Radus: An Optimized Crowdsale Model
This article was written collaboratively by first-year members Alexis Gauba, Ravi Jotwani, Sara Reynolds, and Aditya Sripal, about their recent research project for a sustainable crowdsale model.
Many companies, capitalizing on the current ICO hype, raise exorbitant volumes of money quickly without bearing any measure of accountability. Gnosis, for instance, sold $12.5 million worth of its token in 10 minutes during its crowdsale, obtaining an unreasonably high valuation of $300 million. Our research project focused on building a new model, Radus, to introduce both stability and accountability. The introduced phase based Reverse Dutch Auction system, modified from the Gnosis model, encourages investors to settle on a reasonable, fair evaluation by prioritizing the growth of the product itself.
In this revised crowdsale, we want the ICO incentives to align with the objectives of the founding team. Ideally, we create an ICO that promotes stability and fairness to the investors. Previous ICOs have had their valuations rise exponentially on the basis of unwarranted hype and investor fear of missing out (FOMO). While this sounds like a good thing, a responsible founding team wants their token’s valuation to be tied to the inherent value of platform rather than speculative hype. While the current ICO market rewards any founding team with any ICO model, we believe there will eventually be a market correction that will force founding teams to prove legitimacy before raising vast amounts of money. Even if this correction does not come soon, we believe that it is in the long-term interest of any responsible founding team to have their token valuation tied to the inherent value of the platform that their token grants access to. Thus, we’ve outlined a way to achieve that by setting a token cap based on expected market interest, bringing in institutional investors (such as VC’s and angel investors) early, creating a prolonged release of the token through phases, and by locking trading until the completion of the crowdsale. By opening the sale to early investors, we gain credibility and a fair evaluation of the market and by having multiple phases, the company is incentivized to work on the product during the development periods between phases to to gain investor interest.
At each phase of the auction, we run a reverse dutch auction to sell the allotted tokens for that round. With a reverse dutch auction, we set the initial price per token very high and decrease it as a function of time. Thus, investors can buy once the price drops to a point they feel is reasonable for the token. In our model, we employ three caps to determine when a given phase ends: the money cap, the token cap, and the time cap. The money cap denotes how much money is raisable in a given phase. The token cap sets how many tokens can be sold in a given phase. The time cap is hit if neither of the above caps is hit within an allotted time. If any of these caps are hit, the token sale for that phase ends and every investor up to that point pays the lowest price reached during that phase.
Let’s analyze what this means in practice. Every investor is incentivized to invest at the first instance that he/she feels that the price is reasonable. By the rules of the dutch auction, they will always either pay at that price or less. If a token sale hits the money cap, the company has raised all of the money they desired and need for the next round of development, indicating they are meeting their objectives.
In our modeled ICO, we will have at maximum of 9 rounds, including a “Phase 0”, which we only open up to Venture Capital firms. Phase 0 is unique because we aren’t auctioning tokens, but rather percentages of the total amount of tokens. This allows institutional investors to determine the total amount of tokens in circulation as well as their initial valuation. It also established credibility by bringing in institutions to hold the team responsible for meeting deadlines and development goals.
In each subsequent phase, the sale is open to all investors and will be capped by either the token cap, money cap, or time cap, which are explained in the infographic below.
Here’s an overview of the token sale mode:
We implemented the crowdsale using Ethereum’s smart contracts. The structure of the codebase is as follows:
- RadusToken: ERC 223 token to comply with Ethereum standard and to allow for high levels of security
- DutchAuction: an implementation building off of Gnosis’ reverse dutch auction; accounts for the way each auction will run during each phase
- Phases: oversees and manages the DutchAuction running in each phase
We also built an iteration of a UI that companies could use for a token sale of this nature, allowing buyers to visualize the graph of the pricing function and easily purchase tokens via ETH.
We have been working on formally verifying our contracts to ensure the utmost security. Currently, they are formally verified within the bounds of what securify.ch allows, however, we are looking into services like Oyente for more rigorous proof.
Looking ahead, we plan to keep developing the Solidity implementation of both the Reverse Dutch Auction and Phases smart contracts, making both more efficient and creating seamless interactions between both. We are also currently iterating and fine-tuning the crowdsale specification to carefully account for edge cases.
We were able to deploy RadusToken and Phases to the Rinkeby testnet, but were unable to deploy DutchAuction due to the large contract size. However, Phases cannot function without DutchAuction. As such, while we were able to test the functionality of the model locally, we were unable to do so on the testnet. This means we have been unable to address all issues in the set of smart contracts. We encourage anyone interested in deploying this model to use our work as a framework to develop a more efficient implementation of the ideas laid out here.
Our model provides a mechanism for companies who are serious about achieving their goals to do so, in a way that promotes stability and legitimizes this space over time. While the model still needs fine-tuning and further economic analysis, we believe it is a big step towards making ICOs sustainable. We would like to see companies taking advantage of this model in the near future. Reach out to any of us on Blockchain at Berkeley’s slack if you’re interested in deploying this model, or would like to discuss the model in general.