Introduction to Token Curated Registries
Token Curated Registries are an extremely exciting and rapidly evolving subset of ‘cryptoeconomic primitives,’ protocol based incentive systems that enable coordination and allocation of capital to achieve shared goals.¹ The underlying network assets, more commonly referred to as tokens, incentivize network participants towards some predictable outcome that is beneficial for every network participant. The most popular examples of crypto economic primitives include token curated registries, curation markets, and prediction markets.
While cryptoeconomic primitives is a fascinating topic as a whole, I’d like to focus this post on the concept of Token Curated Registries. Specifically, what Token Curated Registries are and what are their most useful applications. For those that want to learn more about cryptoeceonomic primitives, a great place to start is Andreessen Horowitz’s Crypto Canon Reading List.
What are Token Curated Registries?
Token Curated Registries, TCRs for short, are a type of crypto system in which a native token is used to incentivize high quality curation of lists.² The native token aligns the interests of everyone who derives value from the list towards some common good, in this case, the best curation of information possible.
Lists are a vital part of how humans evaluate information. They’re an especially useful tool for comparing frequently sought after information. Think of how much value we ascribe to popular lists like TIME’s most influential people, Thrillist’s best places to eat, and Business Insider’s best places to work. Lists are valuable insofar that they help people make informed decisions regarding valuable topics.
We often rely on curators to keep lists useful, especially in information sets with a lot of noise. Curators help us by sifting through information and telling us what is valuable and what isn’t. We essentially forgo analyzing information ourselves and instead trust the curator’s opinions.
Trust is a paramount theme in information curation. Problems can arise if curation power is controlled by a central entity. Imagine what would happen to the quality of a list if lower quality subjects could pay a curator to get listed. The quality would quickly degrade as people stop trusting that list as a reliable information source.
Centralized curators can also be biased, which can skew the validity of their claims. If people come to agree that a curator’s decision making is biased, there is a good chance that they’ll stop using them as an information source. Think of Facebook’s algorithmic feeds. They cater more towards what will make them the most ad revenue rather than what the user is actually trying to consume.
Remember that blockchain technology is particularly useful for creating trust in decentralized environments, specifically environments in which actors have some incentive to behave dishonestly. Blockchains incentivize good behavior by rewarding network actors with some scarce asset. By design, the most economically beneficial decision an actor can make is to behave in the best interest of everyone else. This can be extended to curating lists. A network of decentralized curators can be rewarded with a digital asset for quality curation, curation that helps consumers make the best decisions possible.
Mike Goldin was one of the first people to thoroughly document a network in which decentralized actors were economically driven towards curating lists to the best extent possible. Goldin argued that “so long as there parties that want to be part of a given list, a market can be created around what information makes it onto the list.”² ‘Token Curated Registries 1.0’ outlines a voting scheme that assigns curation rights relative to the total of number of tokens held by an entity. The idea is those that hold the most units of the native token have the greatest incentive to curate information to the best extent possible. If the list’s quality deteriorates and users stop utilizing it, token holders lose money. Quality curation keeps the list useful and keeps demand for inclusion on that list high.
Token Curated Registries are experimenting with the idea that a scaled, free-market could produce higher quality curation of lists than today’s algorithmic/centralized feeds. Curation is subjective, meaning it’s value is dependent on it’s interpreter. In the case of news and content today, Google and Facebook have almost total control over what people see. Token Curated Registries is an extremely powerful concept because it allows individuals access to a decentralized alternative to content curation. They subvert the existing business model and provide an economic incentive to curate content in a way that’s optimized for consumers.
Who Uses Token Curated Registries?
Consumers want high quality information because they use lists to make decisions. The more times a list helps consumers make advantageous decisions, the more frequently the consumer will use that list in the future. For example, a list of the best pizza restaurants in New York City. If a consumer finds extremely delicious pizza from that list, they will continue to use that list to discover new places to eat.
Candidates want to be featured on lists, especially if that visibility could result in some monetary gain. Lists are by design exclusive so placement on a list serves as as a form of validation. In the above example, a restaurant wants to be featured on that list because more people would know about it and more people would spend money there. Similarly, restaurants on the list are considered more exclusive and more valuable than those that aren’t.
Holders want the price of their token to increase. Demand for the token is directly tied to the demand from consumers to utilize the list. This means that token holders are incentivized to keep the quality of the lists high so more candidates purchase the token and apply to the lists.
How do Token Curated Registries Work?
