Funding Request for Liquidity Mining Program Extension from Polymarket

:memo: REVISED PROPOSAL :memo:


:rotating_light: :rotating_light: :rotating_light: LIVE POLL :rotating_light: :rotating_light: :rotating_light:


Project: Polymarket

Proposers:

Jared Kehlmann, Market Curator & Community Manager

Liam Kovatch, Head of Engineering

Dylan Golow, Technical Product Manager

Emails:

jared@polymarket.com

liam@polymarket.com

dylan@polymarket.com

Summary

Proposal to extend UMA-denominated liquidity mining rewards on Polymarket.

Description

In January, Polymarket, the largest decentralized prediction markets platform in the world, began supporting UMA’s Optimistic Oracle as the primary resolution source and decentralized oracle for markets displayed on the website. The UMA <> Polymarket integration has been a major success thus far. More than 1000 markets have been successfully resolved by UMA, more than $65M in total open interest (OI) has been secured by the UMA oracle and the number of active participants in UMA’s oracle mechanism has grown significantly.

To support the UMA integration, Risk Labs granted 100,000 UMA over a 10 week period to Polymarket for a small-scale, experimental liquidity mining program. The liquidity mining program targeted subsidies for two core activities: liquidity provisioning and fee rebates. During the pilot program, the reward rule set has been iterated on multiple times to improve its overall function and exclude any efforts to exploit it. The results have been promising. Organic liquidity, in the smart contract AMMs displayed on Polymarket, is being provided at the highest levels ever and trading activity has increased. March, 2022 was the best volume month for the AMM contracts displayed on Polymarket in a year.

From Polymarket’s perspective, between the liquidity program and the resolution integration, UMA has become an important piece of the ecosystem and platform, providing decentralized resolution and critical incentive experimentation. For UMA, we believe Polymarket represents a sizable integration with significant demand that adds economic value to the UMA ecosystem and has grown the community. This is represented by Polymarket’s placement as the first integration listed on the UMA website. Overall, the integration and liquidity mining seems quite symbiotic. As such, and based on the value adds outlined below, Polymarket is requesting a 30 week expansion of its rewards program totaling 495,000 UMA funded by the UMA DAO. This figure represents ~1.3% of the community treasury.

Value Add:

TL;DR:

  • Polymarket significantly increasing the total economic value and activity within the UMA oracle network
  • Polymarket battle testing UMA oracle hardness
  • Polymarket growing the UMA community
  • Polymarket serving as a signal that UMA works at scale for other projects
  • Polymarket’s UMA-secured OI can be supercharged with extended reward experimentation
  • Contributing valuable protocol feedback

Increased Economic Value and Activity for UMA

Polymarket’s UMA integration has significantly increased the total economic activity within the UMA oracle ecosystem including value inflow. Specifically, the integration requires ~15 daily price requests each with a reward that is currently set at $50. This extrapolates to ~$300,000 per year in new incentives for reporters. As Polymarket continues to grow, the economic value contributed by the Polymarket integration to UMA will also increase. During the proposed 30 week extension period, we expect similar or increased frequency of market creation/resolution.

  • 37,510 USDC total fees have been paid to UMA-reporters as rewards since the 2nd week of January, 2022
  • 333 UMA-resolved markets were created between March 1st and 22nd. This is an average of 333/22 ~= 15 markets a day.
  • Markets are currently parameterized with a reward of 50 USDC, this parameterization is expected to stay static over the 30 week extension period
  • Extrapolating to a year 1550365 = 273,750 USDC per year for price proposers
  • Extrapolating over the proposed 30 week extension period, this is $157,500 paid to UMA reporters

Subsidizing Oracle Hardness

While UMA provides an elegant theoretic solution to the oracle problem, there are practical considerations that must be made in order to ensure oracle correctness/hardness. Particularly, there must be an always-online community of reporters and disputers. The more mature this network of agents are, the quicker and more confidently markets can be resolved. In early stages of adoption, UMA’s optimistic oracle is at its most vulnerable position. Polymarket’s early and sizable integration has helped develop this network of reporters and disputers. We estimate that the integration has resulted in dozens of new active reporters/disputers. As the economic scope of Polymarket’s integration, particularly as open interest backed by UMA’s oracle continues to grow, the maturity and thus security of the oracle will also increase as more incentives will be available to reporters and disputers. In this sense, by being the first, Polymarket is subsidizing the oracle hardening process. Concretely, this allows for more aggressive and competitive optimistic oracle parameterization which allows quicker and less expensive resolutions. This dynamic is a sort of fly wheel that should be incentivized around a large integration. We expect the 30 week extension to further advance the practical security of the UMA oracle with unique, hard-to-contrive free market exploitation tests.

