Beyond APY: the vaults.fyi Reputation Score
In the wild west of onchain yield, investors are drowning in options but starving for trust. Onchain earning shouldn't be a solo expedition into the unknown.
The vaults.fyi Reputation Score we introduce today is a tool designed to cut through the noise—transforming the complex history and profile of each DeFi yield vault into a clear, transparent metric.
See our initial scoring below, and read on to learn our methodology:
For more detailed calculations, see spreadsheet here.
What’s a Reputation Score? Why not a Risk Score?
The truth is, producing a reliable representation of all potential risks in any given protocol is an incredibly difficult task. There are countless ways for protocols to fail: contract vulnerabilities, economic exploits, governance failures, oracle issues, phishing attacks, and so on — just check out hacked.slowmist.io for examples.
Protocols built by the brightest minds and reviewed by the best auditors can suffer a fatal blow from a nearly invisible errors, and governance-minimized protocols can suffer from unforeseen market conditions. In lieu of all of this, one of the best and simplest ways to evaluate a protocol is by its track record.
By building our vaults.fyi Reputation Score, we aim to give the community an easy-to-understand metric that evaluates a vault's credibility, performance, and proven track record.
Implications
Our metric will favor the most “Lindy” protocols that have secured large amounts of capital over long periods of time.
Vaults with lower Reputation Scores are not necessarily risky – they simply have not been battle tested yet.
We aim for our Reputation Score to supplement nuanced risk assessments created by leading experts like LlamaRisk, Chaos Labs, Gauntlet, and the like.
Methodology
We break down our Reputation Score evaluation into five dimensions:
1. Protocol Integrity (40%)
Aims to reflect the scale of trust earned by a protocol over its operational history
Measures Total Value Locked (TVL) sustained over long periods of time
Uses a logarithmic scaling to account fairly for massive differences in TVL between the largest protocols vs. upstarts
2. Pool Diagnostics (20%)
Aims to account for liquidity profiles for each individual pool within a protocol
e.g. Ethereum Aave v3 pyUSD ($11M TVL) will have different liquidity characteristics from Aave v3 USDT ($2.9B TVL), given it has ~250x larger TVL
Similar to the Protocol Integrity component, this metric is a factor of how TVL persists over time and is represented using logarithmic scaling
3. Community Adoption (15%)
Aims to measure holder distribution and diversity
More holders indicate broader, more distributed trust
Future improvements will include:
Analyzing holder concentration and ability for large owners to rug liquidity
Filtering out low-value addresses
Distinguishing between individual and protocol holdings
4. Underlying Asset Reliability (12.5%)
Aim to evaluate the stability of stablecoins and pegged assets
We leverage stablecoin ratings from Bluechip.org
For wrapped assets like WBTC and cbBTC, we arbitrarily assign a score of “D”, similar to USDT’s score from Bluechip.
For emerging stablecoins not yet covered by Bluechip, such as crvUSD and USDe, we do not assign a score.
5. Underlying Blockchain Security (12.5%)
Assesses the underlying blockchain's maturity and security
We use L2Beat's Stages framework to assess security of Ethereum L2s
For alt-Layer 1s Polygon and Gnosis, we arbitrarily rank them the same as Stage 0 L2s.
Known Limitations:
We're upfront about our framework's current limitations. We welcome suggestions and collaborations on the following items:
We do not have standardized ratings for all underlying tokens. As of today, we are missing ratings for emerging stablecoins (e.g. crvUSD, AUSD, USD0), wrapped assets (e.g. WBTC, cbBTC), and other tokens.
We are missing network ratings for alt-L1s (we do not have ratings for L1s similar to L2Beat’s stages for L2 networks).
We are not making nuanced risk assessments related to protocol governance or smart contract reviews. We are not aware of a scalable, standardized way to do so.
TL;DR
Measuring a protocol’s TVL over time provides a useful proxy for how “lindy” it is.
Following the point above, this is a reputation score - not a risk metric.
Community input is crucial for ongoing refinement of our framework.
Next steps
Over the coming weeks, we will consider adding these scores to vaults.fyi products, including our app and our API.
To be clear: this initial framework is a V1. We don't claim to have all the answers, but we're committed to continuous improvement.
Have thoughts or feedback? Join our working group on Telegram.
Our objective is to make onchain yields more transparent and easier to navigate. Join us.