How to design sybil-resistant airdrops

(Mostly)

For anyone who has spent any period of time on Twitter recently, it's likely you've been inundated with threads about how to FARM THE LATEST AIRDROP and earn $10,000+ while only using $13 worth of liquidity.

This stupidity presents a number of problems for teams who want to attract high-value users and legitimate technology enthusiasts who want to use the latest and greatest crypto applications.

On the one hand, you have organised and professional sybil farms either running bots or employing people in poorer nations to grind on their behalf.

At the other extreme, you have highly misplaced expectations being peddled by those optimising for engagement. People expect to receive annual salaries for doing one bridge transaction and swapping 0.01 ETH to $17.

This results in a race to the bottom, where legitimate users feel they have to engage in sybiling to not get left behind. Or even worse, they give up trying new projects because "it's only whales farming it" or "it's being botted and isn't worth my time."

Pretty quickly, you end up with highly inflated numbers as the 100,000 or so actual people left operating on chain each have 3 or 4 wallets, and those at the extreme operate sophisticated farms of hundreds.

As sybils become more and more sophisticated, the net return for real users begins to decrease, and the loyalty and attention-based return from the capital expenditure for teams issuing airdrops declines too.

Eventually, if teams cannot determine clever ways to reward real users, it's likely LPs start asking difficult questions to their investors, airdrops decline, and the stimulus-like effect it has reduces. Decentralisation wanes, and the magic of this online world we live in resembles more and more the trad-fi world we all wished to escape, with the wealthy and connected grabbing their share of the pie early and the everyday user being unable to invest until a decade later under the guise of "investor protections" when all the growth and investment returns have been drained from the project.

Okay, that sounds bad: But don't projects today already try to remove sybils?

Some projects have done a great job at this. Blur, for example, rewarded users for historic NFT volume but incentivised listing and bidding on NFTs while taking care not to promote wash trading behavior.

This resulted in actual users getting the lion's share of the airdrop, those who took risks with their liquidity were rewarded, and there was reduced incentive to sybil. Smaller players could win by finding niche areas in the NFT marketplace to bid on that larger accounts did not bother to look at.

Blur's incentive structure allowed them to take Opensea's throne as the go-to NFT marketplace by creating sticky and loyal users who remained post-airdrop.

(Blur volume remains significantly higher than Opensea even months post-airdrop)

Other projects, however, have done a less good job at this.

You can probably remember some of the horror stories over the years of when projects have gotten airdrops wrong.

Sometimes it's from having badly designed sybil prevention strategies that wipe out real users.

Or maybe it's designing the incentives that reward wash trading and fake volumes but dilute real users and token holders.

Or projects that lean too heavily on advertising an "airdrop" as a method of customer acquisition.

Setting expectations too high and then failing to meet them has never gone well for any business, but with crypto's switching costs being so low, it's even more devastating in this industry.

It's clear, then, that a badly designed airdrop or a poorly architected attempt to remove sybils can do irreversible harm to a project. But those that are designed well can springboard a platform in gaining market share, attention, and growth prospects.

So what methods can projects use today to prevent sybils from destroying their community without being so strict that real users get removed?

If you want to attract real users, then it makes sense to make the criteria relevant to things actual users are doing.

The above seems like a pretty simple statement, but it's one I believe to be true.

A total of 7,079 wallets have interacted with AAVE V3, yet just this week alone, 10 times that number have traded an NFT, and over 100,000 wallets hold the current popular Memecoin PEPE.

Yet, strangely, I can think of a ton of serious projects that have airdropped tokens to AAVE holders and voters but not many that have been filtered by NFTs or ERC-20 traders.

Now, of course, you can argue that not everyone wants to trade NFTs or memecoins, but over the past three years, these have been the main use case for on-chain activity.

(This gas burnt since the merge chart shows that overwhelmingly Uniswap and NFT marketplaces are the main use cases on chain)

For this reason, I'd propose using a points-based system with a sliding scale based on volume and wallet age.

Points-based system:

This is not a new idea and is something that's been tried before, but in my opinion, with poor execution.

Gitcoin passbook, for example, is a great idea but has a terrible design.

I know this because my main wallet with tons of history, thousands of transactions that's interacted with basically every major project ever released, and has burnt over 25 ETH in gas fees achieves a lower score than a new wallet I set up, linked a new Twitter and Gmail account, and sent $2 to a random project asking for funding on Gitcoin.

THIS MAKES IT SO INCREDIBLY EASY FOR PROFESSIONAL BOTS TO MASS SPAM AND SYBIL.

In fact, it actively promotes bots and spam accounts and detracts from real human users who don't spend hours of their day thinking about intricate ways to make their account look "real" by connecting multiple social media accounts to some centralised server.

So it's time for a better system:

Because I'm lazy and have spent about seven minutes thinking about this, I'm sure some gigabrain teams could add about a thousand more metrics to make a system far better.

Actually, thinking about it, there's probably a pretty decent business for someone that can build a giant data dashboard and design tailored airdrop incentives with teams under an NDA agreement.

