Trades - Prediction
The BMLL Trades dataset for Prediction Markets is the definitive factual record of executed activity across all four CME prediction-market venues. It captures every print (including corrections, amendments and cancellations) and lines them up under a single normalised schema that matches BMLL Trades for futures and options.
Because the price of a binary contract is the implied probability of the outcome, each trade is also a direct observation of what real capital believed at that moment. That makes the Trades package the cleanest source of probability time-series data available in regulated US markets.
Product Overview
- Price, size and execution timestamp for every trade, on every event contract, every day
- Exchange trade-condition codes preserved alongside the normalised representation
- Amended and cancelled trades flagged and retained, with no silent rewrites of history
- Aggressor side identified using BMLL Level 3 order book reconstruction
- Settlement-date conventions handled for sports contracts (weekend prints arrive in the Monday file)
- Timestamps at or exceeding venue matching-engine precision (nanosecond or microsecond)
Example Use Cases
- Probability Signal Research — Treat the prints as a time series of implied probabilities. Calibrate models, study how priced probability drifts before macro releases, earnings windows, and sporting events.
- Cross-Asset Alpha — Compare the binary print path against the underlying CME futures (ES, NQ, CL, GC, BTC, EUR/USD) for lead-lag and dislocation signals.
- Event-Driven Backtesting — Build histories of Fed decisions, payrolls, CPI, earnings, sporting outcomes, all with regulated, exchange-grade prints.
Product description
| Asset Classes | Prediction contract |
| Coverage | CME Event Contracts across XCBT, XCEC, XCME and XNYM |
| Data available from | December 2025 |
| Delivery mechanisms | API / SFTP / S3 / Azure Storage / Google Cloud Storage / Snowflake / Databricks |
| Available via | BMLL Data Feed and BMLL Data Lab |