OmensHow We Build 1-Second OHLCV Candles From Polymarket Trades
Follow the Omens candle methodology: outcome-token identity, execution timestamps, decimal volume, deterministic ordering, corrections and known collection gaps.
Explore Polymarket 1s OHLCV data
Omens · Updated
A one-second chart is useful only if its prices, sizes and timestamps mean what the reader expects. This guide describes how Omens constructs execution candles and how to reproduce basic checks on our frozen BTC Up/Down sample.
The sample is an observation dataset. It is not an assertion that every Polymarket execution was captured. Its metadata documents the source, capture time, timestamp basis and known gaps.
Start with an outcome token and execution event time
Up and Down have different outcome tokens. We aggregate them separately; combining their prices would describe neither instrument. Within a token, only active execution records contribute to a candle.
For a 1s bucket, Omens floors the execution timestamp in milliseconds to a UTC second. The bucket includes its opening timestamp and excludes the start of the following second. Received time describes when an observation arrived, not necessarily when the execution happened.
bucket_ms = floor(event_ms / 1000) × 1000
Order executions before choosing open and close
Executions are sorted first by event time. When every execution in a bucket has a source sequence, that sequence breaks timestamp ties. Otherwise Omens uses a deterministic ordering key and marks ordering as uncertain.
Deterministic ordering makes recomputation repeatable. It does not establish the true exchange order of same-timestamp observations when the source lacks enough information. Open and close inherit that limitation; high and low do not depend on the order.
Calculate prices, share volume and traded value
Open and close come from the first and last ordered execution. High and low are the maximum and minimum execution prices. Omens sums shares and execution notionals using checked decimal arithmetic. In the observed-trade sample, notional is price multiplied by shares.
Volume measures outcome shares. Quote volume measures value in the source collateral unit, and VWAP divides that value by shares. Neither volume nor trade count measures unique traders or net buying. The sample labels its trade count as observed_executions.
volume = Σ shares; quote_volume = Σ(price × shares); VWAP = quote_volume / volume
Recompute corrections and aggregate larger intervals
A late execution or correction can change a previously displayed candle. Candle revisions let a newer computation replace an older one; the frozen sample keeps its original captured values so the example remains reproducible.
For larger intervals, Omens takes the first available one-second open and last available close, finds the highest high and lowest low, and sums volume, quote volume and trade counts. Recompute VWAP from summed value and shares rather than averaging individual candle VWAPs.
Aggregation cannot repair missing executions. Zero-volume display candles carry a previous price for visual continuity and must not be interpreted as new trades.
Reproduce the checks on a real BTC five-minute round
The linked CSV contains 366 non-synthetic candles for market 5060604 during September 29, 2026, 01:55–02:00 UTC. Prices are on the 0–1 scale. Keep market_id and outcome with bucket_ms when grouping rows.
Download the CSV, metadata and Python loader into one folder. Run python3 load-ohlcv.py to verify the SHA-256 checksum, row count, UTC bucket alignment and price bounds, then summarize each outcome. The script uses Python’s standard library and Decimal, so no package installation is required.
The metadata reports unverified source completeness and three unrepaired Down-token connection gaps. The script prints those limitations alongside the totals. Passing file checks establishes consistency with the frozen export, not completeness against the exchange.
Why a price-history series cannot supply execution volume
Polymarket’s price-history API documents observations with a timestamp and price. Those two fields alone do not supply individual execution sizes. Consequently, resampling that series cannot reconstruct execution-based share volume or every traded high and low.
Execution OHLCV adds those measurements from the trades that were recorded. Coverage still matters: a valid observed candle can omit activity during an outage. Compare sources, timestamp bases and collection windows before treating two datasets as equivalent.
Explore Polymarket 1s OHLCV data
Keep exploring
Polymarket 1s OHLCV Data & Candlestick Charts
Explore Polymarket 1-second OHLCV candles built from recorded trades. View live candlestick charts, trading volume, historical markets, and download a real 1s OHLCV sample.
Follow trading activity. Understand what it measures.
Read Polymarket recorded trading volume, execution counts and event concentration. Learn how OHLCV share volume differs from traded value and liquidity.
Read the statistics behind each round.
Understand prediction market volume, VWAP and historical settlement frequencies. Learn how sample selection and missing data affect round statistics.