EthGPrague $tayRich - Stablecoin Yield-Generating Solution for holded payments applyed in a Booking Mini-App

Everyday DeFi

When day did you realized you were leaving yield on the table when booking your well deserved vacations? The worst, same on the same side, service providers are well getting paid in average 2 months after the service provided. However, anticipating is both good for the tranquility of both. To sum up:

• Service providers often wait up to two months for payments
• Centralized entities capture yield from anticipated payments
• Customers lose potential yield on their advance payments
• Web2 users face complexity in participating in Web3 solutions
• No incentive for early payments in group bookings, mostly for the one nice human that paid everything and finishes up never seeing the money back.

Overview

$tayRich is a Worldcoin mini-app, human-only booking mini-app that restores capital efficiency to customers by redistributing yield from anticipated payments up to the booking date. Bookers receive their fair share of low-risk, risk-adjusted yield—previously captured by centralized entities.

Solution

• Immediate yield generation from day one of booking
• Fair distribution of yield to all participants
• ZK-proof of sufficient funds for Web2 users
• Incentivized early payments
• Guaranteed on-time payments for service providers

Implementation

Core

• Beraborrow CDP positions for yield generation
• NECT stablecoin for risk-adjusted returns
• Revolut web proofs for Web2 user verification

Key Features

1. Early Yield Generation: Pool starts generating yield from day one, even with a single member payment
2. Risk-Adjusted Returns: Utilizes Beraborrow CDP positions and NECT stablecoin
3. Web2 Integration: Revolut web proofs for non-Web3 users
4. Human Verification: Worldcoin-powered human-only system

Yield generator CDP positions tokenized strategy

Risk-Adjusted Yield Score Computation

1. Horizon Scaling

To compute the adjusted APR for a given investment horizon We have figure out that we had a rare information; the T date. We have levaraged the fact that we know when the withdraw will be done, so we can find the best option on the risk ajusted on the Liquid Stability Pool, using it as the default benchmark. $$ \Delta = T - t_{now} $$

$$ APR_{\Delta} = (1 + \mu)^{(\Delta/365d)} - 1 $$

2. Downside volatility (sortino numerator)

Scales volatility for the investment horizon:

$$ \sigma_{\Delta}^- = \sigma^{(\Delta/365d)} $$

3. Expected loss from emergency

We have estimated the expected loss from liquidation risk using a log-normal approximation inspired by the Black-Scholes-Merton model. For having the probability of breaching the liquidation threshold before a keeper can act, based on the volatility over the investment period.

Liquidation Probability:

$$ p_{\text{liq}} = \Phi\left( \frac{\ln(1 - \frac{\text{CR} - \text{CR}_{\text{liq}}}{\text{CR}}) + \frac{1}{2} \sigma^2 \cdot \frac{\Delta}{365}}{\sigma \cdot \sqrt{\frac{\Delta}{365}}} \right) $$

Where:

• Φ(·) is the standard normal cumulative distribution function (CDF)
• CR = current collateral ratio
• CR liq = liquidation threshold
• σ = annualized volatility of the collateral asset
• Δ = time horizon in days

Expected liquidation loss

Once p_liq is estimated, we compute the expected liquidation loss:

$$ L_{liq} = p_{liq} \times 0.4 $$

Where 0.4 is in this example, the assumed haircut in case of emergency liquidation due to slippage, keeper delay, or adverse price execution.

4. Correlation Haircut

Penalizes vaults that move closely with the baseline (lack of diversification):

$$ H_{\rho} = 1 - |\rho| $$

5. Final Risk-Adjusted Yield Score

$$ S = \frac{(\mu_{\Delta} - \mu_{\Delta,SPL} - L_{liq})}{\sigma^-_{\Delta}} \times (1 - |\rho|) $$

A higher score indicates a more favorable risk-adjusted yield.

##Implementation:

1. Risk-Adjusted Tenor Picker:

Inputs: Δ = T – now, metrics from APY Oracle.

For every MLV we fetch:

• μ = forward net APR
• σ = annualised NAV-vol
• p_liq = 7-day liquidation probability (ICR < 200 %)
• ρ = 30-day correlation with the SPL benchmark (sNECT auto-compound)

Score:

$$ S = \frac{Expected\ Yield - Benchmark\ Yield - (Liquidation\ Risk \times 0.4)}{Downside\ Volatility} \times (1 - |Correlation|) $$

We pick the vault with max S; if S ≤ 0 we default to SPL.

Quick rules of thumb (still apply after score test):

• Δ ≤ 7 d → low-drift stable MLV
• 7 < Δ ≤ 30 d → LST-based MLV
• Δ > 30 d → LP/PoL MLV

1. YieldEscrowVault: RC-4626 wrapper that:

   1. Accepts deposits.

   2. Performs a single flash-deposit into the chosen MLV. For lack of time instead of looking for the best optimized proportion, we have leaveraged the long proven market startegy 40/60 portfolio. Inserting a 40 percent of our best solution "risky assets", with the remaining 60 on the Liquid Stable Pool.

   3. Issues Escrow-Shares (NFT or ERC-20) that automatically mature at date T.

   4. Records Δ and the vault address on-chain for auditability.

2. Zero-liquidation loop inside MLV Core: The selected Managed Leverage Vault handles auto-compounding, leverage control, and enforces ICR ≥ 200 % via Beraborrow keepers.

Next: 4. Multi-Sink NECT Deployment:

MLV shares emit NECT → routed by our **StrategyRouter**.

5. Just-in-time exit engine:

   24 h before T we :

   1. redeemIntent queued (zero exit fee).

   2. keeper (own server) callsexecuteWithdrawalEpoch

   3. unwrap to USDC

   4. send USDC to the service provider and harvested yield delta to the traveller

Team

• Hemang Vora
• Tanguy Vansnick
• Carles Cerqueda

Future Plans

• Expand to more service providers
• Enhance yield optimization strategies
• Improve Web2 user onboarding experience
• Add more payment verification methods

License

This project is open source and available under the MIT License.

With love, made in Prague