A specific example of Polymarket effectively catalyzing a “stress test” occurred on Wednesday, March 16th, 2022 when a participant in the UMA Optimistic Oracle’s dispute mechanism attempted to push an invalid resolution through the optimistic oracle. The motive to do this came from the same address 0x785f1800d456888cfeafe99154a113066ab45078, which spent 1940 USDC acquiring “Bryant” outcome tokens of “Who will win Wright State vs. Bryant” market on-chain, after it was known Wright State won. The user both proposed the wrong price and disputed it himself once another user tried to dispute. This event resulted in:

  • a dispute that escalated to the UMA DVM, ultimately resolving correctly,
  • a loss of $4690 for the malicious actor
    • address: 0x00000000de7d6d60dd363b87a76b188154025660
    • -1940 USDC from the market to acquire the incorrect shares -1500 USDC from the proposal fee, -1250 from the loser bond
  • 2125 USDC for the UMA DAO

Growing the UMA Community

Through this integration, UMA has gained access to the mindshare of thousands of monthly active traders (MAUs) that are part of the passionate, technical and engaged Polymarket community. Over the 30 week extension period and based on Polymarket’s user growth, we expect UMA to be distributed to at least 10,000 new addresses.

Providing Signal For Other Projects

Polymarket is, by far, UMA’s largest and most significant integration to date. Polymarket markets have had more than $65M in total OI backed by UMA in under 3 months. The magnitude and success of this integration serves as a signal to other projects that UMA can work and can work at scale. Poymarket advertises UMA as the resolution source on every market page on Polymarket.com. Additionally, Polymarket primarily integrates with UMA on Polygon which also signals strong cross network support. Polymarket has served as a valuable signal before, once representing >97% of all value locked on Polygon and being the first major dapp to use the sidechain network.

Supercharging Polymarket’s Open Interest Secured by UMA

The experimental, small-scale UMA <> Polymarket liquidity mining program has also been a major success. Organic liquidity across markets is being provided at the highest levels ever; this has in turn led to increased trading activity and average trade size since the start of the program (February 17th). Simply put, Polymarket’s users love it and it works. At the same time, Polymarket as a platform is evolving and will soon introduce significant market structure improvements. Polymarket expects the OI of the platform to benefit from the upcoming improvements but also believes that they will need a kickstart to be successful. As such, it is in the interest of both Polymarket and UMA to continue the UMA incentive program to allow further incentive experimentation.

Feedback to Improve UMA

A first-mover integration on the scale of Polymarket offers UMA valuable developer insights. The development team behind Polymarket has worked closely with the UMA team offering feedback, pieces of which have turned into product and protocol improvements. Polymarket continues to make suggestions and help improve UMA. All of the integration is open source and available for other developers to reference.

Deliverables:

The deliverable of this proposal is prolonged incentivization of activity on UMA-resolved Polymarket markets with specific concentration on liquidity incentivization and LP fee rebates for traders. We propose distributing 495,000 UMA tokens over 30 weeks according to the following distribution schedule. This is a first draft of the proposed schedule and for maximum benefit and impact should be expected to evolve during the incentive period similar to how it has during the current program.

Reward Ruleset

Initially the rewards ruleset will stay static, building off the current 10,000 UMA per week emission schedule. The current distribution ruleset is outlined below and is designed to incentivize liquidity provision and subsidize liquidity fees paid by traders.

Liquidity rewards for an LP over an epoch will be calculated as the sum of the following across all qualifying markets:

Screen Shot 2022-04-07 at 12.52.18 PM

This essentially says LPs will earn a pro rata distribution of the reward available each block for a market based on their relative liquidity share.

Fee rebates for a trader over an epoch will be calculated as follows.

