Author: Trusta Labs

On Valentines Day, February 14, 2024, the fifth day of the Chinese lunar calendar to welcome the God of Wealth, the Starknet Foundation (SNF) announced the long-awaited STRK token airdrop plan, the Starknet Provisions Program. Starknet’s native token STRK will be airdropped and distributed through this program starting on February 20, and the collection period lasts for four months until June 20 [1].

New attempt at Pratt Whitney airdrop

As can be seen from the allocation chart of the Starknet Provisions Program, more than 700 million STRK will be allocated to nearly 1.3 million addresses. Compared with the nearly 250,000 addresses that received the first round of Optimism airdrops, the nearly 620,000 addresses that were airdropped by Arbitrum, and the nearly 200,000 addresses that were airdropped by Celestia, this airdrop hopes to be widely incentivized, and the intention of spreading rain and dew is very obvious.

1. More than 50% of tokens are awarded to more than 500,000 individual Starknet users;

2. More than 600,000 dApp users on StarkEx, such as dYdX, ImmutableX, Rhinofi and Sorare, will receive nearly 10% of the token allocation;

3. Starknet is the first L2 project to reward 21.8% of airdropped tokens to independent Ethereum pledgers and Ethereum LSD holders;

4. Reward Ethereum and Starknet developers: EIP authors, Ethereum developers, Ethereum Protocol Guild members and Starknet developers;

5. Non-Web3 open source developers are rewarded based on GitHub contributions. Setting a new benchmark in inclusivity.

It can be seen that the Starknet Foundation is committed to the Provisions Program to cover as many ordinary users as possible, whether in terms of motivating people, covering the ecosystem, or covering non-Web3 developers. At the same time, attracting as many users and developers as possible, the ultimate goal is to bring the project into a more decentralized era. As StarkWare CEO Ben-Sasson said, The Provision plan marks a new stage of more decentralization for Starknet. It stimulates peoples curiosity about Starknet, promotes the learning and use of Cairo, and greatly increases the size of the community. Growth. More and more developers are adopting Cairo, recognizing that a new era of scalability means a new language is needed to enable it.”

Trusta Labs, an independent third-party data service provider for the Starknet airdrop program

How to fairly reward real and valuable project users based on data drive. This is the challenge that lies before the Starknet Foundation and Trusta Labs in August 2023. In the past six months, Trusta Labs has served as an independent third-party data service provider, similar to Nansen in the Arbitrum airdrop program, providing services for the Starknet Provisions Program in the following aspects:

1. Airdrop consulting: Trusta Labs successfully helped Celestia complete the TIA Genesis Drop. Based on this experience, Trusta Labs provides professional, data-driven consulting services.

2. Token distribution based on MEDIA: For inclusive airdrops, a quantitative and scientific value assessment of user addresses is required. Trusta Labs invented an on-chain reputation evaluation system based on MEDIA to calculate the subdivision dimension factor of all addresses for Starknet. Based on this, Starknet singled out important variables and developed token qualification rules and token distribution scores.

3. Anti-witch service: Integrating the anti-witch methods of Trusta Labs and Starknet, a set of AI-First and User-First witch identification solutions were implemented. After multiple rounds of data simulation, it was applied to this Provisions plan.

MEDIA Variable Selection and Application Considerations

In the design of qualification screening and token distribution model, Starknet used the Trusta Labs MEDIA scoring system as a reference [2]. The MEDIA system includes 5 dimensions: Monetary, Engagement, Diversity, Identity and Age, which comprehensively evaluates the reputation value score of an address. Starknet finally selected the following five variables:

1. Engagement - Active Months: Interact with Starknet in at least three different months;

2. Engagement - Interactions: Complete at least 6 or more Starknet transactions;

3. Monetary - Volume: The total transaction amount is not less than 100 U;

4. Monetary - Balance: Hold at least 0.005 ETH on snapshot date (November 15, 2023)

5. Diversity - Unique Contracts Interacted: The total number of different smart contracts that interact;

In the entire MEDIA variable selection, Starknet has implemented its inclusive airdrop idea, and the entire design is simple and broad (compare Arbitrum’s airdrop qualification definition [3]). for example,

• Starknet does not use the Identity variable in MEDIA because this would make addresses with a certain type of identity (such as the owner of stark id) more special and preferred in airdrops;

• Starknet did not use the Age variable in MEDIA, which directly led to complaints on social media that early users were not taken care of. In fact, paying more attention to the interaction between users and projects, rather than judging based on the interaction time, is also an insistence on inclusive airdrops.

• Starknet chose a few simple variables: number of transactions, month, amount, and snapshot day balance to define user qualifications. The thresholds for these dimensions are also very basic. Users need to have made more than 6 transactions in 3 different months, and the total transaction amount is not less than 100 U, they are likely to get the basic airdrop.

