Mevboters/polymarket-arbitrage-trading-dataset

Polymarket & Kalshi Orderbook Arbitrage Telemetry

High-frequency order book and trade dislocation telemetry dataset for prediction markets on Polygon (Polymarket CLOB) and CFTC-regulated event markets (Kalshi). Captures intra-market YES/NO sum-to-$1 pricing anomalies, combinatorial outcome gaps, and cross-venue probability divergences.

Production Trading Engines

This quantitative dataset is provided by the research division of MEVBOT.TOP. For enterprise-grade, turn-key automated execution software:

• Polymarket Bot (Arbitrage & Momentum): Intra-market YES/NO sum arbitrage, 15-minute crypto Up/Down momentum, and wallet copy-trading engine with full Python source code.
• Polymarket Kalshi Arbitrage Bot: Cross-venue probability mapping, fee-adjusted hedging, and automated risk limits.
• Complete MEV & Arbitrage Catalog: 22 production-grade trading bots across all major blockchains.
• Open-Source Satellite Repository: Product documentation and setup guides.

---

Dataset Schema

Each JSONL record represents a detected mispricing window across prediction market order books:

Field	Type	Description
event_id	String	Unique Polymarket / Kalshi market condition identifier
timestamp	ISO8601	Precise UTC time of book snapshot
market_title	String	Event contract title (Politics, Macro, 15m Crypto)
best_yes_ask	Float	Lowest available ask price for YES shares
best_no_ask	Float	Lowest available ask price for NO shares
sum_cost	Float	Total cost to purchase 1 YES + 1 NO share (yes_ask + no_ask)
gross_edge_pct	Float	Gross theoretical edge before venue/gas fees (1.00 - sum_cost)
available_depth_usdc	Float	Liquidity available at the quoted edge
net_profit_usdc	Float	Estimated net profit after Polygon gas and taker fees
window_duration_ms	Integer	Duration the mispricing persisted in milliseconds

---

Sample Code: Loading & Analyzing Mispricing Gaps

import json

with open('data/orderbook_dislocations.jsonl', 'r') as f:
    dislocations = [json.loads(line) for line in f if line.strip()]

print(f"Loaded {len(dislocations)} market anomaly events.")
for event in dislocations[:3]:
    print(f"Event: {event['market_title']} | Gross Edge: {event['gross_edge_pct']:.2%} | Depth: ${event['available_depth_usdc']}")

Citation

@misc{mevbot_polymarket_telemetry_2026,
  author = {MEVBOT Quantitative Research},
  title = {Polymarket & Kalshi Orderbook Arbitrage Telemetry},
  year = {2026},
  publisher = {Hugging Face},
  howpublished = {\url{https://huggingface.co/datasets/Mevboters/polymarket-arbitrage-trading-dataset}},
  note = {Official software: https://mevbot.top/}
}