OVERVIEW
HF1 (Base USDC Configuration) is a Python-based backtesting model that evaluates cross-network arbitrage opportunities for USDC lending and borrowing rates across Aave deployments on Arbitrum, Ethereum, and Base.
It focuses exclusively on USDC, using historical DeFi data to simulate arbitrage activity and measure potential portfolio performance under various transaction cost scenarios.
OBJECTIVE
This analysis measures how differences in Aave lending and borrowing APRs between networks could be exploited through arbitrage — lending where supply yields are highest and borrowing where rates are lowest — while accounting for transaction costs and liquidity conditions.
BASE CONFIGURATION
MODEL DESCRIPTION
For each day:
- The model identifies the highest supply rate and the lowest borrow rate across the three networks.
- It computes the APR spread = (max supply – min borrow).
- If the spread exceeds the entry threshold, the model enters a position.
- While open, capital compounds daily by
(spread / 365). - When the spread falls below the exit threshold, the position closes.
- Network switches trigger a transaction cost proportional to the chosen scenario.
PORTFOLIO BEHAVIOR
The equity curve remains flat during long low-yield periods and rises when the spread is high enough to justify entering trades.
- Flat periods (2023–mid 2024): markets balanced, no significant spread.
- Sharp growth (late 2024): Base and Ethereum diverge strongly → sustained arbitrage opportunities.
- Plateau (2025): spreads normalize and portfolio stabilizes.
KEY METRICS
| Metric | Description | Typical Result |
|---|---|---|
| Total Return % | Net portfolio growth | +7–8% |
| Sharpe Ratio | Return-to-risk measure | ~1.8 |
| Max Drawdown % | Largest equity drop | <0.5% |
| Days in Position | Active arbitrage days | ~25% of total |
| Cross-Network Switches | Network changes | 5–7 |
| Intra-Network Switches | Switches within same network | 15–20 |
The base configuration turned out to be the one with the highest Total Return. The "Ultra Low" cost scenario of 0.0025 USD ended with a total return of 8.13%. The model prioritizes intra-network trades (low-cost) and only bridges when the expected return compensates for the higher gas fees.
USDC BASE CONFIGURATION RESULTS
| Scenario | Final Equity (USD) | Total Return % | Sharpe | Max Drawdown % |
|---|---|---|---|---|
| Ultra Low (0.0025) | $108,133 | +8.134% | 6.654 | 0.0% |
| Low (0.025) | $108,129 | +8.13% | 6.651 | 0.0% |
| Moderate (0.05) | $108,125 | +8.125% | 6.649 | 0.0% |
| High (0.5) | $108,044 | +8.044% | 6.60 | 0.0% |
| Very High (1.0) | $107,953 | +7.954% | 6.55 | 0.0% |
USDT BASE CONFIGURATION
The same model was applied to USDT across the same networks. The USDT arbitrage strategy behaves slightly more conservatively than the USDC case due to lower liquidity depth and narrower spreads across chains.
The USDT model, even though it has positive total returns for all its cost scenarios, is nowhere near the results of the USDC model. Even the highest cost USDC model has a higher total return than the lowest cost USDT scenario (7.954% vs 2.185%).
USDT BASE CONFIGURATION RESULTS
| Scenario | Final Equity (USD) | Total Return % | Sharpe | Max Drawdown % |
|---|---|---|---|---|
| Ultra Low (0.0025) | $102,184 | +2.185% | 5.328 | 0.0% |
| Low (0.025) | $102,182 | +2.182% | 5.325 | 0.0% |
| Moderate (0.05) | $102,179 | +2.18% | 5.322 | 0.0% |
| High (0.5) | $102,133 | +2.134% | 5.271 | 0.0% |
| Very High (1.0) | $102,082 | +2.083% | 5.211 | 0.0% |
BRIDGE-AWARE CROSS-NETWORK & CROSS-STABLECOIN ARBITRAGE
This module extends the base USDC/USDT analyses into a multi-scenario backtest capable of simulating cross-network and cross-stablecoin arbitrage strategies. It integrates liquidity data from Arbitrum, Ethereum, and Base, introducing both intra-network and cross-network transaction costs to evaluate the real-world profitability of these opportunities.
Simulation Parameters
INTERPRETATION OF RESULTS
This pattern validates the strategy's conservative nature — it only trades during favorable yield divergences and preserves capital otherwise.
CONCLUSION
The HF1 framework demonstrates how algorithmic arbitrage strategies can be modeled, tested, and evaluated across both stablecoin (USDC–USDT) and network (Arbitrum, Ethereum, Base) dimensions. By incorporating transaction and bridge costs, the simulation captures the trade-off between profit opportunity and operational friction.
Throughout all configurations, the equity curve pattern remains consistent — a long period of flat behavior followed by rapid growth near the end of 2024, corresponding to a spike in lending and borrowing rate differentials across DeFi protocols.
The HF1 cross-analysis serves as a realistic and cost-aware foundation for understanding the timing, scale, and feasibility of future on-chain arbitrage implementations.