polydata5/polymarket-updown-5m-ws

Polymarket Crypto Up/Down 5-Minute Markets: Raw WebSocket Capture

A continuous recording of Polymarket's public WebSocket feeds for the rolling 5-minute "Up or Down" crypto markets: the full order book and every trade of each market, plus the Chainlink price feed the markets settle on. Every frame is stored exactly as received with a local receive timestamp. Nothing is filtered, deduplicated, aggregated or resampled. A new UTC day is added every day.

Coins	BTC, ETH, SOL, XRP, DOGE, BNB, HYPE
Period	2026-08-21 to 2026-11-09 (UTC), 75 days with data
Frames	5,338,120,631
Size	431.8 GB of gzipped JSON Lines
Updated	daily, the morning after each UTC day. Last update 2026-11-10 01:05 UTC

The markets

Every 5 minutes Polymarket opens a new binary market for each coin, for example "Bitcoin Up or Down - October 21, 11:00PM-11:05PM ET" (slug btc-updown-5m-1792638000, where the number is the window start in Unix seconds). That makes 288 markets per coin per day. Each market has two outcome tokens, Up and Down, which trade between $0 and $1 on Polymarket's central limit order book (CLOB). The winning token pays $1. In Polymarket's own words:

This market will resolve to "Up" if the time-weighted average price (TWAP) of Bitcoin, generated by Chainlink, of the time range specified in the title is greater than or equal to the price at the beginning of that range. Otherwise, it will resolve to "Down".

The resolution source is the Chainlink <COIN>/USD 60-second TWAP data stream, which is recorded here too.

What is recorded

Two streams per coin, stored as separate files and dataset configs:

• <COIN>_clob: the CLOB market channel for the Up and Down tokens of the coin's live market. This covers full order-book snapshots (book), every price-level change (price_change), every trade (last_trade_price) and tick-size changes (tick_size_change).
• <COIN>_rtds: Polymarket's Real-Time Data Socket price feed for <coin>/usd. It carries the Chainlink price (crypto_prices_chainlink) and the Chainlink 30 s and 60 s TWAPs (crypto_prices_twap_thirty, crypto_prices_twap_sixty). The 60 s TWAP is what the markets resolve on.

Possible uses include order-book reconstruction, market-making and latency backtests, price discovery between the oracle price and market odds, calibration of short-horizon prediction markets, and end-of-window effects.

How the capture follows markets: the collector follows one market per coin at a time. At each 5-minute rollover it closes the connection and opens a new one subscribed to the new market. The data shows this as disconnect/connect rows and a new conn_id. The switch takes about 0.5 to 12 s, so the first seconds of each market can be missing, while the previous market's last frames run a few seconds into the new window. Each new subscription starts with a full book snapshot of both tokens.

Files

data/<YYYY-MM-DD>/<COIN>_clob.jsonl.gz   order book + trades, one UTC day
data/<YYYY-MM-DD>/<COIN>_rtds.jsonl.gz   Chainlink price + TWAPs, one UTC day
data/<YYYY-MM-DD>/coverage.png           per-minute coverage chart of that day
coverage.png                             coverage chart of the whole dataset (above)
manifest.json                            per day and file: hours, line count, size, missing minutes, repairs

Each daily file is that day's hourly capture files concatenated in time order. The result is a multi-member gzip, which any gzip reader handles, with one JSON object per line. There is one dataset config per coin and stream (e.g. BTC_clob, ETH_rtds) covering all days.

Record format

field	type	meaning
recv_ts	float64	local receive time, Unix seconds (UTC)
recv_mono	float64	collector monotonic clock, seconds (arbitrary origin; for precise intervals)
hour_start	int64	Unix seconds of the UTC hour of the capture file the line came from
frame_i	int64	line index within the hourly capture file (restarts each hour and after a collector restart)
frame_scope	string	scope of frame_i (hour_file)
conn_id	int64	WebSocket connection number within the collector process
conn_i	int64	frame index within that connection
reconnect_n	int64	reconnect counter of the connection
event	string	connect / disconnect / reconnect on connection lifecycle rows, else null
raw	string	the text frame exactly as received (JSON text); null on lifecycle rows

Messages also carry Polymarket's own timestamp (ms) inside raw, so you can measure feed latency against recv_ts.

