FiRM Liquidation Factor Adjustment for Select Volatile Collateral Markets

Summary

Following analysis performed on the October 10th, 2025 liquidation cascade - the largest in cryptocurrency history - the Risk Working Group (RWG) proposes systematic liquidation factor (LF) increases across five volatile collateral markets to align FiRM’s safety parameters with empirical evidence from extreme market stress. Throughout this event, Chainlink oracles operated correctly by design, but network congestion created on-chain publication delays that produced step-wise repricing, compressing multiple minutes of price movement into single oracle updates and overwhelming liquidation execution in markets with lower LF settings.

Proposed Changes:

Market Current LF Proposed LF Change
wstETH 60% 75% +15%
wBTC 60% 75% +15%
wETH 40% 75% +35%
cbBTC 50% 75% +25%
CRV 60% 100% +40%
No Change Current Proposed
INV 50% 50% Maintain
st-yETH 100% 100% Maintain
CVX 100% 100% Maintain
cvxCRV 100% 100% Maintain
st-yCRV 100% 100% Maintain

These adjustments strengthen FiRM’s resilience to oracle latency during extreme market conditions while RWG collaborates with Chainlink to individually backtest each volatile asset’s oracle performance against October 10th data, verifying that fine-tuned OCR (Off-Chain Reporting) configurations produce validated improvements that would successfully pass the stress test conditions observed during the event. To clarify, no user action is required at this time.

By securing existing markets, RWG establishes the foundation to confidently pivot toward growth-facing initiatives.

Background

On October 10, 2025, crypto markets experienced an unprecedented liquidation cascade. Within hours, over $19.2 billion in leveraged positions were liquidated across DeFi and CeFi platforms. Altcoins were significantly impacted, many losing 50-60% of value within 10-minute intervals.

FiRM processed 79 liquidations across three markets (CVX: 72, CRV: 5, cvxCRV: 2), clearing approximately $717,000 DOLA - one of the largest 24-hour liquidation periods in protocol history. CVX dropped 71% in 27 minutes (from $2.94 to $0.85), resulting in $110,310 in bad debt, since repaid through a generous contribution from CVX co-founders C2tp and Winthorpe. RWG published a comprehensive analysis of the CVX/USD Chainlink oracle performance during the event, which led to governance action increasing CVX LF from 60% to 100% to secure the market. Markets with 100% LF (cvxCRV, st-yCRV) experienced zero bad debt despite exposure to similar volatility. FiRM remains free of bad debt and fully operational.

The October 10th liquidation cascade - an isolated, unprecedented market shock - occurred against the backdrop of broader Q4 2025 market deterioration. Throughout October and November, DeFi has experienced a series of high-profile protocol exploits (Moonwell, Balancer) and stablecoin depegs (Elixir, Stream, Yala), with on-chain indicators revealing defensive positioning across DeFi ecosystems. The combination of an isolated extreme volatility event layered on top of already fragile market conditions amplified systemic stress, justifying heightened risk management across longtail volatile collateral.

The Role of Oracle Latency

RWG is conducting comprehensive analysis of Chainlink oracle performance across volatile and stable assets listed on FiRM to understand the relationship between oracle latency and liquidation outcomes during the October 10th event. Chainlink oracles maintained full data integrity throughout the event with no gaps or invalid rounds, demonstrating that the issue was feed provider latency-driven rather than FiRM failure-driven, and responsiveness normalized as volatility subsided. RWG is now working with Chainlink to conduct similar detailed analysis across all other collateral markets to understand oracle behavior patterns and validate whether current LF and CF methodologies provide appropriate safety margins for each asset’s specific oracle configuration and latency profile.

Proposed Parameter Changes

While 100% LF across all FiRM volatile markets is the most protective configuration for protocol safety, borrower experience and loss scaling represent critical trade-offs that must be balanced against risk minimization. In FiRM’s model, the liquidation incentive (10-12% for volatile assets) applies to whatever portion of debt gets liquidated - when LF is 100%, that penalty applies to the entire position, amplifying user loss and typically ending borrower retention altogether. Partial liquidations, by contrast, give borrowers the opportunity to re-collateralize or repay without losing their entire position, maintaining borrower relationships even during market stress.

