A Multi-Agent Polymarket Strategy for Detecting Capital Flows and Managing Risk

Summary

This article presents an automated strategy for binary event contracts on Polymarket. A scheduled workflow first filters markets by liquidity, activity, spread, and price range, then examines hourly candles for patterns such as gradual price advances, rising volume, narrowing pullbacks, breakouts, and volume surges. Candidates with multiple anomalies proceed to news search and a set of isolated AI roles that assess market activity, news support, and implied probability. A final decision stage buys only when the combined assessment indicates undervaluation; the execution layer selects the strongest candidate and uses a limit order with a timeout.

A separate monitoring path handles redemptions and trailing stops, with an entry-price floor as a fallback. The article reports one run in which screening reduced 141 candidates to three, with one reaching execution, and describes false breakouts as a known issue. This is an illustrative observation, not evidence of durable profitability. The author characterizes the system as an early framework; binary contracts can lose their full value, and signals, news analysis, parameter choices, and near-expiry probability estimates all have limitations.

Key ideas

• The strategy screens prediction markets before applying price and volume anomaly tests to identify possible capital flows.
• News analysis and price analysis are assigned to separate AI roles before a final decision combines their conclusions.
• The execution logic selects one highest-ranked buy candidate and submits a limit order with a timeout.
• A separate process manages redemptions and trailing stops with an absolute fallback level.
• The reported example is limited evidence, and binary contracts carry the risk of losing their full value.

Tags

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.