Message types

raw holds the frame exactly as Polymarket sent it. Shortened real examples:

CLOB (<COIN>_clob). Every message has market (condition id of the 5-minute market) and asset_id (outcome token id). Prices are in USDC per share (0 to 1), sizes in shares, all as strings.

// book: full order book of one outcome token. Sent on subscribe as a JSON array of snapshots, later as single objects
{"market":"0x3dd0…009b","asset_id":"8071…6826","timestamp":"1792638010067","hash":"2eb9…53c6",
 "bids":[{"price":"0.01","size":"10574.26"},{"price":"0.02","size":"3936.12"},…],"asks":[…],"event_type":"book"}

// price_change: new total size at a price level (size "0" = level removed). BUY = bid side, SELL = ask side
{"market":"0xc8dd…00a5","price_changes":[{"asset_id":"3732…8911","price":"0.99","size":"71327.52","side":"BUY",
 "hash":"23f2…5c37","best_bid":"0.99","best_ask":"1"},…],"timestamp":"1792637999671","event_type":"price_change"}

// last_trade_price: a trade print
{"market":"0x3dd0…009b","asset_id":"5134…1540","price":"0.69","size":"14.492754","fee_rate_bps":"0","side":"BUY",
 "timestamp":"1792638010535","event_type":"last_trade_price","transaction_hash":"0xeee9…b095"}

// tick_size_change
{"market":"0x3dd0…009b","asset_id":"8071…6826","old_tick_size":"0.01","new_tick_size":"0.001",
 "timestamp":"1792638301170","event_type":"tick_size_change"}

RTDS (<COIN>_rtds). Chainlink price and 30 s / 60 s TWAP updates, about one per second each. payload.timestamp is the price time in ms, and full_accuracy_value is the value scaled by 1e18.

{"topic":"crypto_prices_chainlink","type":"update","timestamp":1792638000267,"connection_id":"geSy…hA==",
 "payload":{"symbol":"btc/usd","timestamp":1792637999000,"value":82841.44,"full_accuracy_value":"82841440000000000000000"}}
{"topic":"crypto_prices_twap_sixty","type":"update","timestamp":1792638000285,"connection_id":"geSy…hA==",
 "payload":{"symbol":"btc/usd","timestamp":1792637999000,"value":82844.49278626572,"window_s":60,
 "full_accuracy_value":"82844492786265713803264"}}

Each RTDS (re)connection also starts with "type":"subscribe" messages for crypto_prices, crypto_prices_twap_thirty and crypto_prices_twap_sixty, holding the last ~50 s of values in payload.data.

Lifecycle rows have no raw; they only carry event: connect, disconnect or reconnect.

Usage

Stream a whole config with datasets:

from datasets import load_dataset

ds = load_dataset("polydata5/polymarket-updown-5m-ws", "BTC_clob", split="train", streaming=True)

One day as a DataFrame (needs huggingface_hub for hf://):

import pandas as pd
df = pd.read_json("hf://datasets/polydata5/polymarket-updown-5m-ws/data/2026-11-09/BTC_rtds.jsonl.gz", lines=True)

Rebuild the order book from book + price_change. Replaying the deltas this way reproduces the next book snapshot exactly:

import json

books = {}  # asset_id -> {"bids": {price: size}, "asks": {price: size}}

def apply(msg):
    for m in msg if isinstance(msg, list) else [msg]:
        if m.get("event_type") == "book":
            books[m["asset_id"]] = {"bids": {l["price"]: float(l["size"]) for l in m["bids"]},
                                    "asks": {l["price"]: float(l["size"]) for l in m["asks"]}}
        elif m.get("event_type") == "price_change":
            for c in m["price_changes"]:
                if c["asset_id"] not in books:
                    continue
                side = books[c["asset_id"]]["bids" if c["side"] == "BUY" else "asks"]
                if float(c["size"]) == 0:
                    side.pop(c["price"], None)
                else:
                    side[c["price"]] = float(c["size"])

for row in ds:
    if row["raw"] and row["raw"][0] in "[{":
        apply(json.loads(row["raw"]))

