Designing Gatekeepers — Using Token Curated Registries And Work Tokens To Manage Access Control
Where some form of access is approved by a limited set of experts a Token Curated Registry can be used to curate a list of those experts. Think a Dragon’s Den TCR (Shark Tank if you’re in the US) curating Kickstarter Projects to an approved register.
TCR’s are a relatively new crypto-primitive getting plenty of attention. In this article I am going to look at how we shaped the economic design pattern to overcome the limitation of minimum economy size that potentially limited TCR utility for our use case.
To anyone developing in this space right now you’ll appreciate the draw of working with the canvas of programmable incentives and how these can deliver solutions to previously hard to solve real world problems. This is one such story.
If you are new to TCR’s the articles below are a good primer.
Ryan Selkis, CEO & Co-founder of Messari, a distributed crypto data library using a TCR summarises the concept nicely.
The TCR concept is a simple, elegant, enabling technology that will open up new information services business models. They tokenize inherently valuable information signals, and put a price on the “social scalability” of a given curated list.
Overcoming The Issue Of Minimum Economy Size
One of the open questions about TCR’s is their Minimum Economy Size. Mike Goldin raises this in TCR 1.0.
What is the minimum size of an economy necessary to support the decentralized curation of a list in this way? Is it economical to decentrally curate a grocery list?
To incentivise curators (token holders) token prices should increase driven by demand from list candidates. This means that short lists are inherently un-economical. When a TCR has an economy size that is too small curators have less, if any, incentive to curate. That’s TCR economics.
It is the “social scalability” of TCR’s that underpin AdChain, Messari, District0x, Ocean and Civil. Minimum economy size sets a theoretical threshold below which a TCR might not be able to exist, even if there is a need for it. This speaks to how and where in the real world TCR’s can be used.
On the flip side, and if your building in this space essential reading, Simon de la Rouviere explores TCR maximum economy size in this article.
In Need Of A Gatekeeper
Gatekeeper is a term used in social analysis to refer to persons who are able to arbitrate access to a social role, field setting or structure.
The following are examples of gatekeepers. Doctors arbitrate an ill person’s access to a sick role. Publishers arbitrate a writers access to the profession of author. Art critics and dealers act as arbiters of acceptable or desirable (collectable) paintings.
In general gatekeepers are not censors, as such, but are involved in negotiation and impression management. This negotiation revolves around the conflict between those seeking access and the organisational concerns of those whom the gatekeepers represent.
This role can be played by a TCR of any length. However, to curate a short list of experts the original TCR incentive design requires some variation on the original game.
I’m fascinated by the potential of a decentralised academic research funding network, I’ll get into this in more detail at a later date and time but for now I wanted to share one of the building blocks I’ve been working on that utilises a Gatekeeper TCR.
Academic Peer Review Plays A Gatekeeper Role
Research funding bodies effect access to research funds and thus arbitrate the nature of research activity. Peer review plays a pivotal role in this.
Peer reviews are used to decide who gets funded by including the opinions of experts. This is at the heart of an annual $400bn of non-industrial research funding meant to ensure integrity and consistency to a robust, quality process for all funding initiatives.
But a healthy dose of biases, oligopolies and private interests associated with peer review mean decisions might not be reliable. The process fails the academic research community over and over again. The bottom line is that traditional peer review has never worked as well as the sector imagined it to, which when you consider that peers are not paid or otherwise rewarded is no surprise. Peer review is ripe for decentralisation.
Using Demand For Access To Scale Short TCR Incentives
In the case of short lists, social scalability is replaced by another function to produce the same output.
The Gatekeeper TCR design uses demand from a large number that wish to gain access to incentivise a small number that arbiter that access. The work done by a short list becomes the curation incentive that moves the price of the TCR token. For example, peers sit between the number of projects submitted for review and the number approved as fundable and released to our research exchange.
The number of experts curated to a Gatekeeper TCR therefore depends on how productively each listee plays their role based on the demand for that role.
Example
To achieve a monthly volume of 1200 reviews, where each review requires 6 peers and each peer will do exactly 2 reviews, our list is required to be 100 peers long. Change any variable and the needed list length varies.
Productivity is used in place of scale to retain the original TCR design and incentivise curators to curate a high quality list of peers (experts) in a strategic and economically rational way.
Paired Token Incentives — Registry Tokens Plus Work Tokens
Replacing social scale with productivity requires pairing a Token Curated Registry Token (RT) with a Work Token (WT). Our goal remains to incentivise curators of our TCR to maintain a high quality list but also one that works productively. To achieve this we use the price equilibrium of a WT as the main design signal.
