Token-curated Registry with Depreciative Stake Model
TLDR: We developed a TCR implementation with depreciating stake over time. This reduces overall cost of staking, encourages more participants and creates effective information curations on ephemeral, real-time data.
At Band Protocol, we are building a set of crypto-economics tools to enable effective information curation in decentralized communities. One of the tools offered is Token Curated Registry (TCR). TCR mechanic allows network participants to collectively build a list of entries on the topics they are interested in. It is not a silver bullet though — for one, each entry requires an owner to take the risk of staking tokens, which he/she may lose through the challenge. Moreover, TCR is designed toward curation of items with a fixed value, making its use cases further limited.
In standard TCR, the amount of stake to apply and maintain listing is fixed. This incentive structure no longer works when data is real-time, and the value of each entry depreciates as new alternatives emerge. In this article, we are going to explore a practical example of curating real-time data with TCR, discuss potential problems, and present a solution using TCR with Depreciative Stake Model.
Use Cases
Depreciative stake model allows TCR to power curations on a subset of information we call ephemeral data. Ephemeral data is characterized by being short-lived and continuously replaced by more recent alternatives over time.
A few examples:
- Live Information Feed — News articles, weather forecasts, traffic, notifications. This category includes any entries that are regularly superseded with a constant stream of newer, more recent reports of the world’s state.
- Product Announcement — Imagine token-curated ProductHunt. Each product launch is short-lived, as the launch of more recent technologies will always get more attention from audiences.
- Top Movies / Music Chart — Best entries present high value. However popular though, they only receive peak interests for at most a few weeks before gradually dropping off.
Curating ephemeral data using standard TCR can be a challenge. Once an audience consumes an entry, its value depreciates, and so does the incentive for its applicant to keep it in TCR. Depreciative stake model makes sure that the crypto-economics of the TCR aligns with applicants, enabling effective economic of data curation.
Practical Use Case: News Curation TCR
Let’s consider the case where we want to curate daily crypto contents (hint: check out CoinHatcher.com). With TCR, we allow publishers to apply their published contents to the list by staking tokens, which we will henceforth call CoinHatcher Token (CHT). Readers can observe and get notified every time new content is listed, so they get to see fresh content as it comes out. CoinHatcher Token holders make sure that the list comprises of credible, thoughtful, and well-written pieces. High-quality contents drive more readers, which in turn drives more demand from publishers and writers to stake and get their contents listed on CoinHatcher website for discovery — resulting in value appreciation of CoinHatcher Tokens.
Assuming Bob works at a fictitious news publishing company — CoinInsights Inc. Bob proposes his recently published news to the TCR and stake min_deposit = 1,000 CHT. Alice, as a CHT holder, acknowledges that the piece is great and let it get listed in the TCR. The problem arises when we consider that for regular readers, more recent news would present more value to them since they are only interested in discovering news that they have not yet read. After a couple of days, Alice notices that Bob's listing presents little to no value to regular readers since it has become outdated. She speculates that other token holders have the same view, so she challenges Bob's listing and subsequently wins. Bob's stake is now gone, distributed among Alice and other token holders who vote against Bob.
Misaligned Incentive Structure
To avoid being challenged, Bob must withdraw his listing at some point. In fact, Bob would find it challenging to keep staking after listed for a while as the majority of community members would already have discovered his listing. The longer he keeps the news listed, the less marginal benefit he will receive.
Even though Bob’s news may still be accurate and perfectly valid at the time of challenge, the usefulness of his listing to readers remains subjective to token holders’ judgments. He also will not have time to closely monitor his submission every single day. For Bob, this is a deal breaker, so he would rather not risk anything and no longer participate in this TCR. The incentive structure of the TCR fundamentally works against Bob in this scenario.
We present a solution to this problem by introducing the concept of a depreciative stake model — a mechanism that allows TCR to account for depreciation of value in the listed entries.