It is worth mentioning that the calculation of account balances in the MEDIA system includes Eth and multiple stable currencies. However, due to the emphasis on Ethereum, Starknet has put forward the need to calculate and use the ETH balance. Funds such as USDT are not included in the balance. At the same time, if the funds are deposited in DEXs LP or in a lending agreement at the time of the snapshot, they will not be included in the qualification scope of the snapshot. Trusta Labs observed that some users complained that due to the ultimate use of funds, 0.005 ETH was not retained in the account on the snapshot date (November 15, 2023), which affected the price of short investment. This parameter setting of currency, amount, and time window reflects the project party’s airdrop preference. Trusta Labs will also insist on making suggestions to future project parties and consider using a more comprehensive fund balance form and more diverse gradient rewards. standard.

Anti-Sibyl Service in Starknet Airdrop

Fighting witches is a very challenging task, and the root of this challenge comes from the definition of witches. Trusta Labs never believes that personal multi-account activities on the chain are a witch behavior, even if the purpose of the interaction is to obtain airdrop rewards. However, using tool scripts to interact in batches and compete unequally with real users for airdrop resources, even if it is a so-called premium account, violates Starknets principle of inclusive airdrops.

In [4], Starknet colleagues briefly explained the anti-witch program jointly developed by Trusta Labs and Starknet. This is a two-stage clustering scheme,

• Phase 1 uses multi-category methods to cluster addresses from aspects such as capital network aggregation, first behavior aggregation, token Approve behavior aggregation, and batch cross-chain operations. Phase One is essentially an improvement on a similar Sybil identification scheme found in Arbitrum airdrops. Arbitrum only considers community detection mining for the address fund network, and more address aggregation dimensions are considered in the Starknet airdrop. However, contrary to what everyone has guessed, Trusta Labs will not analyze similar behavior in a single Odyssey event, for example. Similarly, Trusta Labs does not have the ability to obtain user IP, device fingerprint and other Web2 information from anywhere.

• Compared with the Arbitrum solution, the new phase 2 purifies clustering through the K-Nearest-Neighbor based on spatial distance and the Longest Common Sequence method based on sequence distance, minimizing the chance of accidental killing of good people who accidentally enter the cluster.

The solution for witch identification involves relatively professional machine learning and AI knowledge, which we will explain in another article.

A sample address to obtain STRK

Let’s take the address 0x01fbfdc792c9d169f3672e7389c4f406dbbb917cd8a58994cac8bb4ddbc4a3a8 as an example. Let’s first look at his MEDIA score on Trusta Labs’ Trustgo product, which is as high as 96 points.

The Monetary, Engagement and Diversity scores of this address are as high as 99 points, which shows that the address performs very well on relevant variables in these Starknet selected dimensions. Interested readers can go to [5] to see for themselves the Starknet performance of this address.

In the end, the address obtained the maximum 10,000 STRK that a single address can obtain.

Conclusion

An airdrop solution that satisfies everyone may be an infinitely closer dream. Between early users and users who continue to invest, between so-called minimum guarantee accounts and high-quality accounts, between real individual users and group-controlled batch users, the project side must make choices and cannot have both. Starknet’s inclusive airdrop idea may be creating a new paradigm of airdrop incentives. On this road, Trusta Labs will help all users who interact authentically and make the right efforts to enjoy the results of incentives.

For STRK short-term investment qualification inquiry, you can not only check your short-term investment qualification on the official website [6]. Trustgo’s Airdrop Checker product [7] has also provided one-stop support for various recent airdrop eligibility queries including STRK.

In addition to this STRK provisions program, Trusta Labs is also a program partner of the Celestia Genesis Drop. Through the technological accumulation of AI+Web3, Trusta Labs is becoming a provider of objective and scientific on-chain reputation scores and Proof of Humanity verification. Provide project parties with a set of solutions that can fairly evaluate users, effectively screen users, and efficiently motivate users.

【 1 】Starknet Provisions Program：https://www.starknet.io/en/content/starknet-provisions-program

【2】Trusta Labs’ MEDIA scoring system: https://medium.com/@trustalabs.ai/media-score-as-the-infrastructure-for-on-chain-user-value-assessment-c37f68eeb198

【3】Arbitrum airdrop distribution: https://docs.arbitrum.foundation/airdrop-eligibility-distribution

【4】A brief description of Starknet’s anti-witch: https://github.com/starknet-io/provisions-data/blob/main/sybil_report.md

【5】A sample address to obtain STRK: https://trustgo.trustalabs.ai/dashboard/0x01fbfdc792c9d169f3672e7389c4f406dbbb917cd8a58994cac8bb4ddbc4a3a8?chainId= 23448594291968336

【6】Starknet airdrop query:https://provisions.starknet.io

【7】Trustgo’s Airdrop Checker: https://trustgo.trustalabs.ai/booster/0x085ed975a8b6b860de3c2b871da60a3f9f48a5b8?chainId= 324