The equilibrium price is the price at which the quantity demanded equals the quantity supplied. It is determined by the intersection of the demand and supply curves.
The system is in equilibrium when work token prices remain stable and TCR token prices increase.
Here’s how this works to maintain the TCR incentive design.
The TCR Token (RT) — A Bonding Curve
Where a paired token model is used, the network token is staked and as the candidate is accepted to the list (assuming no challenge) a corresponding amount of non-transferrable TCR tokens (RT) are minted and issued to the new listee. These are priced on a bonding curve (new tokens issued at higher prices). The stake is held as a deposit if the candidate is successfully listed.
An initial list will need to be bootstrapped perhaps with the issue of RT to a small number of founding curators. Post the initial TCR creation, as with normal TCR design, curators will stake their RT to challenge and if successful the network token backing these will mint new RT and issue these proportionally to the challengers.
The Work Token (WT) — Supply Curve
Those wanting access must be reviewed by the Gatekeeper TCR. To do this the network token is staked to mint non-transferrable Work Tokens (WT). The same amount of network tokens will always mint the same number of WT regardless of the WT price (we’re considering top-ups for more complex peer reviews). If you haven’t come across work tokens, this is a good overview.
WT’s are priced on supply curve (they are a commodity). They can be minted (increased supply and decreased price) or burned (decreased supply and increased price).
A Balance Of Mint And Burn — Incentivising List Curation
Gatekeeper TCR listees conduct reviews by staking RT to claim WT. The probability that a listee will lead a review is proportional to the number of tokens staked as a fraction of total tokens staked by all peers. (There are parameters for claiming a review that will be governed by network token holders periodically).
A review is activated when all WT are claimed. RT and WT are then locked. The price of the unlocked supply of both tokens reflects this lock-up.
Reviews use a consensus model where access (positive consensus) and no access (negative consensus) have the same value to the peer in terms of the WT reward. This is important to avoid incentive-caused bias. No consensus, where there is a split decision, is the least desirable outcome.
If a review reaches consensus then the WT are burned by the TCR contract and the network token stake held for the WT mints new RT. These are then distributed to the reviewers proportionately. In this way a productive list incentivises listees doing work by allocating more RT as well as increasing existing RT prices.
Incentivising consensus above no consensus directs peers to conduct reviews where they believe consensus is more probable, regardless of whether the outcome is positive or negative. Where there is a split decision, applicants will need to tighten up their funding proposals in order to re-submit and achieve consensus. Those that achieve negative consensus will have to re-stake for a new review sharpened by the loss of their WT deposit.
If a review achieves no consensus WT and RT are unlocked. Essentially the trade is unwound at the contract price, which may be above or below the current market. There is an opportunity cost to reviewers that incentivises them to look for projects that have a higher probability of reaching consensus.
If the holder of WT’s no longer wishes to proceed and re-submit for review they will sell their WT and redeem their deposit. WT are priced on a spread. TCR listees are strategically incentivised to hold RT if they expect that the TCR will generate a high number of consensus outcomes on a growing number of review submissions in the future.
The Productivity Effect
TCR listees aim for maximum productivity from a list of any size. It is the demand for the list to conduct work timeously that acts as a proxy for scale. A productive list will see the value of TCR tokens increase assuming continued demand for the access rights listees arbiter. In other words, demand for a high quality list.
This incentivises each peer to reach their highest level of productivity on behalf of the TCR. The greater the number of projects reviewed by each peer the greater the impact of productivity on the TCR token price.
Why Who’s Curating Whom Is Important
Of course, productivity is influenced by the quality of those performing the work, so maintaining a high quality list is fundamental. Listees will use the registry token, and not the network token, in order to challenge low quality candidates and set the parameter of list size based on its productivity.
Where productivity replaces social scale a better fit for curators is a cooperative model. Only listees are curators of the Gatekeeper TCR. Messari uses this method.
“I Don’t Want to Belong to Any Club That Will Accept Me as a Member”, Groucho Marks.
Curator identity, quality, expertise, authority and experience are essential components where lists generate decision signals for high stake games like funding. Academic peers listed are best equipped to curate the best list of academic peers. Peers curate peers to achieve high quality short lists that will perform the gatekeeper role productively to their benefit and to the benefit of all network token holders.