Candidates have to stake the network’s intrinsic token to be considered for listing. Token holders then have some amount of time to vote on whether or not that candidate belongs on the list.
There is a clear economic incentive for both the candidate and the token holders to behave honestly. Candidates won’t apply to the list if it’s likely they’ll lose their deposit. For a basic whitelist where the outcomes are binary, it’s likely obvious for all parties whether or not something belongs on a list. Token holders will ensure that only quality candidates get accepted to the list in order to keep the list useful. The more that consumers use that list to make decisions, the more demand there is from candidates to be on it.
The most simple implementation of a TCR calculates votes based on the total number of tokens held by the voter relative to the total supply. The idea is that those who have the most at stake are the most incentivized to act in the network’s best interest.
Many TCRs use a commit-reveal scheme in which voting is divided into two phases: the commit phase and the reveal phase. Voters lock up tokens to a specific vote in the commit phase and those tokens are unlocked during the reveal phase. The two phases are separated to avoid coordination attacks. This disables any one entity from having undue influence over the voting process. The vote results could become skewed if an entity were able to openly broadcast their votes during the commit phase.
Why are Token Curated Registries Innovative?
Blockchains are particularly well-suited for voting applications because network participants can feel confident about the accuracy of the vote count. Votes are just transactions that can be audited and verified so it’s easy for network actors to confirm that the system behaves the way it’s intended to. For TCRs specifically, users can feel confident that every listee was accepted justifiably through a transparent voting process.
TCRs are self sustaining because they don’t rely on a central third party to maintain them and can operate normally in the absence of the their creators. This is accomplished through the use of an intrinsic token that aligns the behavior of every participant towards the greater good of the network. As long as there are economic incentives to apply to and curate a list, a TCR will be actively maintained and made useful for end consumers.
Blockchains in particular enable this characteristic because they allow market actors to coordinate without a central entity. Transactions are settled and cleared autonomously, and smart contracts are able to run with no down time. Blockchain technology obviates the need for any third party to maintain the network’s financial agreements.
Anyone can enter into the network and participate freely. A lot of popular information sources have become ‘pay to play.’ Think ICO listing sites. Applications aren’t evaluated based on the project’s technical merits, but rather whether or not they paid an arbitrary listing fee. In many cases, roadblocks like listing fees degrade the quality of curation and make lists less useful for end consumers. The TCR system on the other hand, allows anyone to enter into the network and participate in value creation. Similarly, anyone is allowed to consume the list as well as participate in the arbitration process at no cost. The more token holders there are, the more the voting base becomes decentralized.
Projects Utilizing Token Curated Registries
AdChain is really where a lot of the thinking around Token Curated Registries first started. The adChain Registry is a set of smart contracts on the Ethereum blockchain which stores domain names accredited as non-fraudulent by adToken holders.³ A listing on the adChain Registry indicates that a domain gets real human views rather than bot traffic. Digital ad buyers can use this list to strictly service bid requests from its registrants at no cost. Token holders are incentivized to verify domains because they want more publishers utilizing the list. The more that publishers use the list, the more demand there is from candidates to acquire adToken and apply to the registry.
Distric0x is a network of decentralized markets and communities built on top of a set of smart contracts and front-end libraries.⁴ Districts are just online communities in which users gather to accomplish certain tasks. Districts generally allow users to post listings, search, rank and pay in a peer to peer manner, similar to craigslist or reddit. District0x leverages the District Registry to manage which districts get access to the platform. The District Registry is a decentrally-maintained whitelist of districts that have been granted access to the district0x network. Every district in the District Registry has been accredited by DNT holders as non-malicious and value-additive to the entire network.
Messari is an open source database designed to track financial and operational disclosures for crypto assets. The idea behind Messari is to be the EDGAR of crypto, a unified platform that
provides easy access to asset-specific information. Messari is a direct response to the industry’s lack of self regulation. There is currently little incentive for projects and fund managers to volunteer certain info. Similarly, information about crypto assets is scattered across various silos and changes frequently as teams iterate on dilution and vesting schedules.
Messari’s proposed solution is a Token Curated Registry that lists high quality projects with vetted information disclosures. The TCR would serve as a token based self-regulatory organization. Projects will pay a listing fee to be included in the list, similar to how public companies pay fees to file with EGAR. Messari token holders will vote on whether or not the project belongs on the list. Projects are incentivized to be on the list because it is a signal of quality and transparency. Assuming demand for Messari scales, it will be hard for unlisted projects to be successful. In the Messari ecosystem, token holders are incentivized to provide high quality curation because they’re entitled to a share of all the application fees. The better the curation, the greater demand from consumers; the greater the demand to consume the list, the greater the demand for applications.