Screen Shot 2022-04-07 at 12.52.14 PM

In effect, fee rewards are distributed pro rata based on a trader’s share of total fees paid over an epoch. The right to disqualify wash trading from the reward distribution is reserved and has been exercised effectively already, setting a precedent that deters others from doing the same.

Over the 30 week period, it is expected that the above ruleset will be further developed and tuned in order to accommodate evolving market structure and to increase the program’s efficiency. Generally, the rewards will be used to create a healthy, well-balanced, two-sided marketplace which in turn will result in additional value secured by UMA.

Distribution Mechanic

Distribution currently happens via a fork of SushiSwap’s merkle distributor (itself a fork of Uniswap’s). Every week, rewards are calculated according to the defined rule set and a merkle root representing the epoch’s rewards is stored in the contract. User’s submit proofs to this contract to claim their rewards. The total distribution might total a little more or less than the expected distribution each week due to the complexity of parameterization (can’t anticipate how long all markets will last). Any remaining UMA will be distributed in additional epochs at the discretion of the multisig wallet signatories.

Total Budget Requested:

495,000 UMA delivered to a multi-signatory Polygon wallet of which the composition will be decided between the UMA DAO and Polymarket. The wallet address will be made public once it is deployed.


Update 4/29/22:

The following multi-signatory wallet on Polygon is the proposed destination of the granted UMA:

0x8cBefc47656eCb2e9E2cb89731283CC1e8cD7f4E

The wallet is a 2 of 3 gnosis safe multisig. Two signers are Polymarket representatives and the third is Risk Labs’ multisig. Risk Labs’s position as a signer in the multisig does not represent an endorsement of the proposal; they are a neutral party.


Update 5/4/22:

The multisig address added in the 4/29/22 update has been updated according to a suggestion from @pemulis1 to instead include an Ethereum mainnet multisig. The new address has the same policies and signers as the original Polygon wallet.

Team/Project:

Polymarket is a leading decentralized finance (DeFi) project, with a global, technical team. It is the largest decentralized prediction markets platform in the world by volume.

9 Likes

Note that under budget
485,000 UMA delivered to a multi-signatory Polygon wallet of which the composition will be decided between the UMA DAO and Polymarket. The wallet address will be made public once it is deployed

This proposal cannot go forward to a poll until a wallet address is added for the funds to be transferred to.

(NOTE : posted as forum admin)

1 Like

Thanks for flagging @Mhairi! Polymarket will be sure to get a wallet address added before we start a poll. We intend to gather feedback and build some consensus for a period before we initiate a poll. Additionally, a small error in the calculations were noticed this morning, the total request has been adjusted to 495,000 UMA up from 485,000 when the post was originally made.

My first thought is whether you have considered paying proposer rewards and requiring proposer/disputer bonds in UMA tokens rather than USDC?

This would lead to UMA tokens from the UMA DAO filtering down to active participants in the proposer/disputer layer, which in turn encourages their participation in DVM votes, and overall seems like it would strengthen the oracle system. These recipients also seem more likely to hold onto their UMA tokens and use them after receiving them. Another side effect is that you could hold onto more USDC for other purposes, like paying contributors or liquidity providers who aren’t necessarily interested in participating in UMA’s oracle system and would probably sell their UMA, anyway.

A second thought is whether there is any way for the Polymarket community or team to use the UMA tokens in their treasury to participate in the voting system. A long-term goal is for the major users of UMA’s oracle (Polymarket being one, Across being another example) to hold UMA tokens and participate in voting, since they are even more strongly incentivized than a typical voter to get answers correct, since oracle correctness is critical to their system. Voting rewards would also provide income to your treasury and increase the number of UMA tokens available for LP rewards.

3 Likes

@pemulis1

whether you have considered paying proposer rewards and requiring proposer/disputer bonds in UMA tokens rather than USDC?

This was considered, but the conclusion was that it reduced the set of possible proposers/disputers, which in turn compromises oracle integrity. With that said, as the UMA distribution on Polygon continues, this could be reconsidered. Additionally, the price dynamics of UMA make it more challenging to parameterize rewards for long-term markets, though markets could, of course, be selectively deployed with USDC/UMA rewards. I agree with your points that this structure might encourage participants in the liquidity rewards program to hold on to the UMA they are rewarded.

