Experimenting with Futarchy* for Optimism Grant Allocation Decisions

[elizaoak]

NOTE: Launch of our Futarchy Grants Contest was delayed ~a week due to external events. The new launch date is Monday, March 10.

*A key difference is that forecasters in Optimism’s Futarchy Grants Contest won’t deposit real money, so there is no real loss to users. Rather, forecasters will use PLAY tokens. The amount of PLAY tokens a forecaster receives is tied to reputation via OP attestations, and top forecasters (based on PLAY token winnings) will be eligible for OP rewards.

TL; DR: Optimism is experimenting with Futarchy during Season 7 to test new ways to more effectively make growth incentive grant decisions. Anyone with a Farcaster and Telegram account is encouraged to participate as a forecaster by signing up here.

• All forecasters will have the opportunity to influence additional grant decisions
• Optimism governance contributors with OP attestions (including Citizens, Top 100 Delegates, Council and Committee members, Retro Funding Guest Voters, and NumbaNERDs) are eligible for higher amounts of PLAY tokens (voting power)
• Top forecasters will receive OP rewards and will be highlighted on a public leaderboard in June, once we’ve measured the impact (increase in Superchain TVL) generated by projects

Futarchy: What and Why?

• To recap, futarchy is a form of government proposed by economist Robin Hanson where prediction markets (more precisely, decision markets) are used to determine which policies will have the most positive effect.
• Decision markets are similar to prediction markets but rather than simply predicting whether an event happens they directly decide what happens.

Historically, making grants decisions or other predictions has been difficult for DAOs. Current approaches such as Council structures are less scalable, less transparent, and more centralized, so at Optimism we we’re exploring more effective approaches to making predictive governance decisions.

• Futarchy is said to lead to more accurate decisions by incentivizing accuracy: When decision-makers are held accountable for decisions (i.e., they are rewarded for accurate decisions and penalized for inaccurate decisions) this should incentivize thoughtful, rather than random or biased, decisions.
• In addition, a permissionless futarchy model encourages broader participation — wisdom of the crowds — rather than, say, a centralized and closed decision-making body.

Optimism’s Futarchy Experiment: Grants Allocation

At Optimism, we want to understand: Is Futarchy a more effective way to make grant allocation decisions than existing approaches? Specifically, do projects selected via Futarchy see a greater increase in Superchain TVL than projects selected by the Optimism Grants Council?

To do this, we are experimenting with elements of Futarchy* in our upcoming “Futarchy Grants Contest” which will launch on Feb 27 and run for 3 weeks.

*A key difference is that forecasters in Optimism’s Futarchy Grants Contest won’t deposit real money, so there is no real loss to users. Rather, forecasters will use PLAY tokens. The amount of PLAY tokens a forecaster receives is tied to reputation via OP attestations, and top forecasters (based on PLAY token winnings) will be eligible for OP rewards.

• The experiment opens March 10 and runs until March 20, in parallel with the Grants Council cycle. Forecasters can sign up to join at any time during this period. At the end of the 3-week experiment, the top 5 projects will receive a 100K OP grant each.

• The question forecasters will be trying to accurately answer: If given a 100K OP grant, which protocol will have the highest increase in Superchain TVL?

• At the end of Season 7 (June 2025), top forecasters will be eligible for rewards, based on the accuracy of their predictions.

*Note – The impact metric of Superchain TVL is based on the H1 Collective Intent. From an experimental design perspective, truly isolating the effects of an OP grant on Superchain TVL is difficult because there will likely be spillover effects among protocols: E.g., shared dependencies mean improvements in one protocol that receives grant can affect others (even if no grant). We recognize this and hope our experimental pilot will descriptively show how the Grants Council vs a permissionless play money market identifies high-impact projects within the same interdependent ecosystem.

Sign up to participate in gov innovation and influence grants allocation here!

FAQ

Who can participate as forecasters?

Forecasters are the voters in this grants contest. Anyone can participate by signing up on this link. You’ll be asked to link TG and Farcaster and some other identifying information for sybil protection. If you have received an OP attestation, you should use the same wallet as you are eligible for additional PLAY tokens.

In a real market, forecasters deposit varying amounts based on conviction and knowledge. We wanted to recreate that dynamic and use involvement in Optimism to proxy that given our play money design. Special thanks to Robin Hanson for helpful discussion here.