There has to be some level of initial demand from candidates to apply to the list, without any content, a list is useless for consumers. Empty lists or lists with a small set of submissions don’t have enough value to attract more submissions in order to make that list useful. A potential solution is to collaborate with existing governing bodies within the specific domain that the registry classifies. For example, AdChain is working with the Data & Marketing Association to help bootstrap the adChain Registry with the organization’s core members.³
Not all information requires a free market for curation. The best use cases are those in which a sustained economy of consumers, candidates, and voters can be formed around a topic. The topic in question must display some form of information asymmetry so that you can craft an economic incentive to eliminate it.
Candidate demand also needs to grow over time in order to ensure that there’s a continuous stream of applications. More applications lead to an appreciating token price which results in an engaged group of voters who judiciously curate the list. If candidate demand gradually plateaus or declines, voters will have no incentive to maintain the list. That’s why projects like AdChain and Messari are interesting, the cost to create web domains and crypto assets continues to decrease which has resulted in an abundance of supply. In an ever growing sea of noise, a proper incentive mechanism might be the best way to filter out the value.
What happens if an application fee is more expensive than the value a listee generates from being on the list? Crypto markets are still too nascent to be efficient. Speculation on a highly influential list might drive up the price of the token to the point where the application fee is greater than the value derived from being on the list. Demand for the token would decrease which could have an adverse effect on the number of stakeholders willing to curate the list. This isn’t a large issue as free market forces are still at play. It’s likely better for the market to come to a natural consensus on the value of the list.
A Token Curated Registry’s parameters refer to the unique variables that govern its economy. These variables vary by list and are intended to facilitate the most honest curation economy possible. The most basic parameters include the amount of time token holders have to commit votes to a challenge, he amount of time token holders have to reveal their votes to a challenge, and the percentage of votes necessary for a certain outcome to take effect.
Since TCRs haven’t been tested in production environments, the most optimal parameters are still unknown. Situations can arise in which existing parameters disrupt market functions. For example, the amount of time token holders have to challenge an application could be so long that users forget to cast their votes at all.
Registries can have their own reparameterization process in which token holders can propose new parameters. Changes to the parameters are reviewed and voted on in a similar fashion to how to new applications to the registry are evaluated. To propose a new value for a parameter, a user stakes the intrinsic token behind their claim and submits the application to token holders to vote on. Applications are evaluated the same way that applications to the registry are.
There are also a number of potential attacks that adversaries could pursue to deteriorate a registry’s utility. The P + Epsilon Attack refers to a scheme in which an attacker can take over a mechanism to vote in whatever favor they want at zero cost. An attacker credibly commits to pay out voters to vote in their favor so that even if the voter doesn’t vote in consensus, they’re still economically incentivized to vote in favor of the attacker. If every voter evaluates the game the same way, the money they would make from the bribe would outweigh the money they would lose by voting incorrectly, so everyone votes in favor of the attacker.
Another potential attack is called registry poisoning. Registry poisoning occurs when an entity degrades in value after being accepted to the list.² If consumers utilize that list and find that some of the listed items are of poor quality, they will stop using that list to make decisions. Token holders should be vigilant and actively challenge any listed entities that they feel shouldn’t be on the list. The UX of TCR applications has to be intuitive enough for voters to easily manage existing listings.
2018 is without a doubt the year that the crypto community at large learns about and experiments with Token Curated Registries. TCRs are appealing because they’re simple to understand. Why don’t we use a scarce digital asset to incentivize decentralized actors to honestly curate a list of important information? Users aren’t forced to pay the curators an access fee to consume the list because market forces enable curators to profit from their good behavior.
Curation is also an important theme throughout the digital world as a whole. The rise of the internet drastically decreased the costs associated with creating and sharing ideas, which made it extremely difficult to filter signal from the noise online. While companies who focus on curation have become particularly great at doing so, problems naturally arise when a few entities control the flow of information. Token Curated Registries could usher in a new model by which information is ranked and sorted without bias and in the best interest of the consumer.
If you enjoyed this post, you can follow my thoughts on twitter!
 The Emergence of Cryptoeconomic Primitives — Jacob Horne
 Token-Curated Registries 1.0 — Mike Goldin
 The AdChain Registry — Mike Goldin, Ameen Soleimani, James Young
 district0x Network — Matus Lestan, Joe Urgo, Alexander Khoriaty