Find which market and outcome a token belongs to with Polymarket's public Gamma API:

import json, requests

m = requests.get("https://gamma-api.polymarket.com/markets",
                 params={"condition_ids": market_id, "closed": "true"}).json()[0]
m["slug"], json.loads(m["outcomes"]), json.loads(m["clobTokenIds"])
# ('btc-updown-5m-1792638000', ['Up', 'Down'], ['8071…6826', '5134…1540'])

Coverage

A minute counts as captured if at least one frame of that stream was received in it. Every other minute from 2026-08-21 on is missing, including the days before a coin's capture started. Each day folder has its own per-minute chart, linked from the per-day table below.

series	first frame (UTC)	captured	missing	short gaps (<15 min)	longest gap
BTC_clob	2026-08-21 00:36	74.8%	20 d 10 h 48 min	44 (111 min)	18 d 4 h 39 min
BTC_rtds	2026-09-09 23:45	72.8%	22 d 46 min	3 (15 min)	19 d 23 h 45 min
ETH_clob	2026-08-28 17:29	86.9%	10 d 14 h 5 min	235 (419 min)	7 d 17 h 29 min
ETH_rtds	2026-08-28 17:29	86.9%	10 d 15 h 38 min	265 (451 min)	7 d 17 h 29 min
SOL_clob	2026-09-09 23:45	72.7%	22 d 2 h 5 min	44 (94 min)	19 d 23 h 45 min
SOL_rtds	2026-09-09 23:45	72.8%	22 d 51 min	3 (20 min)	19 d 23 h 45 min
XRP_clob	2026-09-09 23:45	72.7%	22 d 2 h 28 min	51 (117 min)	19 d 23 h 45 min
XRP_rtds	2026-09-09 23:45	72.8%	22 d 52 min	4 (21 min)	19 d 23 h 45 min
DOGE_clob	2026-08-28 17:29	87.1%	10 d 11 h 1 min	250 (430 min)	7 d 17 h 29 min
DOGE_rtds	2026-08-28 17:29	87.0%	10 d 12 h 27 min	247 (427 min)	7 d 17 h 29 min
BNB_clob	2026-08-28 17:29	87.3%	10 d 6 h 27 min	247 (423 min)	7 d 17 h 29 min
BNB_rtds	2026-08-28 17:29	87.0%	10 d 12 h 13 min	261 (451 min)	7 d 17 h 29 min
HYPE_clob	2026-09-09 23:45	72.6%	22 d 3 h 49 min	71 (198 min)	19 d 23 h 45 min
HYPE_rtds	2026-09-09 23:45	72.8%	22 d 51 min	3 (20 min)	19 d 23 h 45 min

Gaps of 15 minutes or more (UTC)

