UMA 2.0 is coming
2023 will see sweeping upgrades to the UMA ecosystem. The underlying theories of UMA have proven themselves, and over the coming year the Risk Labs foundation supporting UMA will propose to transform the oracle so that it is more robust, has a stronger voter base, and a more flexible security model that allows for securing more value–all with an aim to supporting the next generation of real-world applications.
The Original V1sion
Four years ago, the Risk Labs foundation started out on a path to design synthetic derivatives, but we were dissatisfied with the blockchain data provisioning solutions that we had to choose from.
So we decided to build that instead.
This led us to Vitalik Buterin, who had already done important research and thought leadership to fix the oracle problem (among others such as Paul Sztorc of Truthcoin, Augor, and Kleros). At that time, no one had yet taken on his challenge to build an oracle capable of using a game theoretical concept pioneered in the 1960 for war games: Schelling-point coordination theory.
This theory states that if you cannot coordinate with other players, but share a similar instruction, you can still expect to gather around a short list of likely outcomes. The classic example is if you’re asked to “Meet someone at noon tomorrow in New York,” you’re likely to end up at one of the obvious landmarks, and have a much higher-than-random chance to meet.
Vitalik realized that we could use this coordination scheme not just as a means to meet each other in a location, but to meet at truths on-chain as well.
UMA has put this thesis to the test. Did it work?
UMA’s Path Towards the Optimistic Oracle
For the first test of Vitalik’s design, UMA supported a long list of tradable synthetic assets. Here are some of the more interesting highlights:
- A synthetic Ethereum gas derivative
- uStonks, an index based on the Wall Street Bets community
- Decentralized ETHBTC
- SpaceX rocket insurance
- Event-based incentive tokens
Following these, UMA pioneered the term “Optimistic Oracle,” which refers to an oracle that validates data provisions optimistically, and only requires the Schelling point vote in cases of a dispute instead of for every data request. You could retroactively think of this as UMA 1.1, as it opened the door to powerful protocol integrations such as:
- Smart contract insurance
- Prediction markets
- Cross-chain bridges
- DAO-incentive platforms
Since launch, UMA’s oracle has secured billions of dollars worth of markets without exploit.
As usual, it seemed that Vitalik was onto something.
UMA 2.0
Having proven this thesis, Risk Labs foundation is proud to announce the multi-phased release of UMA 2.0 over 2023, the first to be complete in *January. (Risk Labs is the foundation and core team behind UMA and Across Protocol.)
*See Editor’s Note at the bottom of the article
These changes together represent the maturing of the oracle–With a robust voting system and increased participation, simplified ways for dapps to plug into the oracle, and further research into the security model of the Data Verification Mechanism (DVM).
Here are 3 milestones to watch for in the 2023 UMA 2.0 roadmap:
1. UMA Token Staking
UMA is a work token.
The work is to secure the oracle. Over 99.9% of this work is done in the background, thanks to the design of the optimistic system. This system only works if voters show up for the .01% case (the pessimistic case), which is when there is a dispute that needs review.
This is the time when the DVM needs tokenholders to do the work to resolve what the truth is: the OO needs you to care when there is a vote, and put in some thought to get the answer correct. Previously, this was accomplished through discrete inflation payouts each vote. In UMA 2.0, rewards will be streamed continuously and will begin just as soon as you stake $UMA.
The protocol’s reward emissions rate is expressed as $UMA/second, with the initial amount at .18 $UMA–Adjustable by governance vote. This will be shared among all staked $UMA, so the APY will be variable. If participation matched the voting rates we see today, the reward rate would be right at 30% annualized.
Rewarding good work
While tokens are staked, voters will also experience reward reallocation depending on if they vote consistently and accurately or not. Each time there is a vote, a small fraction of tokens will be redistributed from the stakes of unsuccessful voters to successful voters. These reallocations are discrete, per-vote events. This design ensures that there is sufficient motivation to vote accurately, i.e., to do your work well.
This change also introduces an unstaking cooldown period, set at 7 days to start. If you want to go from voting to unstaking your $UMA, you’ll need to wait the duration of that cooldown.
This will impact the circulating token ecosystem, as it forces voters to commit to their role as a voter and take on the “risk” of being unable to sell in a flash. More importantly, the new design also reduces the attack surface of the oracle by making it impossible for voters to both be bribed and cash out their holdings before the impact of the bribe is felt in the UMA token price.
2. Simplified Oracle Interactions
Also in January, Risk Labs will roll out a new website with developer tutorials and new voter dapp with more tools for voters and a token staking dashboard.
On the protocol level, oracle interactions will be simplified by changing the request pattern to one of optimistic assertions. Instead of asking a question, a protocol or person will submit a statement or piece of code “as true.” So a request is also a proposal, to use the current language.
During the dispute window, anyone can dispute and claim that the assertion is actually false.
3. Security Model
Until now, UMA’s security model has been based on maintaining the total value secured under half the marketcap of the $UMA token. This means it is always more expensive to bribe tokenholders than the value of the funds you could possibly steal. This model is elegant, but limiting for a number of reasons, including that it would require UMA to charge fees soon after getting initial traction, when the priority is growth and adoption.
This hasn’t presented a security problem because the incentives of UMA tokenholders are highly aligned with truthful outcomes. In addition to the economic incentives inherent in the system’s design, there are benefits from the reputation of known tokenholders (plus even the threat of legal recourse) to protect the system in the short term. However, as UMA matures, it will move away from this backstop.
The next step will be for UMA to broaden voter participation, including by encouraging protocols and DAOs that use the oracle also have a stake in $UMA and vote.
Securing More Value
The upgrades coming in January will introduce new patterns that will allow the oracle to secure more value, allowing protocols to plug their own layers of security and arbitration in, together with the DVM and optimistic oracle
These updates will be explained in detail closer to their release.
Inevitable Technology
Forget about UMA for a moment and just ask yourself if you think ‘this thing’ should exist:
A thing where you can ask a credibly neutral crowd of people to both answer and validate information
This thing’s credibility is possible because you can see on-chain that every member of the crowd has skin-in-the-game.
Because the data ends up on-chain, it opens a new design space for coordination layers:
- Contracts enforced without governments
- Incorruptible information markets
- Grants, incentives, employment, investments, and etc. — With no counterparty risk
- News fact checking
- The ability to decentralize entire protocols by disintermediating the core team
That is the vision. UMA’s own mission of “Universal Market Access” is a bold one: to ensure that opportunities to improve one’s life be more equitable the world over. This is the yet unfulfilled promise of DeFi, we will work for as long as it takes on this design because we believe the world deserves it.
Finally, we’d like to sincerely thank our loyal community members who have stuck with Risk Labs and in the UMA ecosystem through all seasons and strategies. Many came during up markets, but stick around for the mission. We think that UMA is onto something big, and we are glad you’re along for the ride.
If you haven’t joined us yet, and want to spend time with some the brightest and kindest anons, hop into discord.umaproject.org.
See you in 2023, UMA 2.0.
Editor’s Note
UMA’s engineers are always looking for ways to improve their code. In the case of UMA 2.0, they had a couple of tweaks to make it even more awesome through streamlining some of the UMA 2.0 code, post initial audit.
Security is always UMA’s highest priority, so although the changes are minor, we are amid a full re-audit from Open Zeppelin, just in case there are any interdependencies. We’re now expecting to begin the governance action to implement UMA 2.0 in late February.