Price Movements And Signals
The gatekeeper token model is designed to maintain work token price equilibrium when there is a balance between the number of access requests and the number of listees that arbiter that access. There has to be a sufficient number of listees to encourage new requests for exchange access. There should also be a sufficient number of access requests available in order to incentivise listees. In-balance is the signal for changes in behaviour based on tokenised incentives.
These signals encourage listees to participate rationally in order to maximise their return and generate continued demand for access.
Below are scenario’s that illustrate this and the utility of TCR Gatekeeper design.
Where WT Prices Decline vs RT
A decline in the WT price occurs when there is an increasing supply. More WT are minted than are burned. This occurs when too few WT are claimed by peers, meaning demand for access is greater than the numbers of access decisions given ( I will cover the scenario of no consensus briefly at the end).
As WT prices fall relative to RT, holders of RT are incentivised to claim WT as they will earn the same amount by staking fewer of their own RT. Peers can take on more reviews with the RT they own, the list is incentivised to become more productive.
Where 100 WT must be claimed to activate a review:
Day 1–100 RT = $100 and 100 WT = $100 [100 RT required to claim 100 WT]
Day 2–100 RT = $100 and 100 WT = $90 [90 RT required to claim 100 WT]
Falls in the relative price of WT also signal the need for list size increase when the TCR is at maximum productivity and can no longer meet review demand. In this case the continued decline in WT incentivises TCR curators to judiciously expand the list size (by voting to change the list parameter for maximum list size) enabling new peers to apply, be invited and/or be sponsored to join the list to claim the excess supply of WT.
Should TCR curators not respond strategically, but rather tactically by waiting for WT prices to fall in order to use less RT to claim reviews, then over time this will encourage project owners to withdraw their projects and recover their stake as the likelihood of their project being reviewed is low. This resulting contraction of WT supply will create upward price pressure.
Early TCR list entrants are incentivised to curate a list that expands productively to review a growing number of projects, as list size increases to achieve this productivity the RT price will rise.
When the WT price increases and RT remain stable the design is in an optimally productive state. In this scenario increased numbers of peers will productively achieve review outcomes on all proposals for access at the same rate as new ones are submitted.
The WT price becomes an early signal that the list must grow and be populated with high quality candidates. Where a list is poorly curated the risk of no consensus increases. Low quality listees, unable to informed decisions reduce the probability of consensus. This will lead to an excess of WT in supply.
An excess of WT supply also occurs when a high percentage of projects are likely not to achieve consensus on the opinion of peers. This in-balance will also result in those projects eventually being removed, resulting in supply contraction
Importantly, price squatters that hold RT and do not review when the system is out of balance will see their stake slashed and their TCR inclusion possibly challenged.
Where WT Prices Increase vs RT Prices
A rise in the WT price occurs when there is a decreasing supply of these tokens as more WT are burned than are minted. This is a result of a high number of WT claimed by peers, meaning the number of access decisions given is greater than the demand for new access decisions.
As WT prices increase relative to RT, holders of RT are less incentivised to claim WT as they will earn the same amount but have to stake more of their own RT. Peers can take on less work with the RT they own, the list is likely to become less productive.
An increase in the WT is an early signal that there are too many peers relative to the volume of access submissions. This illustrates why curators must not increase the list size too quickly or it will become increasingly unproductive. Too many peers and too few projects.
Increases in the relative WT price will encourage existing peers to leave the list, as there are too few projects available to justify tying up capital in the TCR pool, by selling their RT and redeeming their deposit . Early entrants have a high incentive to sell some or all of their RT if they can potentially make a profit. Later list entrants (those paying the highest RT price) are incentivised to leave the list first to avoid losses.
Of course, there will be periods of in-balance that are tied to the responsiveness of researchers and peers. Peers may also feel that the time they can commit has changed and the number of RT they own is too high. IN this case they may simply decide to reduce their workload and sell RT.
In Conclusion
Where time is a precious commodity in the academic research community, new tokenised incentives are likely to hone behaviour over time. The base line for the game remains intact, where list curators are incentivised to judiciously curate a high quality list of sufficient size for it to be optimally productive.
Note: parameters such as minimum time to review, active review period, lock-up, re-submission period are all parameters that will be set to manage the relationship between these two tokens.
I’m tinkering with the economic model under different probabilistic assumptions and working through how to overcome the challenge of tuning the system. Thanks to Matt Lockyer for his feedback. If you’d like to know more twitter.com/jonthnH.
