10.1 LMSR-Specific Risks
Protocol Subsidy Risk (Eliminated with Phased Dynamic LMSR)
Traditional LMSR Risk:
LMSR typically operates as a subsidized market maker, meaning the protocol can lose money if traders collectively profit against the market maker.
Path Protocol's Solution:
Path Protocol implements a Phased Dynamic LMSR system that eliminates protocol subsidy risk:
BOUNDING Phase:
Only BUY orders allowed
Pool accumulates from trader deposits
No protocol capital required
If market never activates: Full refund (REFUNDED status)
ACTIVE Phase:
Dynamic virtual_b always equals real_pool
Prices reflect true liquidity
Settlement uses pool (trader money)
Winners paid by losers (zero-sum)
Protocol never adds money
Settlement Risk: ZERO
Mitigation:
Phased architecture eliminates protocol capital requirements
Dynamic virtual_b ensures prices match pool
BOUNDING timeout protects early traders (full refund)
Parimutuel settlement ensures zero-sum (winners from losers)
No protocol reserves needed
Residual Risk:
Zero settlement risk. Only trading phase risk exists (bounded and manageable), but even this is minimized as virtual_b dynamically adjusts to match real_pool.
Compute Optimization Complexity
Risk:
LMSR requires exponential and logarithmic calculations, which are computationally expensive on Solana.
Issues:
Approximation errors in exp/ln calculations
Potential compute unit (CU) limit exceedance
Precision loss in fixed-point arithmetic
Mitigation:
Extensive testing of math library (>10,000 test cases)
Taylor series approximation with 10 terms (<0.01% error)
Lookup tables for common values
Continuous monitoring of CU usage
Residual Risk:
Edge cases with extreme parameter values could cause transaction failures or slight pricing inaccuracies.
Parameter Selection Risk
Risk:
Choosing inappropriate liquidity parameter target_b can affect market transition timing and initial liquidity.
Too Low target_b:
Market transitions to ACTIVE quickly
But may have limited initial liquidity
Prices may have higher impact initially
Too High target_b:
Market may take longer to reach ACTIVE
Risk of BOUNDING timeout (full refund)
But provides deeper liquidity once ACTIVE
Mitigation:
Recommended target_b ranges by market type:
Binary markets: 500-2000 USDC
Multi-outcome (3-5): 1000-2000 USDC
Multi-outcome (10+): 2000-5000 USDC
Dynamic virtual_b in ACTIVE phase adapts to actual pool size
Governance can adjust target_b if needed
BOUNDING timeout protects traders if market never activates
Residual Risk:
Low. Even if target_b is suboptimal, dynamic virtual_b in ACTIVE phase ensures prices always reflect true liquidity. BOUNDING timeout provides safety net for unsuccessful markets.
Last updated