TCR with Depreciative Stake Model
In standard TCR, applicants stake a fixed min_deposit amount of tokens to propose an entry to the list. By making the min_deposit a decreasing function of the amount of time that an entry is listed, we mitigate the risk of applicant losing more stake than the listing’s value. It also encourages more applications since the listings cost less overall. We call this Depreciative Stake Model.
To demonstrate a simple use case of depreciative stake model, we choose the following formula for the stake curve:
Now, let’s apply the curve to the daily crypto news example. With the new min_deposit model, Bob would still be staking 1,000 CHT at the time of application. A day after his news being listed, Bob will be able to withdraw 100 CHT from his initial deposit with the rest still locked up as stake. After seven days, Bob only has to stake 300 CHT onward to stay listed.
In the case that Bob hasn’t withdrawn his stake, let’s suppose Alice wants to challenge Bob’s listing at day 3. She would be staking 700 CHT to match with Bob’s min_deposit at that moment. If she wins, only 700 CHT from Bob's stake would be distributed to Alice and voters. The remaining 300 CHT belongs to Bob.
With this new min_deposit model, the longer Bob's news is in TCR, the less risk he takes of losing his staked tokens.
Why is this Model Better?
From applicant perspective, the opportunity cost of staking tokens at any given time should never exceed the utility of the entry being included in the TCR. For example, if Bob believes there is a 10% chance that his listing will be challenged and rejected, for him to maintain the listing at day 3, he has to value that being listed is worth more than 70 CHT (10% * 700 CHT, the expected value that Bob will lose his stake).
In standard TCR, to maintain the inequality mentioned with a fixed stake, the probability of the news being rejected must decrease over time as the utility decreases. It is improbable since the readers hardly receive any utility from old news, and thus the token holders would have little incentive to keep the old listings in the TCR.
On the contrary, by modelling the stake to decrease over time, we can balance the inequality by adjusting the stake based on the change in utility and rejection rate.
Curve Shape and Parameterization
As the change in rejection rate and utility (page view, click-through-rate) over listing time should be publicly available, the community should be able to accurately model the curve for min_deposit to suit its curation model. For example, one can choose to implement the following curve based on historical data:
Without further experimentation though, it would be difficult to pinpoint the exact curve that applies for general use cases. For the starter, we will be testing linearly-declined curve, a generalized version of the curve used in the example above, described in the formula below:
Community determines initial_deposit, terminal_deposit, and t_decay_period parameters through community voting.
Back to Solidity World
We finish this article by discussing a possible concrete implementation of the depreciative stake model on Solidity. The implementation can be done by adding a view method that returns the current min_deposit of a given entry at the time of function evaluation. With that in place, any computation that requires min_deposit simply refers to the result of the function, as opposed to using a predefined constant value. Using Band Protocol's Equation library, any arbitrary decay function of min_deposit can be added to a TCR contract only with a few lines of code!
Band Protocol is also using this model to power the upcoming release of CoinHatcher! Our in-progress implementation of TCR with depreciative stake model is available for inspection here.
Conclusion
In this article, we discuss the situation where TCR mechanic with a fixed min_deposit does not work. Depreciative Stake Model, which defines min_deposit as a decreasing function of the amount of time that entry is listed, solves the problem with TCR entries that have depreciative utility value over time. Lastly, we showcase how we implement such functionality in our TCR smart contract.
Advantages of TCR with depreciative stake model:
- Effective curations on ephemeral, real-time data
- Encouraging more applications, as the overall cost of listing decreases
- More entries stay in TCR, thus more data points and token value accumulation
Although the effectiveness of depreciative stake model remains to be seen, we believe it has broadened the set of problems that TCR may solve. To validate our idea, in the upcoming weeks we are launching a simple TCR with depreciative stake model in our crypto insights curation website — CoinHatcher.com. The solution is implemented with Band Protocol, which supports token issuance via Bonding Curve and various curation mechanism. Stay tuned for further updates via our telegram channel.
Let us know what you think! If you’re interested in implementing the Depreciative Stake Model in your smart contract, don’t hesitate to shoot us an email at connect@bandprotocol.com
