See the data before you subscribe.
Explore actual Market Events v1 rows from a recorded observation and its market context to the reference-price movement that followed. These selected preview rows demonstrate the schema and missing-data treatment, not representative trading performance.
Actual sample observations
6 preview rows from 2026-09-13; full daily release: 120 observations. Built 2026-09-13 14:03:21 UTC. Calculation event-study-six-spot-1.2.
| Market | Observation (UTC) | Type / trigger | Feature availability | 1h price return / status |
|---|---|---|---|---|
| DOGEUSDT | 2026-09-13 13:30:55 UTC | event · bollinger | locally_known | — · pending |
| DOGEUSDT | 2026-09-13 13:30:55 UTC | event · volume | locally_known | — · pending |
| XRPUSDT | 2026-09-13 13:30:55 UTC | event · bollinger | locally_known | — · pending |
| BTCUSDT | 2026-09-13 14:00:00 UTC | control · Hourly control | locally_known | — · pending |
| ETHUSDT | 2026-09-13 14:00:00 UTC | control · Hourly control | locally_known | — · pending |
| SOLUSDT | 2026-09-13 14:00:00 UTC | control · Hourly control | locally_known | — · pending |
These rows show real schema and missingness, not a representative performance sample. Retrospective context was acquired later; pending and gap outcomes must not be replaced with zeros.
- Events are recorded public rule triggers; controls are scheduled hourly observations, not matched controls or proven negative signals.
- Retrospective features were recovered later and must not be treated as known at the event time.
- Forward returns use the next full five-minute candle open. Missing and pending outcomes are not zero. Fees, fills and slippage are not modelled.
- CSV for spreadsheets, Parquet for notebooks, JSON for API queries. Fixed plans include this dataset, not the internal archive.