RFP (Idea): Increased Capital Efficiency and Risk Management with RociFi

Author: RociFi

Key terms

Credit scores: numerical scores representing a user’s (wallet address) creditworthiness and trustworthiness based upon their on-chain transaction history. The scale is 1 to 10 with 1 being the lowest risk (best score) and 10 being the highest risk (worst score). The scores are calculated looking at numerous DeFi protocols across 8 blockchains - not simply Radiant.


RociFi is a DeFi protocol enabling under-collateralized lending via on-chain credit scores – issuing 35,000 credit scores and 1,300 under-collateralized loans to-date.

Our credit scores are real-time, composable, and scalable to allow any protocol in DeFi to integrate them seamlessly. Integration via loan pool parameter recommendations and risk dashboard could allow Radiant Capital to increase capital efficiency and risk management.

To-date, we have integrated our scores with CyberConnect, Relation Labs, and ConvoSpace; and received grants from Lens and Aave to explore the same.


Capital efficiency and risk management are at the core of all DeFi lending platforms. Simply put, if protocols have better information, they can make better decisions for their token holders, which can lead to greater revenue with less risk.

Our credit scores have been battle-tested under the difficult conditions of blockchain-native under-collateralized lending; generating an 82% repayment rate. Having access to battle-tested user level credit risk metrics could allow Radiant to become safer and more capital efficient, while building the foundation for differentiated products like user customized lending in the future.


The integration of credit risk analytics into Radiant could drive greater revenue to the treasury and less risk to depositors. It further could allow Radiant Capital to differentiate itself in the future with new product offerings like customized lending, which fits with the ultimate vision of growing the utility of the Radiant Capital ecosystem. Furthermore, our credit scores have multi-chain coverage, meaning that they can grow with the Radiant ecosystem is it expands to other blockchains.

Radiant using RociFi credit scores for user level, risk-based loan parameter optimization, offers three key benefits (ranked by priority):

  1. Increased Capital efficiency – more revenue to the treasury based upon loan-to-value ratio optimization in the lending pools. Near-term priority.

  2. Risk monitoring on a user level granularity for pre-emptive spotting of red flag situations – proactively decrease loan-to-value ratios or other risk mitigating actions. Near-term priority.

  3. Customized lending terms for every user based on credit scores – product differentiation and revenue boost. Mid-long term priority.

Capital Efficiency Simulation - Radiant Capital

We ran a complete capital efficiency and risk simulation for Radiant Capital using our credit scores over an evaluation period from 2022–08–05 to 2022–10–31.

We did an analysis where we defined “good” accounts as those that did not default on their loans due in this period, and “bad” accounts as those that did. To better understand if more creditworthy borrowers are less likely to get liquidated, we can look at the distribution “Actual Goods” vs. Actual Bads” for the broader ecosystem first.

To get a better understanding of the impact specifically to Radiant we can take a look at the distribution of Radiant borrowers. Any borrowers who have only ever taken 1 loan are excluded from this analysis.

For Radiant, the distribution of borrowers tends to lean more to the right. Roughly 81% of “bad borrowers” get score 10 with 65% of borrowers who did not default also ending up on score 10. About 95% of “bad borrowers” end up on scores 7–10. Part of the reason for 65% of “good borrowers’’ ending up on score 10 has to do with both length and depth of credit history. 50% of these borrowers had fewer than 5 loans and all of them had accounts less than a week old at the time of scoring.

The bad loan rates for the broader ecosystem and Radiant are as follows:

On the broader ecosystem, the risk of liquidation remains well below 1% up to score 6, however for Radiant borrowers this risk is generally much higher, with the risk already at 1.05% on score 2. On both we can see that the risk is increasing with score, with the exception of some confusion around score 3 & 4, 6 & 7 on Radiant.

An LTV curve by score is proposed below based on an assumption around liquidations slippage, implied default rates, and asset standard deviations. For reference, RociFi’s credit score scale is 1-10 with 1 being the best and 10 being the worst. Furthermore, the LTVs are intentionally aggressive to generate the most aggressive liquidation scenarios, thus most conservative projections.

For stable collateral assets (USDC / USDT):

For volatile collateral assets:

The formula for picking the LTV curve here was as follows:

The idea here being assuming that we don’t want LTV’s to exceed 100%, using a 5% margin of error, we scale back the LTV by the Expected Default Rate, and 2x the daily standard deviation of price returns. The scaling factor was selected in each instance so that whenever volatile collateral is involved, in the worst case we assign an LTV of 60%, this is consistent with the current maximum LTV on Radiant. Given default risks and price volatility, we believe this LTV curve conservatively gives sufficient runway for fast price movements that may cause problems for liquidators in light of default risks.