• 2026-08-21 00:00 to 2026-08-21 00:36 (36 min): BTC_clob
• 2026-08-21 00:00 to 2026-08-28 17:29 (7 d 17 h 29 min): ETH_clob, ETH_rtds, DOGE_clob, DOGE_rtds, BNB_clob, BNB_rtds
• 2026-08-21 00:00 to 2026-09-09 23:45 (19 d 23 h 45 min): BTC_rtds, SOL_clob, SOL_rtds, XRP_clob, XRP_rtds, HYPE_clob, HYPE_rtds
• 2026-08-21 15:55 to 2026-08-21 16:20 (25 min): BTC_clob
• 2026-08-22 14:42 to 2026-08-22 15:00 (18 min): BTC_clob
• 2026-08-22 15:29 to 2026-08-22 16:00 (31 min): BTC_clob
• 2026-08-22 16:33 to 2026-08-22 17:00 (27 min): BTC_clob
• 2026-08-22 17:28 to 2026-08-22 18:00 (32 min): BTC_clob
• 2026-08-22 18:17 to 2026-08-22 19:00 (43 min): BTC_clob
• 2026-08-22 19:06 to 2026-09-09 23:45 (18 d 4 h 39 min): BTC_clob
• 2026-08-31 13:22 to 2026-08-31 13:50 (28 min): ETH_clob, ETH_rtds, DOGE_clob, DOGE_rtds, BNB_clob, BNB_rtds
• 2026-09-02 14:34 to 2026-09-02 14:49 (15 min): ETH_clob, ETH_rtds, DOGE_clob, DOGE_rtds, BNB_clob, BNB_rtds
• 2026-09-05 13:20 to 2026-09-05 17:47 (4 h 27 min): DOGE_clob, DOGE_rtds
• 2026-09-05 15:27 to 2026-09-05 17:06 (1 h 39 min): BNB_rtds
• 2026-09-05 16:18 to 2026-09-05 17:19 (1 h 1 min): ETH_rtds
• 2026-09-05 17:50 to 2026-09-05 18:32 (42 min): ETH_clob, ETH_rtds, DOGE_clob, DOGE_rtds, BNB_clob, BNB_rtds
• 2026-09-05 19:15 to 2026-09-06 02:57 (7 h 42 min): ETH_clob, ETH_rtds
• 2026-09-05 19:46 to 2026-09-05 21:15 (1 h 29 min): DOGE_rtds
• 2026-09-05 21:36 to 2026-09-05 23:15 (1 h 39 min): BNB_rtds
• 2026-09-05 23:54 to 2026-09-06 01:15 (1 h 21 min): BNB_rtds
• 2026-09-06 02:18 to 2026-09-06 02:57 (39 min): BNB_rtds
• 2026-09-07 09:46 to 2026-09-07 10:32 (46 min): ETH_clob, ETH_rtds, DOGE_clob, DOGE_rtds, BNB_clob, BNB_rtds
• 2026-09-07 10:33 to 2026-09-07 10:50 (17 min): ETH_clob, ETH_rtds, DOGE_clob, DOGE_rtds, BNB_clob, BNB_rtds
• 2026-09-07 13:47 to 2026-09-07 15:15 (1 h 28 min): ETH_clob, ETH_rtds, DOGE_clob, DOGE_rtds, BNB_clob, BNB_rtds
• 2026-09-07 15:20 to 2026-09-07 15:59 (39 min): ETH_clob, ETH_rtds, DOGE_clob, DOGE_rtds, BNB_clob, BNB_rtds
• 2026-09-08 19:03 to 2026-09-08 19:23 (20 min): ETH_clob, ETH_rtds, DOGE_clob, DOGE_rtds, BNB_clob, BNB_rtds
• 2026-09-09 21:40 to 2026-09-09 22:03 (23 min): ETH_clob, ETH_rtds, DOGE_clob, DOGE_rtds, BNB_clob, BNB_rtds
• 2026-09-09 23:24 to 2026-09-09 23:45 (21 min): ETH_clob, ETH_rtds, DOGE_clob, DOGE_rtds, BNB_clob, BNB_rtds
• 2026-09-11 13:56 to 2026-09-11 14:26 (30 min): BTC_clob, BTC_rtds, SOL_clob, SOL_rtds, XRP_clob, XRP_rtds, HYPE_clob, HYPE_rtds
• 2026-10-03 16:15 to 2026-10-03 16:35 (20 min): all series
• 2026-10-12 13:15 to 2026-10-14 10:47 (1 d 21 h 32 min): all series
• 2026-10-21 10:55 to 2026-10-21 11:15 (20 min): all series
• 2026-11-01 04:56 to 2026-11-01 07:00 (2 h 4 min): all series