The image above comes from the Futarchy Experiment Leaderboard, which will publicly display individual top forecasters and leading categories. It currently reflects synthetic data but the categories are accurate — if you fit one or more of categories 1-6 listed above you are eligible for additional PLAY tokens. Unaffiliated users will receive the baseline amount to participate.

Who can participate as projects competing for grants (protocols)?

We have a list of 23 protocols recommended by the OP Growth team who have agreed to sign up. We have decided to focus this round on projects within a specific Superchain TVL range to get the most meaningful insights. Once you sign up to participate as a forecaster and access the user interface, you’ll see the 23 protocols listed. You do not need to bet on all of the projects.

How do protocols receive grants?

Five of the 20 projects will ultimately receive a 100K OP grant each. These will be the top five protocols expected to increase TVL the most. The top 5 projects ranked by the GC will also receive a 100K OP grant each. If both projects are selected, the max award is still 100K OP and Grants Council will select their next-ranked projects for the grant (the overlap will be accounted for in any comparison of project selection by GC versus Futarchy).

What precisely are forecasters estimating?

The overall question is: If given a 100K OP grant, which protocol will have the highest increase in Superchain TVL? The top 5 projects here receive grants.

What do the UP and DOWN tokens mean?

For a given project, a forecaster can predict the increase in TVL. Forecasters can use their PLAY tokens to buy UP tokens if they estimate a project’s increase in Superchain TVL at the end of Season 7 (June 12 2025) will be higher than the current estimate, or buy DOWN tokens if they estimate it will be lower. Top forecasters will receive OP rewards based on how close a forecaster’s estimate is to the final outcome, and the amount of PLAY tokens backing the estimate.

This image is a screenshot from (a working version of) Butter’s interface, where forecasters will estimate a project’s Superchain TVL increase using PLAY tokens.

[dilowe]

Burdayım ve öğrenmek.

[salman]

Love this initiative!

Polynomial is geared and excited for the futarchy experiment.

Curious what the collective thinks of our OP Futarchy Proposal - Polynomial

[Jrocki]

Any other protocols planning to share their plan for the 100k OP grant?

[joanbp]

Ok. So, at the risk of making a complete fool of myself by displaying my lack of understanding… One thing I don’t understand is this:

If we are trying to estimate a future increase in Superchain TVL, measured in $USD, won’t this mean that in a scenario where the $USD price of locked ETH rises over the next few months, a protocol with much locked ETH would automatically achieve a large increase in Superchain TVL ($USD)?

To give an example, take Extra Finance: https://defillama.com/protocol/extra-finance

All of their TVL seems to be on OP Mainnet and Base, so Superchain TVL.

From January 1 to March 13, this year, their TVL (ETH) fell from ~47K to ~36K. A decrease of about 23%.

In that same time frame, though, their TVL ($USD) fell from ~157M to ~67M. A decrease of 57%.

It stands to reason that much of their Superchain TVL ($USD) decrease in this period is due to the change in the exchange rate between ETH and $USD.

Now, let’s say that up until June 12, the ETH price returns to the level it was at by January 1. Then we might see the Superchain TVL ($USD) locked by Extra Finance increase dramatically, even if no additional ETH is entered into the protocol. In this case, the grant they might achieve from Optimism, would not be the cause of the increased TVL.

A different protocol with no previous Superchain TVL, on the other hand, would have to actually do something to get value into the protocol to achieve a similar increase in Superchain TVL ($USD). A grant from Optimism might bootstart them.

Here is my question: Isn’t the Futarchy experiment designed such that Extra Finance is more likely than the protocol with no previous Superchain TVL to achieve an increase in TVL, and so voters are more likely to pick Extra Finance for the grant - regardless of whether giving them a grant would actually have any impact on that increase or not?

[franz]

Hello all

We are excited to be part of this initiative and would like to share how Beets and Balancer plan to increase Superchain TVL: Futarchy - Beets & Balancer

The big buzz word of our initiative is Boosted Pool. Boosted pools automatically double every $ that is LPed into them. How you ask? Easy: 100% of the TVL is rehypothecate into a lending protocol, such as Aave, creating both trading AND lending liquidity at the same time.

We denominated everything in ETH to account for market movements but still have $USD values written down in the article shared above.

