Original HIP-4 research

HIP-4 market data, with the caveats left in.

A reproducible snapshot of Purrdict's independently indexed Hyperliquid outcome-market data. See cumulative notional, fills, parsed market classes, leading outcome tokens, and exactly where the dataset is incomplete.

Snapshot generated August 3, 2026 UTC · historical research data, not an execution quote

This build contains 1,818 HIP-4 outcome-token rows representing $403,183,505.4047749 in cumulative indexed notional and 5,723,936 fills. Those are archive totals from Purrdict's index, not claims about current liquidity or unique prediction-market questions.

01 / Snapshot

What the Purrdict index contains

Outcome-token rows1,818
Indexed notional$403,183,505.4047749
Indexed fills5,723,936
Parsed classes427 rows (23.5%)

The unit of analysis is an outcome token, identified by a HIP-4 coin such as #1731. It is intentionally not labelled a “market count.” A binary question generally has two outcome-token rows, while a multi-outcome question has one row for each possible outcome.

The figures refresh when the landing site builds. They are descriptive of what Purrdict has indexed at that moment; they do not estimate missing intervals or silently discard rows with incomplete metadata.

02 / Coverage

Market-class coverage is measurable—and incomplete

Purrdict parses human-readable HIP-4 descriptions into recurring price binaries, price buckets, named outcomes, and fallback rows. Historical descriptions are not always complete enough to classify. The archive keeps those rows and labels the gap rather than forcing them into the wrong category.

Parsed classRowsIndexed notionalFills
Unclassified1,391$216,664,945.743248522,162,357
Recurring price binary72$96,736,630.551509962,058,553
Named outcome140$76,938,965.931627541,237,442
Price bucket214$12,840,341.345073303265,532
Fallback1$2,621.8333155852
Known metadata gap

1,391 rows are currently unclassified. Their fills and volume remain in the archive totals, but Purrdict does not infer a market type without sufficient description data.

03 / Distribution

Largest indexed outcome rows by cumulative notional

This ranking is historical. A large cumulative row can already be settled, and average execution price is not the current midpoint. Use the live Purrdict app or Hyperliquid's own order book before placing an order.

OutcomeCoinClassIndexed notionalFillsRow tradersAvg. price
Argentina#1731Named outcome$10,374,317.4003877,4482830.7611868497572565
Argentina#1730Named outcome$8,816,103.6669556979,9601,0990.24169579302626312
Spain#2121Named outcome$8,482,727.37891133,7443220.7875820469703314
Spain#2120Named outcome$6,736,698.38941262143,9861,2600.23388513440383094
France#1891Named outcome$5,990,841.35061106,7623070.682929736797737
France#1890Named outcome$4,129,262.25390169107,1071,0560.3057663280932152
Portugal#2051Named outcome$3,680,740.9549815,2141590.8930680649401866
Recurring#1541Recurring price binary$3,495,337.6079869,5693000.6787103298883124
Recurring#200Recurring price binary$3,369,601.9448947,3436310.7370036592526878
Unnamed outcome#5100Unclassified$3,185,476.890113274,4523910.7947072027448338
Unnamed outcome#8131Unclassified$2,873,304.570719,6494150.5097525836874287
Recurring#601Recurring price binary$2,552,490.8891985,3343780.5570411555769096
04 / Methodology

How this snapshot is produced

  1. Index fills. Purrdict's data pipeline ingests HIP-4 outcome fills and preserves the protocol's outcome-token identifiers.
  2. Group by outcome token. The overview endpoint aggregates cumulative fill count, indexed notional, per-row trader count, and average execution price for each coin.
  3. Parse descriptions defensively. Known description shapes become named classes. Missing or ambiguous metadata remains unclassified.
  4. Generate a static research page. The landing build fetches the JSON distribution, calculates only transparent sums, prints the UTC build date, and exposes the source.

The page does not manufacture a protocol-wide unique-trader number by summing overlapping row counts. It also does not turn a token-row count into a unique-question count. Both would be visually impressive and analytically wrong.

05 / Limitations

What the data cannot tell you

  • Current executable liquidity. Cumulative notional does not reveal today's spread or depth.
  • Protocol-wide unique traders. Addresses overlap across outcome-token rows.
  • Unique questions from row count. One question can produce two or many tradeable rows.
  • Complete metadata coverage. Some historical rows lack a parseable public description.
  • Missing-data estimates. The totals include only successfully indexed records; no backfilled estimate is added for unavailable intervals.
  • Investment performance. Historical volume, fills, and average prices do not predict future returns.

For protocol identity and settlement mechanics, use the canonical HIP-4 guide. For current market prices, use Purrdict's live mainnet interface or query Hyperliquid directly.

06 / Access

Download, inspect, and cite the source

The complete overview distribution is public JSON at https://api.purrdict.xyz/api/markets/overview. The response exposes the same fields used on this page, so every displayed aggregate can be reproduced independently.

The public Purrdict HIP-4 market-data repository preserves dated snapshots, a machine-readable methodology, and a tested generator. Download the latest JSON snapshot or CSV snapshot under CC BY 4.0.

FormatsLive JSON API plus dated JSON and CSV snapshots
Live archivehip4.fun/markets
Production data APImute.sh

Suggested attribution: “Purrdict HIP-4 Indexed Market Overview, snapshot generated August 3, 2026 UTC.” Link to this landing page so readers get the methodology and limitations alongside the numbers.

07 / FAQ

HIP-4 market-data FAQ

How many HIP-4 markets does this dataset contain?

The current build contains 1,818 outcome-token rows. A row is not the same as a unique question: binary questions normally create two outcome tokens, and multi-outcome questions create more.

Is the volume figure live trading volume?

No. It is cumulative indexed notional across the rows returned by the Purrdict data API at build time. It is not open interest, TVL, current liquidity, or an execution quote.

Why are some rows unclassified?

The index preserves rows even when historical description metadata is missing or cannot be parsed into a known market class. Unclassified does not mean invalid; it means the public classification layer cannot make a reliable claim.

Can I sum the trader counts?

No. One address can trade several outcome tokens, so per-row trader counts overlap. The page deliberately does not present their sum as protocol-wide unique traders.

Where can I get the raw HIP-4 data?

Open the JSON distribution at https://api.purrdict.xyz/api/markets/overview. Builders that need streaming, webhooks, or normalized historical tables can use mute.sh.