Per day: % of minutes captured, clob / rtds

day (UTC)	BTC	ETH	SOL	XRP	DOGE	BNB	HYPE
2026-11-09	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100
2026-11-08	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100
2026-11-07	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100
2026-11-06	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100
2026-11-05	99.9 / 99.9	99.9 / 99.9	99.9 / 99.9	99.9 / 99.9	99.9 / 99.9	99.9 / 99.9	99.9 / 99.9
2026-11-04	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100
2026-11-03	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100
2026-11-02	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100
2026-11-01	91.4 / 91.4	91.4 / 91.4	91.4 / 91.4	91.4 / 91.4	91.4 / 91.4	91.4 / 91.4	91.4 / 91.4
2026-10-31	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100
2026-10-30	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100
2026-10-29	99.4 / 100	99.8 / 99.9	99.4 / 100	99.4 / 100	99.4 / 99.9	99.5 / 100	97.4 / 100
2026-10-28	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100
2026-10-27	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100
2026-10-26	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100
2026-10-25	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100
2026-10-24	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100
2026-10-23	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100
2026-10-22	99.9 / 100	99.9 / 100	99.9 / 100	99.9 / 100	99.9 / 100	99.9 / 100	99.9 / 100
2026-10-21	98.5 / 98.5	98.2 / 98.2	98.2 / 98.2	98.2 / 98.2	98.2 / 98.2	98.2 / 98.2	98.2 / 98.2
2026-10-20	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100
2026-10-19	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100
2026-10-18	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100
2026-10-17	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100
2026-10-16	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100
2026-10-15	100 / 100	99.7 / 100	100 / 100	99.7 / 100	100 / 100	100 / 100	100 / 100
2026-10-14	55.1 / 55.1	55.1 / 55.1	55.1 / 55.1	55.1 / 55.1	54.8 / 55.1	54.8 / 55.1	55.1 / 55.1
2026-10-13 (no data)	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0
2026-10-12	55.2 / 55.2	54.9 / 55.2	55.2 / 55.2	55.2 / 55.2	55.2 / 55.2	55.2 / 55.2	55.2 / 55.2
2026-10-11	100 / 100	100 / 100	100 / 100	99.4 / 100	100 / 100	100 / 100	100 / 100
2026-10-10	100 / 100	100 / 100	100 / 100	99.7 / 100	100 / 100	100 / 100	99.7 / 100
2026-10-09	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100
2026-10-08	99.9 / 100	99.9 / 100	99.9 / 100	99.9 / 100	99.9 / 100	99.9 / 100	99.9 / 100
2026-10-07	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100
2026-10-06	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100
2026-10-05	99.7 / 100	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100	99.7 / 100
2026-10-04	100 / 100	100 / 100	100 / 100	100 / 100	99.7 / 100	100 / 100	99.7 / 100
2026-10-03	98.6 / 98.6	98.6 / 98.6	98.6 / 98.6	98.6 / 98.6	98.6 / 98.6	98.6 / 98.6	98.6 / 98.6
2026-10-02	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100
2026-10-01	99.9 / 100	100 / 100	100 / 100	99.9 / 100	100 / 100	100 / 100	100 / 100
2026-09-30	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100	99.7 / 100
2026-09-29	99.7 / 100	99.4 / 100	99.7 / 100	100 / 100	100 / 100	100 / 100	98.9 / 100
2026-09-28	99.4 / 100	100 / 100	100 / 100	99.4 / 100	100 / 100	100 / 100	99.7 / 100
2026-09-27	99.9 / 100	99.6 / 100	99.6 / 100	99.9 / 100	99.9 / 100	99.9 / 100	99.8 / 100
2026-09-26	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100
2026-09-25	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100
2026-09-24	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100
2026-09-23	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100
2026-09-22	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100
2026-09-21	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100
2026-09-20	100 / 100	99.7 / 99.7	100 / 100	99.2 / 99.9	100 / 100	100 / 100	100 / 100
2026-09-19	99.0 / 100	98.2 / 100	98.3 / 100	98.3 / 100	98.5 / 100	98.6 / 100	95.7 / 100
2026-09-18	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100
2026-09-17	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100
2026-09-16	99.1 / 99.1	99.1 / 99.1	98.8 / 99.1	99.1 / 99.1	98.5 / 99.1	98.8 / 99.1	98.8 / 99.1
2026-09-15	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100
2026-09-14	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100
2026-09-13	99.4 / 100	99.4 / 100	99.7 / 100	100 / 100	100 / 100	99.4 / 100	100 / 100
2026-09-12	99.4 / 100	100 / 100	100 / 100	100 / 100	99.7 / 100	99.7 / 100	99.2 / 100
2026-09-11	96.6 / 97.9	98.6 / 99.9	96.6 / 97.9	96.6 / 97.9	98.6 / 99.9	98.6 / 99.9	96.6 / 97.9
2026-09-10	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100	100 / 100
2026-09-09	1.0 / 1.0	94.3 / 93.8	1.0 / 1.0	1.0 / 1.0	94.2 / 93.8	94.2 / 93.6	1.0 / 1.0
2026-09-08	0 / 0	97.4 / 97.0	0 / 0	0 / 0	97.1 / 96.9	97.2 / 96.4	0 / 0
2026-09-07	0 / 0	84.4 / 83.8	0 / 0	0 / 0	84.3 / 83.8	84.2 / 83.9	0 / 0
2026-09-06	0 / 0	86.2 / 85.6	0 / 0	0 / 0	98.6 / 98.5	98.5 / 90.3	0 / 0
2026-09-05	0 / 0	75.3 / 71.0	0 / 0	0 / 0	76.5 / 69.9	95.1 / 80.6	0 / 0
2026-09-04	0 / 0	98.5 / 97.8	0 / 0	0 / 0	97.9 / 97.8	98.3 / 97.8	0 / 0
2026-09-03	0 / 0	98.2 / 97.6	0 / 0	0 / 0	98.0 / 97.7	98.0 / 97.5	0 / 0
2026-09-02	0 / 0	96.1 / 94.9	0 / 0	0 / 0	96.0 / 94.7	96.0 / 95.3	0 / 0
2026-09-01	0 / 0	97.3 / 96.6	0 / 0	0 / 0	97.3 / 97.0	97.3 / 96.7	0 / 0
2026-08-31	0 / 0	96.7 / 95.9	0 / 0	0 / 0	96.7 / 96.4	96.7 / 96.1	0 / 0
2026-08-30	0 / 0	98.7 / 97.8	0 / 0	0 / 0	98.6 / 98.3	98.6 / 97.9	0 / 0
2026-08-29	0 / 0	99.6 / 99.0	0 / 0	0 / 0	99.6 / 98.8	99.6 / 98.9	0 / 0
2026-08-28	0 / 0	27.1 / 27.0	0 / 0	0 / 0	27.1 / 27.1	27.1 / 26.7	0 / 0
2026-08-27 (no data)	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0
2026-08-26 (no data)	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0
2026-08-25 (no data)	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0
2026-08-24 (no data)	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0
2026-08-23 (no data)	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0
2026-08-22	68.1 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0
2026-08-21	95.8 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0