Following these LTV’s we simulate lending revenue’s using approximate interest rates based on what was observed on Radiant, those APRs used for simulation are as follows:

We estimate if Radiant had used RociFi’s LTVs as proposed, it could have generated $12,117,061.89 in revenue.

For comparison, we also look at the revenue we estimate Radiant would generate using current base LTV’s and the interest rates proposed in this document. Since July, we estimate this would have been $11,560,362.28, annualized.

Obviously, increasing the leverage borrowers are allowed to take enhances the potential return given the same interest rate. The increase in revenue is about 5.19% increase, which represents the effective increase in capital efficiency to the Radiant ecosystem.

We can take the difference in these quantities, and look at the incremental revenue by month:

Overall, we estimate that using the RociFi NFCS, Radiant could have generated an additional $556,699.61 in revenue.

Loan Distributions – currently, we rank 89% of Radiant borrowers as 10. The risk observed among Radiant Borrowers is generally much higher than the broader ecosystem. Generally speaking the feature values observed for Radiant borrowers do indicate they are of higher risk than the average borrower that uses AAVE / Compound. This is not to say all Radiant borrowers are risky, but a sizable minority are. Regardless, the RociFi NFCS allows for credit migration based on user behavior, both good and bad. Those Radiant borrowers who use the protocol responsibly, along with other protocols, will see their scores will increase over time.

Although most Radiant borrowers score closer to a 10, the majority of borrowing volume comes from less risky better ranked borrowers. 47% of borrowing volume is coming from borrowers ranked a 4 or better.

Potential Losses – using the proposed LTV curve principal losses were minimal. We observed only 1 instance of principal loss estimated at $2,917.49, i.e. bad debt.

The full simulation report can be found here. Also, an additional example that we ran for Moonwell, lender on Moonbeam, as well.


RociFi is requesting an initial 6-month pilot before both parties commit to a longer term engagement. The proposal covers the following deliverables:

  1. Credit Analytics – Dedicated instance and custom API made available for the protocol and stakeholders to query in near real-time. Includes credit scores and other key metrics for current borrowers. Example information below:

{‘CreditScore’: 2, ‘Address’: ‘0xdf0944d413f83abeba6bc23891bc183bb9d6a77b’, ‘features’: {‘count_borrow’: 103, ‘total_borrow’: 304940.38192349765, ‘total_repay’: 304550.954891286, ‘count_repay’: 95, ‘count_liquidation’: 0, ‘total_liquidation’: 0, ‘days_since_first_borrow’: 648, ‘count_deposit’: 238, ‘total_redeem’: 220858.1913208516}

  1. Data integration – API data into a custom dashboard for real-time monitoring, plus continued maintenance.

  2. Parameter Recommendations – Parameter recommendations based on simulations of expected liquidations across credit scores in Radiant lending pools; subject to price and liquidity volatility that may cause problems for liquidators.

  3. Documentation of key terms and metrics.

These deliverables will be split into four main milestones and detailed in the following section.

Steps to Implement, Timeline, and Cost

Project team: Full-stack engineer, Project Manager, Data Engineer, and Data Scientist.

Monthly cost: $25,000 USDC per month with a 6-month term. All milestones must be approved by the protocol before moving onto the subsequent.


  1. Advance of first month’s payment of $25,000 USDC to cover setup costs (1-2 weeks)
  • Setting up K8 cluster for the custom instance of the Underwriter API and its services
  • Setting-up data lake in PostgresSQL
  • Configure Airflow instance
  • Configure data collectors (Airflow DAGs)
  • Backfill data in the data lake
  • Logging and alerting via Grafana and Prometeus
  • Radiant Capital data integrated into data lake
  1. Credit Analytics, Setup and Testing (2-4 weeks)
  • Training Credit Risk model on the data lake
  • Configuring and customizing feature engine
  • Testing new model version on the reference set
  • Deploying new model to production cluster
  1. Data Integration, Setup and Testing (1-2 weeks)
  • Production API deployed and tested
  • Live dashboard tested and deployed for end user
  1. Documentation and Parameter Recommendation (1 week)
  • Simulation environment of expected liquidations across current users (credit scores) in Radiant lending pools
  • Stress test simulation against historical price and liquidity volatility of collateral used in lending pools
  • Curate specific loan-to-value recommendations for each pool based on results
  • Full documentation of key words and terms


We’d like to gauge interest before moving into a formal proposal.

