BTC FILE AVAILABILITY79 days of capture
97.52%22,093 / 22,654 slots* Final bar is a partial two-day window. File availability does not imply uniform tick density inside every file.
ARCHIVE PREVIEW
INDEPENDENT ARCHIVE · ACCESS SUBJECT TO RIGHTS & REVIEW
HISTORICAL PREDICTION-MARKET DATA · CRYPTO + WEATHER
See the market
before the outcome.
Structured Polymarket observations that preserve the state of the book, the timing of the capture, the final outcome, and the evidence needed to judge data quality—organized for analysis and simulation.
APACHE ARROW
ROUND-LEVEL FILES
QUALITY FLAGS
ONGOING CAPTURE
53,432STANDARDIZED ROUND FILES
22,093DETAILED BTC ROUNDS
7CRYPTO ASSETS
DAILYCAPTURE CONTINUES
AUDITED SNAPSHOT · 10 AUG 2026 · COVERAGE VARIES BY COLLECTION, DATE, FIELD & CAPTURE QUALITY
/// 01 — COVERAGE
A growing archive,
described honestly.
Coverage is not uniform. Crypto and weather have different market structures, capture histories, fields, and quality constraints. Any future release will document those differences instead of flattening them into a misleading single completeness number.
FINALIZED FIVE-MINUTE ROUND FILESCoverage by asset
24 MAY → 10 AUG 2026
ASSETSTARTCAPTURE WINDOWROUNDSSLOTS
BTC24 MAY22,09397.52%
ETH18 JUL6,49398.02%
SOL22 JUL5,37799.37%
XRP22 JUL5,37899.39%
BNB24 JUL4,90999.09%
DOGE25 JUL4,57898.49%
HYPE25 JUL4,60499.05%
WEATHER · EARLY CAPTUREDaily temperature markets
New York City highest- and lowest-temperature contracts, market states, forecast context, observations, and resolution evidence.
≤ 79°F0.08
80–81°F0.19
82–83°F0.31
84–85°F0.25
≥ 86°F0.17
ILLUSTRATIVE DISTRIBUTION · NOT AN ARCHIVED RECORD
COLLECTION_01ACTIVE
Five-minute crypto
BTC, ETH, SOL, XRP, DOGE, BNB and HYPE round history with market-state observations, outcomes, timing, and capture-quality evidence. Detailed schemas and source density vary by period.
- BTC detailed capture
- 24 May 2026 → present
- Additional BTC history
- Earlier, mixed schemas
- Other crypto assets
- July 2026 → present
- Cadence
- Five-minute markets
COLLECTION_02NEW
Weather markets
A new collection beginning with New York City daily highest- and lowest-temperature markets, pairing market observations with reference-weather evidence and settlement context.
- Initial location
- New York City
- Market families
- Daily high & low
- Collection start
- 10 August 2026
- Status
- Early capture
WHY THE QUALIFIERS MATTER
A timestamp in the inventory does not guarantee a complete event tape, full-depth book, or independently observed value for every second. Coverage will be published by collection, date, field, source type, and observation quality before access is offered.
/// 02 — DATA COLLECTIONS
Different markets.
One evidence standard.
| COLLECTION | DATA CONTEXT | PLANNED DOCUMENTATION | STATUS |
|---|---|---|---|
| CRYPTO_5M | Order books, probability movement, underlying-market context, outcomes | Schema, provenance, observation age, resampling and missingness | MATURE CAPTURE |
| WEATHER_NYC | Temperature ranges, market repricing, forecast and station evidence | Market lifecycle, source timestamps, reference observations, settlement notes | EARLY CAPTURE |
| FUTURE_COLLECTIONS | Additional event categories only when capture and rights are supportable | Published separately rather than silently merged into existing schemas | PLANNED |
ARROW FILE ANATOMY
One round.
One round.
Nested evidence.