Collector restarts

These hours were captured in more than one part around a collector restart. The parts are concatenated in time order, and frame_i/conn_id restart in each part.

• 2026-09-09 BNB_clob: 23:00
• 2026-09-09 BNB_rtds: 23:00
• 2026-09-09 DOGE_clob: 23:00
• 2026-09-09 DOGE_rtds: 23:00
• 2026-09-09 ETH_clob: 23:00
• 2026-09-09 ETH_rtds: 23:00
• 2026-09-11 BNB_clob: 13:00, 14:00
• 2026-09-11 BNB_rtds: 13:00, 14:00
• 2026-09-11 DOGE_clob: 13:00, 14:00
• 2026-09-11 DOGE_rtds: 13:00, 14:00
• 2026-09-11 ETH_clob: 13:00, 14:00
• 2026-09-11 ETH_rtds: 13:00, 14:00

Repaired lines

Invalid lines removed from crash-truncated hour files (NUL padding left by a hard crash):

• 2026-10-21 BNB_clob: 1 line(s)
• 2026-10-21 BTC_clob: 1 line(s)
• 2026-10-21 DOGE_clob: 1 line(s)
• 2026-10-21 ETH_clob: 1 line(s)
• 2026-10-21 ETH_rtds: 1 line(s)
• 2026-10-21 SOL_clob: 1 line(s)
• 2026-10-21 SOL_rtds: 1 line(s)

Other known issues

• 2026-09-11 14:00-14:26 UTC (BNB, DOGE, ETH): two collector processes wrote the same hourly file, so frame_i interleaves two sequences and some frames may appear twice.

Source and license

Captured by the dataset author from Polymarket's public WebSocket feeds (CLOB market channel and RTDS). This is a capture of publicly broadcast data, released under CC-BY-4.0 for this compilation. Upstream data remains subject to Polymarket's terms. Nothing here is financial advice.