Please post your thoughts in the comments section.

We are open to feedback or questions regarding the proposal. Thanks!

The recent RFP from rocifi is very well thought out. This proposal could be a great way for people to build a credit history for Radiant lending unlocking more capital efficiency for all. Also, it may promote longer term thinking for lenders / borrowers and ultimately help keep toxic flows out of the system.

I strongly recommend to check Atadia Twitter handle: @atadia_io. this is a data scientists project on Solana and they’ve already built a more sophisticated wallet based credit score that result in a much lower default rate that tat 18% of ROCIFI. they already shipped a product called Atadian pass that users mint and get frozen in a certain wallet and tracks all on chain activity to provide their algorithm with all the data needed to provide a much detailed credit score. the guy tested their product already under their defi protocol “the lending lab” and they’ve been lending unsecured loans in $SOL for over 8 months now. last I checked their default rate was something between 5-10% remember those are completely unsecured loans with zero collateral capped at a certain ticket size per each borrower to diversify risk. I know the founder (twitter handle: @puppetatadia) talked to him about Atadia. and he seemed really interested to connect with radiant. I believe it’s worth a shot the guys are really innovative, data science driven and received a grant from Solana foundation to develop risk management practices and protocols on Solana. already have an NFT project with 1.5 Mn MCap down from an ATH of 9.6 Mn last May and it’s one of the most esteemed and proven projects in Solana NFTs.

1 Like

I think credit scores and more efficient lending is a very worthy avenue of exploration.

The ArcX Defi Credit Score is another one worth assessing in this space.

One of their takeaways was that it is very important to present the user with their credit score, and to give them actionable steps to improve the score.

In other words, if the score is too complex or a blackbox, the user doesn’t understand how to get the score from 2 to 10. Or alternatively doesn’t understand why they’ve received a penalty on their score and now can’t borrow under-collateralised. So the opacity of the criteria / model becomes a pain point for the user rather than having a simple score with a message saying “do this, this and this” and you’ll be back to under-collateralised borrowing.

Another concern here is simply the web2 architecture of their system, and Radiant taking steps to have a centralised API as a key input into its lending system. Relatedly, I also think a big question here is Intellectual Property. This is setting up Rocfi to be very embedded in the product and to keep raising their prices, since the blackbox working of their model is not supplied. So to what extent can Rocfi be paid as a consultant to set up the model for Radiant to run without Rocfi.

To summarise broadly speaking credit scores is a great avenue to explore. Specifically my intuition is paying them to setup a bunch of cloud services with a project manager is overkill at this stage, and potentially not in the strategic interests of Radiant product longer-term. For example, all transaction data is transparent, will an open source web3 credit score product emerge that every service can reference?

From Chris_RociFi:

I can appreciate that you have ties to Atadia project and seem to like them. They look like a sharp team.

A couple clarifying notes on your default rate comments:

Per Atadia’s https://atadia.gitbook.io/wp/start-here/investment-memo,

Our application basis default rates have come down from ~26% in Phase 1 to roughly 3-4% in Phase3. Volume basis default rates have come down from 38% in Phase 1 to around 3-4% in Phase 3.

We calculate our default rate on the total loans issued all-time; not at a point in time, i.e. phases. If we broke ours down into phases, the most recent phase would be <1% default rate. We present this way so users have an understanding of the risks over the entire history, not just a point in time. Furthermore, their Lending lab appears more as a closed or controlled beta, which will distort results. In their own words – “our lending outcome dataset remains small and our model still lacks external validity (i.e. it is still not yet advanced enough to predict financial credibility of people outside of our sample with sufficient confidence).” Our protocol is on mainnet and open to the ‘wild’ so to speak.

Regards to sophistication and smartest minds:

It’s extremely hard to make an apple to apples comparison here. But, I can assure you that RociFi is one of the most innovative and hardest-working teams in the space, and has been working to solve this problem since 2021. We’ve also received grants from Aave, Lens, and Polygon as a reference checks.

Happy to discuss more on the differentiation between our projects.

1 Like

From Chris_RociFi

From understanding ArcX still has DeFi credit scores but pivoted to a different analytics angle.

Re: Complexity and Blackbox:

We do a lot of education around this for our users via our blog. We show many easy ways users can improve their scores while understanding why they received the score they did.

Why ML versus human analyst? For better or worse, a machine has proven to better spot risks that are less interpretable to the human eye. However, we provide enough actionable info to users as possible while preserving the model from abuse by bad actors.