But anyway, I'm getting distracted again, so here are some of the metrics I would include when designing an airdrop:

Each of the below would have a different points value that would operate on a sliding scale increasing with volume and wallet age

Has the user traded on Blur or Opensea?

A huge portion of users that exist on-chain will have at some point in their crypto journey purchased or traded an NFT. It's also unlikely that many mass sybil accounts will have done this because of the gas fees involved.

Has the user staked ETH with LDO/RPL/EIGENLAYER?

You can, of course, increase this to include any number of LST providers and add in Solo Stakers. But it's very, very unlikely a sybil user will be holding any StETH because there has been no incentive to do so historically, and holding it for no reason would be inefficient and cumbersome even for farmers who predicted this criterion. Real users, however, may want to earn yield on their ETH.

Average ETH balance over 12+ months

Sybil users will likely cycle ETH between accounts to maximise efficiency, so it's unlikely that over a prolonged period their balance will have remained significant. Real users, on the other hand, are more likely to have a continual ETH balance in their wallets.

Lifetime gas spend

This is the hardest area to sybil. If you are literally burning ETH just to possibly be included in a future airdrop, you probably have something wrong with you. Legitimate users, however, will likely have been using the chain during periods of mania and will have spent a significant amount of capital solely on gas fees. New wallets that have been spawned solely to spam LayerZero and ZkSync are unlikely to have done this.

Uniswap trades

Similar to NFT trades above, if you've been in crypto for any real amount of time, it's overwhelmingly likely you've interacted with Uniswap at least once. For new sybil accounts, however, the token was airdropped years ago, so there is less incentive for fake wallets to use it.

Traded to ERC20s outside of ETH + stables

Sybil farmers hate taking any kind of risk. The most common route I've observed is users depositing from an exchange to Arbitrum, bridging to zkSync, and swapping back and forth between ETH and USDC. Users that have made trades outside of these two pairs are probably more indicative of a real user.

Held a memecoin that was later listed on an exchange

This would be a method of attracting users who are most plugged in and active in the crypto space. If a user is early to a token that later shows credibility and gets listed, it's likely that they're a legitimate user who can add value to your project. This doesn't just need to be for "memecoins" but legitimate projects too.

Minted an NFT or was an early buyer of a BAYC/PUDGY/ART BLOCK

This is the same principle as above but for NFTs. You could take any of the top 100 volume NFTs in history and take a snapshot of accounts that were actively trading them close to launch. This both highlights that someone is clearly active in the space, but it also gives advantages to smaller wallets. Checks and Opepens, for example, or even Loot were all free mints that now command high places on total volume leaderboards. A huge number of active NFT participants will have traded these projects in the early days, but a sybil would not.

Rewarding the community

This might be a bit controversial, but I truly believe that using Zealy/Galaxy quests is the worst way to build a community in the world.

In fact, I now go as far as instantly leaving the Discords of projects that heavily utilise these sites.

No legitimate user would want to "invite ten friends" to the Discord or flood their Twitter followers with copy and pasted shill posts about your product.

By doing so, you are optimising for bots and sybils who can create 100 Twitter accounts with fake followers and have zero interest in the project they are promoting.

A more productive approach would be to ask people to walk into the wilderness and start talking to the trees about your project.

Community-based incentives have their place, but they need to be organic and based on a well-orchestrated, original, and specific culture.

Unique ideas work brilliantly here, such as Taproot Wizards' shower video method. It garners a ton of attention, is extremely difficult to sybil, and builds a community around a peculiar talking point and shared activity. Of course, it's not for everyone, but it has received more attention than any Zealy quest ever.

I believe teams would greatly benefit from developing their own unique culture and rewarding community members who contribute and help nurture it. You don't have to ask people to go to the extremes of showering dressed as a wizard, but creating fun and engaging activities for users to participate in can help grow your project and deter sybils.

Smol Brains executed this brilliantly when they simply asked users to draw a picture of a Monkey with a big head and submit the image to the Discord for a whitelist spot. It's a simple and easy task for a normal person, but sybils are notoriously lazy and will literally not even attempt the most trivial thing if it can’t be botted.

In summary

We understand that airdrops are crucial for the success of a project.

It's also clear that airdrop farming is increasingly prevalent, and sybils are growing more sophisticated.

I believe a points-based system that considers historic on-chain activity, acting as a multiplier when combined with verifiable value-add interactions with a particular project, along with specific community-based incentives, is a robust way to maximise the value of airdrops for projects.

By implementing this approach, I'm hopeful that genuine users will be rewarded, communities will thrive, and bot farms will become unprofitable to operate at scale.

However, if you're a project aiming to achieve maximum results using this method, you have to act first.

Once someone deploys this technique successfully, sybil farmers everywhere will analyse the steps and replicate them on a large scale.

Moloch will be in full swing, and the race to the bottom will begin. But for now, let's utilise this fantastic tool to supercharge growth and reward participation to the best of our abilities.