fairlay, or any other number of options. There’s no reason to use a blockchain or the Ethereum network if the most critical part of the system is going to be centralized.
- Decentralized oracle- what they are calling “The Ultimate Oracle.” From an overall architectural standpoint, in my opinion this is Gnosis’ greatest weakness (with the convoluted token system taking a close second). See the below diagram from their whitepaper for their concept:
So the basic gist is that if you believe an outcome has been reported incorrectly, you can challenge that outcome by posting 100 ETH (within 12 hours of the oracle publishing the outcome), which then triggers “Ether voting,” where anyone can vote on what the correct outcome was by essentially betting ETH on that outcome. After one outcome or another has more ETH voting for it than any other outcome for a full 24 hours, then that outcome wins. Whenever the lead changes, a new 24 hour clock is set. This repeats until an outcome has a full 24 hours of being the leader. At that point, those who bet for the most popular outcome will get to keep the ETH that others bet for any other outcome.
They’ve touted this system as being revolutionary in that “Every ETH holder is the backstop!” Which is theoretically true. But they’re ignoring the fact that:
a)Most likely 99.99% of ETH holders don’t give a fuck about the dispute you’re having about some random market on the Gnosis platform, and
b) For the people who are involved, suddenly in order to defend their bet against an incorrect or fraudulent report by the oracle, they will effectively have to up their bet- by a LOT. Which will put them in a very uncomfortable position.
Let me walk you through one of many scenarios in which this oracle could fail.
Say there is a market open regarding a small town political election- Say you live in Hicksville Alabama, and Candidate A is running against Candidate B. You happen to know that Candidate A is a complete scumbag who murdered the relatives of half the people who live in Hicksville, but got off on some technicality. It is extremely unlikely that Candidate A will be elected, and the odds on Gnosis reflect that- Shares for Candidate A are trading at only the equivalent of 10 cents a share.
Knowing that it is extremely likely that Candidate B will win, you go ahead and buy 1 Eth worth of Candidate B shares at the equivalent of 90 cents/share, thinking you’ve found an easy way to make an 11% profit. The election comes and goes, and lo and behold, Candidate B wins, as you expected. Hurray!
But then, before the market “resolves,” someone challenges the outcome. They go ahead and put up the 100 ETH to challenge the outcome.
So now what? You only wanted to make a 1 Eth bet on the mayoral race. Do you really want to now throw in 101x your initial bet in order to defend your position and “Ether vote” for candidate B? Do you really want to trust that “Every ETH Holder” will come to your rescue? Of course you wouldn’t expect that “Every” ETH holder would get involved here, but would you really expect that any ETH holder would?
From the standpoint of other random ETH holders out there, all they would see is that someone else was willing to put up a 100 ETH bond to challenge the outcome. Despite the fact that there would probably be election results available online or even by calling the town, all of which would show that Candidate B actually won, other ETH holders would have to wonder if there was some part of the story they were missing. Did Candidate B die, or commit some crime right before the election that had disqualified them? Is there some such information that the challenger knew about but no one else voting did? There could be any number of possible scenarios that would give other ETH holders enough reason to doubt that they wouldn’t want to put their own funds on the line in order to make this market resolve correctly, on a market that they didn’t care about to begin with.
And even if YOU decided to make a stand and scrape together 101 Eth in order to “outvote” the challenger, then what? What if they then drop another 100 ETH in there to outvote you back, by a long shot? You gonna scrape together another 100 ETH, bringing your total commitment to 201 ETH?
In reality, the “ultimate oracle” would become the “ultimate bullying platform.” Essentially, the truth wouldn’t matter at all in resolving these markets. All that would matter is who has the biggest pile of ETH with which to challenge and vote with.
And really, from a game theory standpoint, it would be a brilliant thing to do. Find some low-liquidity, low profile markets with longshot odds, and buy cheap shares of the underdog. If the underdog wins, great! If the underdog loses, challenge the outcome with enough ETH to make it uncomfortable and difficult for others to fight you.
And the problem remains the same in a converse way- Say you had 1 Eth bet on a random market, and then after the event is over, the oracle actually does report the outcome incorrectly. Do really want to commit a minimum of 100 ETH in order to open the challenge? Knowing that some other shareholder with deeper pockets could just keep outvoting you until you ran dry?
It’s completely irrational to expect other Eth holders to become involved in disputed markets on a prediction market.
In summary, Gnosis’ planned oracle systems are one of its greatest weaknesses. If they use a centralized oracle, a decentralized platform is pointless. And if they use their proposed decentralized oracle, their system will be rife with scams and bullying.