For better or worse, a little bit of Web2 infra is needed to maintain the scoring system’s privacy for this use case. If it were completely open source, bad actors could simple game the system; leading to unsustainable defaults thus eliminating the utility of the scores.

Re: long-term alignment:

From a pricing standpoint, our original idea to Radiant was that we offer the credit scores via our protocol’s NFCS (NFT representation of credit score) for free. HOWEVER, given we’re on Polygon and Radiant on Arbitrum (BNB incoming), that would increase UX friction. If that is not the case, we are happy to explore that angle if the tradeoff is worth it for Radiant DAO.

Re: open source web3 scoring:

We are that project. Per comments above, fully open source isn’t possible but our scoring is available to any project to query via our open API; as long as they have minted an NFCS previously (UX friction point). Our goal is to share the ‘fire’ with as many lenders as possible to increase capital efficiency throughout DeFi.

And yes, data is transparent but laborious to acquire, clean, and pre-process for human readable tasks. That doesn’t even include the analytics aspect which is where the main value to Radiant resides. Providers of this service will always need to cover expenses either via token or cash.

In short, we believe our goals align with Radiant’s long-term vision. Furthermore, we’re willing to revisit the proposal through a more frictioned but less costly experience if users deem that more satisfactory.

Good responses Chris.

I didn’t mention ArcX as a competitor, they just happened to publish a lot of information about the UX of credit scores when they were very focused on it.

Thinking about myself, if a ML model gave me a low credit score and the solution was to read the blog, it would annoy me. And that was what ArcX surfaced in their customer discovery, which is why they built a very simple model that gave actionable steps. Basically “do this and this and you’re good”.

I didn’t explain in the post, but my sense is that “under-collateralised lending via ML model” is an optimisation once the low hanging fruit of growth is expired. So I think Radiant just has a lot to do yet to deliver on the omni-chain vision without adding complexity to core lend/borrow contracts.

So my feedback is we should deliver on multiple chains working with LayerZero infrastructure, success with dLP plus DAO governance leadership first. This will get good growth without adding to lend/borrow complexity and more UX challenges. Once we’re at that level, ML and under-collateralised lending would be the next leg of growth and we can deal with UX and complexity then.

1 Like

From Chris_RociFi

Understood on ArcX but their product never made it to market. Ours did. With that, we have gotten some negative feedback on our approach for sure, but majority has been positive. Furthermore, when users mint their NFCS (NFT credit score), they are given the same prompts on their profile page as what is driving their score as seen in the blog. The blogs just provide more in-depth details for those that really wanna understand the logic and statistics behind it all.

I just want to make this VERY clear. This proposal is NOT for Radiant to introduce under-collateralized lending. This proposal is for the introduction of credit scores to Radiant for increased capital efficiency and risk management on a pool basis. Example, based on outstanding loan amounts in lending pool A with good scores, LTV can be adjusted from 60% to 62.5% without increasing bad debt risk; or vice versa. The examples in the proposal show how these user-level insights can greatly benefit Radiant’s risk management and capital efficiency. Also, this could be done without touching the smart contracts which is why we chose this as a starting point.

Future product offerings could include customized LTVs for every borrower based on their credit score but would require smart contract integration. Even further beyond that, Radiant could introduce under-collateralized at an even later date. We fully appreciate Radiant’s current roadmap and delivery schedule, which is why we laid out the ‘order of operations’ as we did given it can add value NOW with low hanging fruit, while giving the DAO optionality for future products, as desired.

Hopefully my answers make sense and you get a better understanding of the true spirit of the proposal.

Let’s say I, a single user, have 4 0x’s for different defi strategies;
a) Low Risk High Value only,
b) Medium Risk Medium Value only,
c) High Risk Low value,
d) High Risk Medium Value.

If I apply for a high LTV under-collat loan thru Radiant + RociFi with my (a) 0x, then you’re unaware of just how risky my behavior actually could be. ((1 user))

Radiant would need to CAP the total amount of borrowing available to the Under-collateralized loans Pool, as to avoid the known-unknown of a high volume default happening under these conditions; from a bad actor.

From Chris_RociFi:

Per my other message, I am not proposing under-collateralized loans at this time. Only higher LTVs but still over-collateralized using credit scores on a pool basis. Thus, your comments are not relevant at this time but perhaps in the future. At which time I can comment on how we combat sybil-borrowing such as you described.

This topic was automatically closed after 7 days. New replies are no longer allowed.