[MSilb7]

@joanbp

If we are trying to estimate a future increase in Superchain TVL, measured in $USD, won’t this mean that in a scenario where the $USD price of locked ETH rises over the next few months, a protocol with much locked ETH would automatically achieve a large increase in Superchain TVL ($USD)?

I work at OP Labs, but this is my personal opinion

My take: There is price skew no matter what you do, and USD is easier to conceptualize.

• If you price in ETH, then stablecoins get more valuable as ETH/USD goes down
• If you price in USD, then ETH gets more valuable as ETH/USD goes up
• Today, 40% of Superchain TVL is in ETH+LSTs+LRTs, ~25% in Stablecoins, ~17% in other wrapped assets (i.e. BTC) (Link)

For future: We created a net flows metric (Defillama also does this) as a proxy for “what is attracting/losing value”: (tokens in - tokens out) * some fixed price

• This could be a good metric to use in the future, but there are complexities we’d all need to figure out there first (i.e. do you pick fixed dates or moving averages, can data providers support this).

Also very open to other metrics ideas that we can build! I love that we’re taking it seriously.

[joanbp]

It’s not so much that someone will have an advantage depending on how the increase in TVL is meassured (in USD/ETH/…)

It’s more that we seem to be wanting to use the futarchy to decide who should get a grant, but if price fluctuations account for most of the increase/decrease in a protocol’s TVL, then increase/decrease in TVL seems to not be a great metric. It will not tell us who will use the grant well to achieve an increase in TVL. Thus, it will not point us to the best investment of the OP grant.

[Michael]

Hey Joan,

You are exactly right to be thinking about it this way! If the price of ETH goes up, that is a advantage to all protocols who’s TVL is majority ETH.

But that is a big “if”. Many people thought that the price of ETH was destined to go up months ago, and here we are.

Conversely, any protocol who has TVL in USD and not ETH will have an advantage if ETH/USD goes down. So the correct approach is to make your predictions with your own estimate of the probabilities of the movement of the price of ETH included.

Now in regards to the other part of your question. It is definitely possible that normal operations of projects will have a larger affect on TVL than the impact of the grant, we need to include both into our estimations. But that doesn’t mean that projects with the current highest superchain TVL will automatically get picked. While you think that Extra Finance has an advantage, someone else might see that Superform has been growing a lot recently and rank them higher. Somebody else might be looking at the governance forums of another project and realize that they have a new product coming out on the superchain. This point is proven by the fact that Extra Finance is currently ranked 10… not quite everyone agrees with your analysis. The beautiful thing about this market though is that it is a synthesis of the opinions of everyone participating!

[joanbp]

Right. Thank you! That does make sense. I can see how my assumptions were a bit lacking. It’s good to be learning.

[TheDoctor]

I filled the form to participate, but it seems it’s closed.

[Cecilia_Extrafi]

Hey everyone,

We’re thrilled to be a part of this initiative!
Check out our plans: Extrafi OP Futarchy Proposal
Excited for what’s coming!

[alexsotodigital]

Hi @elizaoak;

Could you provide more clarity on how the “Top Forecasters” will be chosen?

It is said that it will be the accuracy of the predictions as well as the amount of PLAY tokens backing the estimate. But how much does each one weigh?

@elizaoak

At the end of Season 7 (June 2025), top forecasters will be eligible for rewards, based on the accuracy of their predictions.

vs.

@elizaoak

Top forecasters will receive OP rewards based on how close a forecaster’s estimate is to the final outcome, and the amount of PLAY tokens backing the estimate.

My concern is that those who received more PLAY tokens initially (through OP attestations) could have a considerable advantage in becoming the top forecasters.

My assumption is that, to identify (and reward) top forcasters, we should measure “who managed to get more with what they initially had.”; and not simply count the amount of PLAY tokens.

Or am I misunderstanding something?

[Sky]

good questions and good answers too.

Another important factor to keep in mind is that the grant, which is likely to be used for rewards, is in OP token, not USD.

A month ago 100k OP was $120k and now 100k OP is $82k, ~31% less in USD terms. If this continues, the rewards that the projects can use for incentives, could go even lower in USD terms.

Since this whole experiment is using USD as reference, for everything, it would probably be better for the grant to be referenced in USD as well, so all grants are $100k of OP tokens.

It would make things much more clear and aligned with how the results are being measured.

As a thought experiment, if OP token for some reason drops to say $0.10, and each grant is then only $10k (to be used for incentives), the whole experiment gets quite messed up.