The broader DeFi ecosystem has been moving toward more borrower-friendly unwind mechanisms. Aave v3 and Spark implement dynamic close factors that liquidate only 50% of positions when health factor exceeds 0.95, escalating to 100% only in deep distress. Llamalend by Curve pioneered soft liquidations that typically unwind just 1-5% of debt to restore solvency, with these designs showing clear improvements in user retention and becoming major competitive selling points. FiRM currently has room to improve user experience through dynamic or tiered LF logic that scales with health factor or asset-specific liquidity profiles, though such enhancements would require FiRM v2 architecture with new market contract implementations.

CRV ( → 100% LF)

The proposed max increase to the CRV market’s LF parameter represents continuation of FiRM’s longtail asset risk mitigation strategy that began with CVX’s post-October 10th correction to 100% LF. The CRV market on FiRM is rich with operational history and battle-tested solvency across multiple stress events; justifying its continuation. AAVE moved CRV to non-borrowable status following the event, setting borrow caps to 1 and LTV to 0%, citing demonstrated oracle risks and low market efficiency. RWG’s analysis and industry findings support maximum protective parameters.

wETH, wBTC, wstETH, cbBTC ( → 75% LF)

These four markets represent FiRM’s highest-quality volatile collateral. Despite varying oracle feed complexity - from wETH’s direct ETH/USD feed (0.5% deviation, 1-hour heartbeat) to wstETH’s triple-feed architecture (stETH/ETH, stETH/USD, ETH/USD) and bridge asset risk considerations with cbBTC and wBTC’s; these assets share a critical characteristic that distinguishes them from longtail collateral: low volatility relative to high liquidity depth. This fundamental property makes them significantly less vulnerable to the granular price latency issues that overwhelmed liquidation execution in longtail markets during October 10th, as their deep order books and slower price movements provide greater time windows for liquidators to execute before positions become underwater. The proposed 75% LF sits on the higher end of industry standards, but FiRM’s architectural differentiators provide borrower-focused safety features that enable competitive market positioning despite conservative liquidation parameters.

No Changes: Validated Parameters

  • INV market maintains 50% LF as appropriate for its unique risk profile as Inverse Finance’s governance token. The current setting recognizes that partial liquidations reduce market impact for a thin-orderbook, protecting against cascading liquidations in negative feedback loops. Currently, the largest position in the market represents ~81.5% of outstanding debt and maintains heavy overcollateralization, making 50% LF safe for the dominant exposure.
  • CVX market maintains 100% LF following its post-October 10th correction.
  • st-yETH market is being offboarded through separate governance action due to sustained deterioration of its liquid value and operational integrity.
  • cvxCRV and st-yCRV markets were proactively upgraded to 100% LF in July 2025 following RWG’s identification of liquidity decline. Both assets lack CEX presence and rely on thinning on-chain Curve pools with deteriorating pegs (cvxCRV at 41%, st-yCRV 52%). During October 10th, both markets experienced zero bad debt despite exposure to volatility, validating the preemptive risk management.

Comparison to Industry Response

AAVE implemented aggressive risk-off measures following October 10th, documented in their governance proposal “ARFC: Deprecation of Low Demand Volatile Assets” authored by Chaos Labs. The DAO voted on setting borrow caps to 1 (effectively non-borrowable) for 14 assets including CRV, UNI, ENS, ARB, BAL, LDO, and 1INCH, while simultaneously setting LTV to 0% (no collateral value) for the same assets. Their rationale centered on oracle deviations of 15-50% during October 10th combined with low market efficiency creating arbitrage exploitation risk.