A detailed crypto round is represented as a top-level record containing identity, timing, outcome context, and nested observation families. The format keeps captured snapshots distinguishable from normalized fixed-time rows.
outcomedepth_btc_5m_preview.arrow
├─TOP LEVELidentity · timing · outcome · metadata
├─LIVE TICKScaptured snapshots · books · probabilities
├─DECISIONScapture and quality observations
├─SECONDS300 normalized second elements
└─WS EVENTSoptional nested event arrays
1TOP-LEVEL ROW31TOP-LEVEL FIELDSNESTEDLISTS + STRUCTS
FORMAT-FAITHFUL SYNTHETIC EXCERPTDECODED OBJECT
[
{
"market_timestamp": 1786376700,
"market_symbol": "btc-updown-5m-1786376700",
"market_start_utc": "2026-08-10T15:45:00Z",
"market_end_utc": "2026-08-10T15:50:00Z",
"price_to_beat": 118642.31,
"outcome": "Up",
"live_tick_snapshots": [
{"epoch":1786376701,"second":1,
"time_remaining":299,
"up_probability":0.465,
"down_probability":0.535,
"up_bid":0.455,"up_ask":0.475,
"down_bid":0.525,"down_ask":0.545},
… additional captured snapshots
],
"live_decision_snapshots": […],
"seconds": [
{"second":0,"up_probability":0.465,…},
… 299 more normalized elements
],
"ws_events_present": true
}
]Synthetic structure preview. Restricted, private, execution, wallet, strategy, and credential fields are excluded.
market · asset · token IDs · interval
event time · capture time · source age
best bid · best ask · midpoint · spread
price levels · sizes · side · book age
missingness · stale flags · fallbacks · schema
status · winning side · settlement context
/// 03 — METHOD
Reliable data starts with
what was observed.
OutcomeDepth separates captured values, resampled values, derived features, and final outcomes. The goal is not to make every row look complete—it is to make the evidence behind each row clear.
Observation timing
Preserve event time, capture time, source age, and the difference between live observations and fixed-grid representations.
Market structure
Retain visible bids, asks, spread, depth, and market state where the source and redistribution status permit.
Provenance
Identify source families and transformations instead of presenting every calculated field as raw market data.
Outcome evidence
Track provisional and final resolution context, including revisions where they can be verified.
Quality flags
Surface missing books, stale observations, fallback values, schema changes, and heterogeneous capture density.
Neutral tooling
Keep private strategies and execution history separate from the customer-facing data schema.
Timestamped snapshots and event messages actually received.
RAW EVIDENCEConsistent field names, market sides, and time alignment.
ANALYSIS READYResampled or forward-filled rows, explicitly identified.
NOT NEW OBSERVATIONSResolution status and settlement evidence attached later.
REVISION AWARE/// 04 — ANALYSIS WORKFLOWS
Use the stack
you already know.
Work with Arrow data in Python, Rust, SQL, or any other environment that supports columnar files. Filter markets, replay rounds, compare outcomes, test execution assumptions, and inspect data quality without relying on a proprietary viewer.
Build notebooks, charts, diagnostics, and simulations.
Create fast custom analysis and simulation tooling.
Explore local files with columnar query engines.
Use any Arrow-compatible workflow that fits the question.
SIMULATEReplay five-minute paths and test execution assumptions
COMPAREMeasure spread, depth, probability movement, and outcomes
AUDITSeparate observed, normalized, resampled, and resolved layers
BUILDCreate data tools around customer-supplied or properly licensed inputs
/// 05 — LIMITATIONS
What this project
does not claim.
Historical market observations can support analysis, but they are not a guarantee of live execution, market completeness, future performance, or the accuracy of an underlying event outcome.
NO_01
No claim of uniform tick-by-tick coverage across every market and date.
NO_02
No private wallet, execution, balance, strategy, preset, or credential data in any planned release.
NO_03
No affiliation with or endorsement by Polymarket, any exchange, oracle, weather provider, or data vendor.
NO_04
No investment, financial, betting, legal, or personalized trading advice.
NO_05
No promise of dataset delivery until upstream rights and buyer eligibility are confirmed.
/// 06 — CONTACT
Looking for
specific data?
Tell us which collection, period, fields, and intended use matter to you. Join the Discord or contact the OutcomeDepth Team through Telegram or email.
NON-BINDING ENQUIRIES ONLY · NO PAYMENT OR DATA DELIVERY AT THIS STAGE