A second thought is whether there is any way for the Polymarket community or team to use the UMA tokens in their treasury to participate in the voting system.

This is a great question. From a technical perspective, it is 100% possible. If the community wants to make sure this happens, the proposal could be edited, requiring signers on the multisig to commit to ensuring the program treasury participants in all votes. This would mean that the multisig would need to exist on the Ethereum mainnet and tokens would have to be bridged, according to some cadence, to the Polygon network, to fund program rewards (could deploy a Gnosis safe multisig to same address for both networks if desired). Another thing to flag, is that this addition/requirement effectively creates a (decaying) 495k UMA voting delegation to the signers of the multisig.

Strong supporter of this proposal, i think the partnership has worked very well

Would love for this partnership to keep going, seems to have been beneficial for both Orgs!

Great proposal, the level of interest in UMA these past weeks has been absolutely phenomenal owing to this liquidity mining program.

This is a great proposal, a really cool opportunity to get greater adoption/use of UMA through a community of very active users constantly transacting in this space. Should keep this partnership going

I agree, this would be great for both UMA and Polymarket. A win win for everyone here

To your first point, I think the proposer/disputer layer has become pretty active and competitive, so I wouldn’t personally be concerned about a lack of participation if you switched to UMA bonds and rewards. You only need one honest proposer/disputer for the system to work and I don’t see a risk that there will be no one with enough UMA to bond or willing to temporarily swap into UMA to bond.

As for the price dynamics, if both the reward and the bond are priced in UMA, the risk/reward calculation for a proposer/disputer remains the same whether UMA goes up or down in price. You are always risking XX amount of UMA as your bond in order to receive X amount of UMA as a reward. I think it’s worth experimenting with this! You could offer a few markets with UMA bonds/rewards and see how it goes. I think you would see the same level of participation.

For the second paragraph, I think that a 495k UMA voting delegation to the multi-sig is a plus for the voting system! The Polymarket team has the same incentives as any other tokenholder, to vote correctly to protect the value of their UMA and earn rewards, with some additional incentive to make sure the oracle results are correct, useful, and reasonable from the perspective of a protocol-user. The 495k-ish voting power would be significant if there was a truly contentious vote going right down the middle but would not make Polymarket a dominant voter in the system, just one of a dozen or so really big voters.

If the tokens were in a Gnosis Safe, you could also use the Optimistic Governor module that’s going into audit soon to propose and execute votes according to a Snapshot vote of your community. This would take it out of the multi-sig’s hands entirely if that’s desired but in this case I personally don’t see a downside to the team members on the multi-sig voting on behalf of Polymarket as a unit or an upside to running the votes through Snapshot.

3 Likes

Oh, also, typical disclaimer that I’m not speaking on behalf of Risk Labs, just as an UMA tokenholder and member of the UMA DAO.

UMA reporters aren’t proposing anymore as majority of the proposal are being dominated by a BOT. Would be great if the rewards which is intended for UMA reporters goes in the right hands or if the Polymarket team could come up with a way to make the price proposals anti-Bot or something which involved more human touch and which one cannot do via preloading the answers?

I’d love the Superumans to be able to share a pie of the rewards over here. The way it is going now, one wallet is dominating all of the proposals

1 Like

This might be controversial, but I think a bot with pre-loaded answers is an inevitable and positive outcome of the proposal system. If the bot gets something wrong by moving too fast, they can still be profitably disputed.

If they manage to propose correctly 100% of the time, and get those proposals in immediately after the request, that’s the best possible user experience for the requester, and if they’re so efficient and fast that they are the ones answering 100% of the questions, they earned it fair and square.

Edit: This also probably allows the requester to ratchet down the rewards over time since the dominant proposer already did the hard work of creating the bot and isn’t likely to shut it down until it becomes no longer worth it to pre-load the answers.

4 Likes

I just want to provide some initial thoughts and questions first to help further the conversation. I’m not trying to lead anybody to any conclusions or solutions as it’s just really a stream of thinking to generate discussion. There’s clearly a lot of positives for both UMA and Polymarket through this collaboration. The two protocols should definitely continue building this relationship.

