01 / 数据集01 / DATASET
一份记录 BTC 涨跌市场微观结构的归档。A documented archive of BTC Up/Down market microstructure.
数据集记录 Polymarket Bitcoin 涨跌市场的 CLOB 报价和盘口深度，覆盖 5分钟与15分钟窗口。下载前先用这一页了解文件和关联方式。The dataset records CLOB quotes and depth for Polymarket Bitcoin Up/Down markets with 5-minute and 15-minute windows. Use this page to understand the files before downloading them.
采集流程Collection pipeline
采集器接收 CLOB WebSocket 更新，维护本地 order book，并写入多个层次的数据，方便审计和分析：The collector receives CLOB WebSocket updates, maintains a local order book, and writes several layers for inspection and analysis:
CLOB WebSocket → local book → JSONL → Parquet → daily ZIP
5分钟和15分钟市场使用同一套数据模型，通过市场时长和 event slug 判断一行数据属于哪个窗口。The 5-minute and 15-minute markets share the same data model. Their market duration and event slug identify which window a row belongs to.
文件层次File layers
| Layer | Typical path | Purpose |
|---|---|---|
| 实时 JSONLLive JSONL | data/live/*.jsonl | 追加式采集输出，适合审计和增量处理。Append-oriented collection output; useful for audit and incremental processing. |
| 小时 ParquetHourly Parquet | data/hourly/*.parquet | 列式文件，适合更快地查询和研究。Columnar files for faster research queries. |
| 每日 ZIPDaily ZIP | data/daily/btc_*_YYYYMMDD.zip | 按日分发的数据包。Distribution package containing the day’s files. |
| 市场映射Market map | seen_markets.jsonl | 将 event_slug 映射到 market_id、condition_id 和 UP/DOWN asset ID。Maps event_slug to market_id, condition_id, and UP/DOWN asset IDs. |
| 结算结果Outcomes | market_outcomes.parquet | 用于评估和回测评分的事后胜负标签。Post-event winner labels for evaluation and backtest scoring. |
文件如何关联How files link together
研究时的主键是 event_slug + ts。用 event_slug 把报价和深度行关联到市场元数据与结算结果。The primary research key is event_slug + ts. Use event_slug to join quote and depth rows to market metadata and the resolved outcome.
Asset ID 保存在市场映射文件中，而不是重复写入每一行 ticks 和 depth。将文件转换为自己的表时，请保留这层关联。Asset IDs are stored in the market map rather than repeated in every ticks and depth row. Preserve this join when converting the files into your own tables.
查看字段说明Open the field reference这份归档不代表什么What the archive does not claim
报价不是成交Quotes are not trades
这是盘口记录，不是完整的逐笔成交明细，也不能保证展示出来的流动性真的被成交。The data is an order-book record. It does not contain a complete trade-by-trade tape or guarantee that displayed liquidity was executed.
深度不是成交承诺Depth is not a fill promise
展示数量可能变化、消失，或者排在你的模拟订单前面。成交假设应该写进回测模型。A displayed size can change, disappear, or be ahead of your simulated order in the queue. Fill assumptions belong in the backtest model.