Behind Arbitrum's airdrop strategy to reward real users, exclude cheats

The highly anticipated Arbitrum airdrop took place last week, with just over a billion tokens set aside for the most loyal users of the Layer 2 scaling project. Since then, the token has secured a market cap of $1.4 billion, making it the 40th most valuable cryptocurrency and the airdrop one of the largest ever.

Prior to the drop, organizers had faced the bigger challenge of making sure that the distribution of tokens went to the right people, and Offchain Labs, the creator of Arbitrum, worked with crypto data company Nansen to define the distribution and attempt to eliminate airdrop farmers — those who try to game the system in advance to get extra tokens. 

“The objective was to put enough tokens in the right hands so the governance mechanism/decentralization takes over,” Nansen research analyst Aurelie Barthere said in an interview. “That will be the measure of success.”

Barthere, who worked on the distribution since August, said that the airdrop snapshot was originally taken late that month. But it was pushed back because activity on Arbitrum drastically increased during the following months and a snapshot prior to that time would exclude a lot of genuine users who had recently started using the protocol. 

Organizers of the drop also wanted to reward genuine users who would take part in governance and sought to eliminate those who would just sell and move on. Here, the focus was to have minimum thresholds for key markers of activity, to ensure that genuine users, even if they didn’t use Arbitrum much, would receive at least some tokens. But it also had thresholds to increase the rewards for those who had interacted with the network on a larger scale, designed to reward power users.

“Once we had the minimum threshold and once we were sure that all the organic users would get something, then we could really play around with the thresholds to the very active users,” said Barthere.

This strategy helped to ensure that the $1.4 billion network's governance structure couldn't be rigged by cheats, a major problem ever since airdrops took off in 2020 as people realized they can game the system. Some airdrops have been too restrictive in cracking down on cheaters, while some have been too lenient.

Identifying airdrop farmers

While Nansen focused on rewarding genuine users, it was also key to stamp out airdrop farmers. These are entities that used many wallets to interact with the network in the hope that, if an airdrop were to happen, they would get a larger portion of the rewards. It’s also seen as a sybil attack, which uses a node to operate several fake identities simultaneously, on the network.

In this instance, Nansen generated a lot of the data and it was sent to Offchain Labs, which identified clusters of addresses that it reckoned were operated by single individuals. These clusters were not seen as eligible for the airdrop.

Barthere said that the biggest shock for her was when she saw how many wallets had all of their trading activity within just two days, a suspicious sign. She noted more than 1 million wallets met this criteria, making up a large part of the number of wallets that had bridged to Arbitrum.

One of the main ways that airdrop farmers were rooted out was by identifying the use of the Disperse protocol, created by Yearn contributor Banteg. This is a tool that lets a single wallet split up some funds and distribute them among a large number of wallets. It’s a very efficient way to have multiple addresses that have interacted with a certain token, or on a certain network. However, it also made it quite easy to see that all of these addresses were linked. 

“We could see this function associated with sybil behaviors because lots of wallets, sometimes hundreds of wallets that had been funded from this function just went on and did maybe three, four or five transactions and often it was buying and selling to themselves, the same NFTs,” she said.

Barthere noted that it was tricky when some wallets sent funds to centralized exchanges because it’s hard to see what’s happening internally. But in some cases, it was possible to see that some deposit wallets were interacting with the same wallet addresses, implying they belong to the same entity.

As for what Barthere would change if she were to do the distribution criteria again, she said that it depends on the success of this airdrop — and whether it results in good on-chain governance.