Research Summary
The report presents the fifth progress update on the Arcadia Collateral Risk Simulation, focusing on a detailed analysis of the Simulation Output for RETH vs USDC. It also discusses the approach toward developing recommendations for protocol parameters, including Liquidation Factor, Collateral Factor, and Exposure.
Key Takeaways
Simulation Outputs: Single Asset RETH vs USDC
- Simulation Dynamics: The simulation output explored the dynamics of Single asset RETH versus USDC pair across a period from “2023-11-10 00:00:00” to “2023-11-10 12:00:00”. It featured 62 accounts and one liquidator and utilized an empirical borrower initialization model to simulate debt scenarios.
- Account Health: The number of healthy accounts remained constant and unaffected during the simulation period for all parameter sets, indicating a stable environment.
- Liquidation Events: There were almost no accounts being liquidated throughout the simulation timeframe for all the parameter sets, suggesting a low risk of liquidation.
- Active Auctions: Active auctions occurred at the start of the simulation period, with the frequency and magnitude of these spikes varying across the different parameter sets.
Outstanding Debt in Auctions
- Debt Distribution: The heatmap provided a visual representation of the distribution and intensity of debt in auctions over time, allowing for a quick assessment of which parameter sets and time intervals are most associated with auction-related debt activity.
Position-weighted Collateral Ratio
- Collateral Health: The Position-weighted Collateral Ratio (wCR) over time indicates the health of borrowed positions within the simulation, representing the total value of collateral posted against the amount borrowed. Each parameter set exhibited a unique pattern of wCR fluctuation.
Net Insolvent Amount
- Insolvency Patterns: The Mean Net Insolvent Amount spiked at the beginning of the observed time frame, then dropped sharply to a lower level or zero, indicating that the mean insolvency across accounts decreases significantly after the initial period.
Protocol Revenue
- Revenue Generation: Different parameter sets resulted in different revenue generation patterns, with steps where the revenue jumps at certain intervals, suggesting discrete periods of liquidation events.
Actionable Insights
- Refinement and Fine-Tuning: The simulation model can be enhanced by improving visualization tools for clearer data interpretation, updating empirical borrower models to allow multi-asset distribution, and integrating dynamic slippage models to better represent market impacts.
- Extended Simulation Runs: Conducting additional simulations with multi-asset scenarios can help assess inter-asset dynamics and systemic risks.
- Sanity Checking: Collaborating with a Curve Researcher for a thorough review of the models can ensure accuracy and consistency with empirical market behaviors.