Chaos Labs’ recent analysis cited CRV experiencing sustained 58% price dislocation during the October 10th window, with Chainlink SVR (Smart Value Recapture) oracle updates lagging by a constant 5 blocks (approximately 60 seconds) throughout the crash period. Their technical framework focused on oracle-DEX price divergence: when Chainlink pricing diverged from on-chain DEX reality, the dislocation enabled arbitrage exploitation. They documented approximately $200K deficit from this mispricing mechanism, where market participants could supply high-LT collateral, borrow CRV at oracle’s understated price, and immediately sell on DEX venues at higher market prices. This arbitrage could be repeated until either prices converged or protocol liquidity was exhausted.

The oracle-DEX divergence is particularly material for protocols like AAVE and Moonwell (as detailed in a report by Anthias Labs) because CRV and other volatiles were previously borrowable on their platform, creating direct arbitrage vulnerability. When oracle pricing lags behind on-chain reality during volatility, borrowers can extract value from the protocol through the price differential. This exploitation mechanism drove AAVE’s decision to make these select volatiles non-borrowable rather than simply adjusting liquidation parameters.

RWG’s analysis approached the same October 10th event from a different angle, focusing on oracle update latency rather than oracle-DEX divergence. The distinction between RWG and Chaos Labs’ analyses reflects different protocol architectures and resulting vulnerabilities. Both studies identify oracle responsiveness issues during extreme volatility but from complementary perspectives - one focused on price divergence exploitation, the other on liquidation execution timing. Chaos Labs’ observation that SVR oracle lag “could have been insufficient to support timely liquidations, thereby publishing the price updates with a consistent maximum allowed lag” aligns with RWG’s analysis of oracle latency overwhelming FiRM liquidation execution. The convergence of findings from independent analyses examining the same event strengthens confidence that oracle responsiveness during extreme network congestion is a systemic concern across DeFi lending protocols, regardless of specific architecture differences.

Future Considerations

This proposal represents immediate market securing via Liquidation Factor optimization based on October 10th empirical evidence. RWG’s near-term focus following this proposal will be comprehensive collateral factor analysis, particularly for stablecoin LP markets, as previewed in the CVX post-event analysis. Oracle latency during extreme volatility affects not only liquidation execution but also the appropriate collateral factor settings that determine borrowing capacity. The same October 10th oracle performance data being gathered for liquidation factor validation will inform systematic review of collateral factors across FiRM markets, with stablecoin LPs receiving priority assessment given their 97% concentration of DOLA backing.

Longer-term improvements would require FiRM v2 architecture with new market contract implementations. Dynamic liquidation mechanisms would eliminate the tradeoff between safety and user experience inherent in static LF settings, with soft liquidations enabling gradual position reduction and sliding scale factors adjusting automatically based on real-time volatility. Hybrid oracle systems incorporating pull-based pricing from sources like Pyth or Redstone alongside Chainlink push-based feeds would provide redundancy during network congestion. Multi-oracle validation with automated health monitoring could enable real-time parameter adjustments based on oracle performance degradation.

Chainlink is providing detailed oracle performance data for all volatile collateral markets during the October 10th event and historical patterns, which will be incorporated into ongoing risk assessment. This quantitative evidence will enable data-driven validation of proposed parameters and inform future OCR configuration optimizations that RWG is coordinating with Chainlink to pass October 10th stress test conditions.

Completing this liquidation factor optimization during Q4 market deterioration positions FiRM strategically for future growth. By ironing out parameter vulnerabilities and validating oracle feed integrity across stressed conditions now, RWG can confidently reallocate focus from defensive market securing toward growth-oriented initiatives once analysis concludes and markets stabilize. The comprehensive oracle performance data being gathered establishes quantitative frameworks for evaluating collateral factor increases where appropriate and assessing new collateral candidates with validated risk assessment methodologies. This measured approach - securing existing markets first, then expanding thoughtfully from a position of strength - ensures FiRM scales responsibly without compromising the protocol stability that has differentiated FiRM throughout industry-wide stress events.

On-Chain Actions

  • Set Liquidation Factor for wETH market to 75%
  • Set Liquidation Factor for wstETH market to 75%
  • Set Liquidation Factor for wBTC market to 75%
  • Set Liquidation Factor for cbBTC market to 75%
  • Set Liquidation Factor for CRV market to 100%
1 Like