Executive Summary
JIL Sovereign's adaptive enforcement system applies graduated trading restrictions based on the composite entity-level toxicity score. Rather than binary allow/deny decisions, the system uses four enforcement tiers that progressively limit a toxic trader's ability to extract value from liquidity providers.
Four Enforcement Tiers
| Tier | Score Range | Restrictions | Purpose |
|---|---|---|---|
| LOW | 0.0 - 0.3 | Monitoring only, full market access | Normal trading, data collection |
| MODERATE | 0.3 - 0.6 | Spread widening (2x-5x), priority reduction | Reduce edge extraction profitability |
| HIGH | 0.6 - 0.85 | Position size reduction, RFQ-only routing | Force institutional pricing, limit impact |
| CRITICAL | 0.85 - 1.0 | Complete blocking from AMM pools | Full protection of LP capital |
Spread Widening Mechanism
At MODERATE tier, the system widens the effective spread seen by the toxic entity. The widening factor is computed as a function of the toxicity score:
spread_multiplier = 1.0 + (toxicity_score - 0.3) * 13.3
// At score 0.3: multiplier = 1.0x (no widening)
// At score 0.45: multiplier = 3.0x
// At score 0.6: multiplier = 5.0xThis makes toxic flow progressively less profitable while still allowing the entity to trade. If the entity's behavior improves, the score decays and widening reduces automatically.
RFQ-Only Routing
At HIGH tier, orders from the toxic entity are routed exclusively through the RFQ (Request-for-Quote) system rather than the AMM pools. Market makers in the RFQ system can price in the entity's toxicity and adjust their quotes accordingly, providing natural market-based protection for LP capital.
Score Decay and Rehabilitation
Toxicity scores use exponential decay weighting over a 7-day rolling window. An entity that stops toxic behavior sees its score naturally decline:
- After 1 day of clean behavior: ~15% score reduction
- After 3 days: ~50% score reduction
- After 7 days: ~90% score reduction
This encourages behavioral improvement without permanent punishment.