  1. The key question on my mind is how do we justify this amount of UMA? Or what should this amount be? 495k UMA is ~$3MM for a 30 week period.
    a) The proposal has a comparison to rewards for reporters but that’s just $157k for the same period.
    b) Total value secured by UMA from Polymarket is currently $5.22MM. It may not be the best metric given volume is more important for a prediction market; however, TVS is looked at more for oracles (unless revenue is generated per use)
    c) Harder to measure is the strengthening of UMA’s oracle. It is certainly getting tested right now, but what is the value of that? This is a fairly specific use case for the optimistic oracle that doesn’t necessarily translate to other projects using it.
    I think participants in this forum are supportive and may just say yes to any size, but I think there needs to be more thought in this amount and why we are justifying this.
  1. I think incentives and rewards should be more of a shared effort between the two protocols. The proposal may be over stating all the benefits to UMA, but not really highlighting all the benefits to Polymarket. If UMA were to provide $3MM of incentives shouldn’t Polymarket provide something as well? The first incentive program I believe had some rewards from Polymarket. This would be easier if Polymarket had its own tokens.
  2. There should be some deliverables from Polymarket to justify the amount. Some suggestions:
    a) It appears UMA is on every market page, but why is UMA not clearly stated as a partner or “Powered by UMA” on the the home page. For a large number of UMA tokens I think it would not be unreasonable to have UMA on the top of the homepage clearly stating this relationship and showing markets are settled using the oracle.
    b) Can we consider the use of KPI options where UMA tokens are granted to Polymarket depending on the amount of TVL (or TVS). So the higher the TVL the greater the grant.
    c) Can the liquidity mining program that Polymarket is running also incorporate KPI options?

These are just a few thoughts/questions I would like to bring up and would like to get more discussion on before making any decision.

1 Like

My general feeling is that while Polymarket is a power user of the Optimistic Oracle, the request is quite high, with an equivalent dollar value of over half the value secured by the OO. I’m also not very clear what the benefit is of having the LM rewards increase before declining, could you comment on why that emissions schedule was chosen @L-Kov ?

I like @pemulis1 's suggestion that the proposer bond and rewards are denominated in $UMA, as it provides additional utility for the UMA token for those participating in polymarket, and is attractive to those who understand the value of the UMA protocol and want to become active participants in proposing and disputing. Holding $UMA in the polymarket treasury and using it to vote is also something that I think the Polymarket team should consider, both for the voting rewards, which would extend the lifespan of any funding, and also as a demonstration of “skin in the game”.

The distribution of $UMA to 10k new addresses is attractive as a means of decentralising the UMA protocol, however I wondered if there were any stats available from the current LM campaign which indicates how many of those who have received $UMA have actively participated in securing the oracle, either through proposing, disputing or voting.

I don’t think I agree with @kevin that polymarket is too specific a use case to translate to other projects using it, I think that the (relatively) high number of disputes that we have seen recently have prompted some interesting thinking and is a general benefit to the UMA protocol. Some write ups of these and how they were resolved together with the increased visibility of UMA to polygon users that @Kevin suggests would increase the attractiveness of continuing to support Polymarket showcasing the use of the Optimistic Oracle to resolve prediction markets.

I guess my primary concern is that a 40 week programme (incl the initial 10 weeks) creates future expectations, which may see liquidity shift to alternative markets once that ends. Exploring KPI options/Liquidity Mining 2.0 might offer a more sustainable alternative than distributing “raw” $UMA to liquidity providers

I’m also curious about why polymarket doesnt have its own token, and what plans, if any, there are to launch a token in the future.

(NOTE posted a member of the UMA DAO, rather than forum admin)

1 Like

I totally support this proposal, it’s a win win!

A quick note on the request itself
This request today is to the UMA DAO directly. So instead of appealing to the “UMA Team,” Polymarket is asking the tokenholders themselves to unlock funds from the UMA DAO.

Historical comparisons
UMA has done liquidity mining programs before, for products like uPUNK and uGAS, as well as the yield dollar products (somewhat indirectly via developer mining.) This came from Risk Labs, as did the pilot/initial LM program with Polymarket.

