NegRisk Multi-Outcome Intra-Platform Arbitrage

How It Works Polymarket uses a NegRisk mechanism for multi-outcome markets (e.g., 'Which country wins the World Cup?'). In theory, the sum of all YES prices must equal 1.0. When liquidity is fragmented across many outcomes, temporary mispricings occur where the sum drifts above or below 1.0. This scanner identifies those violations in real time and executes offsetting positions across all outcomes to lock in the spread.

Entry Conditions An arbitrage signal is triggered when the sum of outcome prices deviates from 1.0 by more than the round-trip transaction cost (gas + spread). The asyncio-based Python scanner continuously polls the order book, enabling low-latency detection. The position is to buy all outcomes simultaneously when their sum is below 1.0 (guaranteed payout), or sell all when sum exceeds 1.0.

Exit Conditions Positions resolve naturally at market settlement. There is no active exit — the arb is locked in at entry. The key risk is partial fill: if not all legs fill at quoted prices, the position becomes directional.

Verified Results & Implementation Notes The `matthewnyc2/arbitrage` repo explicitly claims 60 passing tests and a paper-trading mode with a live browser demo, indicating a production-grade implementation rather than a toy. It is written in Python asyncio for concurrency. The repo size (1524KB) is consistent with a real, tested codebase. The '60 passing tests' claim is a meaningful quality signal.

Risks Partial fills destroy the arb. Gas costs and Polymarket fee tiers must be modeled precisely. NegRisk arb opportunities may be extremely short-lived as other bots compete. Execution risk is the dominant concern — theoretical arb does not guarantee realized profit.