[thefett]

Super cool experiment! Excited to play around, and thought I’d just put forward a few issues I see with the design (I know I’m late, but would love to hear the discussion)

Issue 1 – It’s not real money. It’s fake money that we may get something out of in the end. This is different than money in that if we have no knowledge of the system, we’re incentivized to just buy both sides near the end (you don’t want the market to move on you) and try to get any rewards out of it.

Issue 2 – Like 1, we’re all forced to play. Even if I have no knowledge of any of these projects, I’m incentivized to just throw money in both up/down and be liquidity. More likely though, people will just pick projects they like as there’s no skin in the game.

Issue 3 – Distributions are limited. Good distribution is great, but since we’re all forced to play, someone with a lot of knowledge has no advantage over someone just guessing. Ideally if someone knows more than the other participants, they would be willing to put more money down (e.g. move the market big time to set the percentage where they know it should go). Under the current set up, I’m kind of limited.

Issue 4 – TVL is more of an indication of broader market sentiment than chain adoption. If crypto prices crash by 50% by June 12th, you win by betting DOWN on almost every market. Conversely if crypto prices double, you should vote UP. Since exposure to broader market sentiment is not equal across protocols (some hold/ trade riskier assets), the contest probably better tracks with which systems are likely to outperform given your overall market outlook, not necessarily whether the project is actually getting more usage or traction.

I think if you allow people to buy into PLAY tokens to fix the distribution or just give the play tokens some expected return greater than just guessing, you can address a lot of the issues. And lastly, maybe controlling the TVL by token price or adjusting it somehow to be market price neutral would probably help in getting around just betting on market sentiment.

[adaw]

Hey frens! Maverick is excited about this initiative!

Here’s how Maverick can help bootstrap Superchain TVL effectively: Futarchy -Maverick Protocol

We look forward to your thoughts and votes - Let’s build!

[Milo]

I had fun with this. A few things:

• I was a bit confused with how many OP-PLAY tokens I had in total. I thought I was spending them, but then it seemed to reset for each project. I’m still a bit confused tbh!
• Requiring 6 transactions for each bet was a lot, and I frequently encountered errors forcing me to refresh and try again.
• It did seem to be more ‘guess the trajectory of the general crypto market’ rather than ‘guess who will use the grant effectively’.
• I was also a bit confused as to what happens when I bet on a protocol that doesn’t receive a grant - would I lose everything I bet? Would I get my OP-PLAY tokens back 1:1 with what I put in, or at the final prices? I was wondering if I should be considering conservatively retaining my OP-PLAY tokens.
• I really don’t have any special insight into any of these protocols, and I’m not sure how much I’m contributing. I feel like I’m diluting the influence of people with more insight. It feels like it would only become an efficient market after years of redistributing tokens to those who are proven to be best at allocating them. So I think it’s a bit of a shame that the entire budget for this experiment was used in one go, I would like to have seen smaller grants distributed over more rounds, with forecasters’ winnings or losses amplified between rounds. Right now it seems like it would be more testing the average opinion rather than testing people’s opinions weighted by their excellence at forecasting. Does this really count as futarchy, or is it just a poll?

[elizaoak]

Thanks to everyone for participating in our Futarchy Grants Contest! We are excited to announce the top 5 winners of 100k OP grants selected via Futarchy.

Top 5 Futarchy Projects

1. Rocket Pool: $59.4M
2. SuperForm: $48.5M
3. Balancer & Beets: $47.9M
4. Avantis: $44.3M
5. Polynomial: $41.2M

Butter will follow up with the full TWAP calculation, which they’ll open-source so anyone can verify the results.

---

You may also recall that the Optimism Grants Council simultaneously ranked the same set of projects to determine 5 winners who will also receive a 100K OP grants.

Interestingly, Rocket Pool, SuperForm, and Balancer & Beets were initially included in the Grants Council’s top 6 rankings. However, because the max award for a project is one 100K OP grant, their final list includes some runner-ups:

Grants Council Project Selections:

1. Extra Finance
2. Gyroscope
3. Reservoir
4. QiDAO
5. Silo

Congratulations to all winners!

---

And of course, thank you to all the participants for joining! This was only possible thanks to the great engagement of our forecasters. Please continue to share any candid feedback on how we might improve future iterations here in the comments of this forum post.