These programs were largely discontinued due to it appearing to be farm-and-dump, and generally the products tended to exist for the sake of liquidity mining rather than the product itself.

One important thing to note was that these programs were under monthly review. E.g., we had messaged it as if we could discontinue the program on any given month.

Sizing and Liquidity Mining 2.0
My opinion is that this is quite a lot of tokens, so that means signing onto a really long program without any “check in.” The 30-week window is a reasonable pace, but signing on for that many weeks is more than I’m comfortable with.

Luckily, UMA has actually built financial products to address this concern:

Basically, you can create KPI Options that align interests better. The “downside” is that those KPI Options aren’t necessarily liquid unless you create a liquidity market for them (which has been done in the past.) But until that time you could create an estimated APY range depending on the KPI payout range.

I think that if we’re talking about this many tokens, it would be much better if the UMA DAO were paying them out based on conditional outcomes that are appealing to UMA tokenholders. This could be governance participation or Polymarket volume – Those two metrics come to mind. I’d love to see Polymarket liquidity providers voting with UMA and for them to have a vested interest in UMA DAO outcomes.

Final notes
I’m an outspoken champion of the partnership between UMA and Polymarket. The conversations in our community around disputes have been extremely stimulating and useful for the oracle. Both teams have put a lot of effort into this working well, and it is indeed a groundbreaking relationship for UMA’s oracle.

I do hope that it’s mutually beneficial: that UMA’s logo on Polymarket is as good for Polymarket as it is for UMA. We are striving to deliver both the actual performance and the brand image of trust and decentralized resolution.

Brass tacks
I understand that PM wants to maintain momentum, and so would like to get this done quickly. For that reason, I’d support a request for 8 weeks of liquidity mining, maybe lighter tokens-per-week though, just to maintain momentum.

I’d like to see that request followed up with a Liquidity Mining 2.0 program that uses some kind of KPI Options or Success tokens.

1 Like

Excited for extending the relationship between UMA and Polymarkets. Resolving polymarket price requests has been a really fun way to gain experience interacting directly with the optimistic oracle. I agree with @hiten_UMA that competing with bots has taken the fun out of it for those of us who were posting bonds. I have stopped trying because in the rare occasion i can beat a bot i felt like i was rushing too much and put myself at risk of making a $4k mistake. I would like to see a KPI option used for the liquidity mining and feel it would give us some assurance that we are getting the expected results from the investment UMA DAO is providing to Polymarket. I would also like to see UMA better advertised on the Polymarkets homepage as @kevin suggested.

Hi. Thank you @L-Kov for putting up a detailed proposal.
I am writing this as a member of UMA DAO. To begin with, the integration between UMA & Polymarket has definitely tested the OO which has delivered and proven its strength.

Towards this proposal, I would like to put down some of my comments as a token holder. These in no way mean that I don’t see the value in this integration and hence the funding request.

First and foremost, I agree with a KPI-based liquidity mining program as has been suggested. It leads to the alignment of goals and brings more value. The payout made in UMA will lead to additional rewards for the users of the oracle who hold onto UMA tokens and use them for voting. They become the guardians of the oracle and get rewarded for it.
The same goes for the treasury that can use UMA for voting. It will guarantee the safeguarding of the oracle as suggested by @pemulis1. It creates an opportunity to earn voting rewards and the community can vote off-chain through Snapshot. The only problem is the lack of token for the community to vote on snapshot(?).

The long-term plan with respect to Polymarket’s token and how it would fit and create value for UMA DAO also needs to be brought out. With a grant of this magnitude, UMA DAO holders would want some reciprocation in the future when and if that happens.

The justification of the amount as suggested by Kevin needs to be brought out through more deliverables. The initial grant was $100000 for 10 weeks. The proposed grant is $3MM for a 30-week period which is 10 times more.

The stats as suggested by Mhairi-UMA about the new addresses that have actively been involved in supporting the Oracle than just dumping it would be crucial. From the perspective of UMA DAO, we want to grow the community and see more valued users being part of it.

As suggested by Clayton, a short-term request will keep the momentum going for Polymarket. Towards this, KPI options or Success tokens can be rolled out for liquidity mining. It will provide a testing ground and eventually conclusive data to iron out some of the comments/observations that have been made.

1 Like