Logo

Liquid Terminal

HIP-4 documentation

Testnet · Exploratory documentation — not official Hyperliquid documentation.

Research timeline

Structured recap of the multi-day sprint and the public X thread list.

For long-form prose, diagrams, and raw API notes, keep the markdown file open in parallel.

Sources

Three-day sprint

One block per calendar day — bullets are factual milestones, not a tutorial.

  • Day 1 — March 24
    • HyperCore / HyperEVM architecture: EVM inside Core, EvmRawTx in MessagePack
    • “2 worlds, 1 universe” — decoded V1 txs: deposit, createContest, finalizeContest
    • V2 found via heimdall-rs; 24/24 selectors; full V1 vs V2 comparison
    • HIP-4 as two layers: CLOB (HyperCore) + parimutuel (HyperEVM) — “two jars” theory
  • Day 2 — March 25
    • V1 + V2 research published on Liquid Terminal + X
    • Live testnet: YES on HYPE 15m market
    • VoteGlobalAction discovered — native L1 settlement (no Merkle claim)
    • registerTokensAndStandaloneOutcome — full L1 lifecycle mapped
    • Early settlement signers 0xe92d… / 0xc25c… — later refined to 8-wallet cluster (Day 4)
    • First winning prediction (BTC > 70,836); article draft + diagrams
  • Day 3 — March 26
    • V2 contests 865–867; EvmRawTx dual-identity (relayer vs EVM signer)
    • Deployer funding → Block Theory Cap theory; EVM likely not HL team
    • Multi-outcome = independent YES/NO pairs; pair minting confirmed live
    • EnigmaValidator joins; outcomeMeta, coin mapping, WebSocket research
  • Day 4 — March 27
    • Eight linked L1 addresses traced as one cluster (@quertyeth, @valtitudexyz)
    • Mapped: 0xc25c… HIP-4 operator (create / settle / questions / descriptions / named outcomes)
    • Mapped: 0xe92d… oracle (price feed); historical settlement tx proof; recent settle often 0xc25c…
    • Six cluster members unmapped (role TBD — hypotheses only)
    • No overlap with EVM parimutuel deployer 0xe21c…

X archive (10 posts)

Ordered roughly as published. If two entries share a URL, treat the later one as a reply in the same thread.

  1. Tweet 1 — HIP-4 uses HyperEVM (later corrected)

    First take: HIP-4 contest logic on EVM; trading on Core # markets. Later corrected — official flow is L1-native.

  2. Tweet 2 — V1 full research

    Reverse-engineered ABI, events, reverts, storage, lifecycle, bridge — links liquidterminal.xyz/hip4.

  3. Tweet 3 — V2 contract found

    Second genesis contract, OZ hardening vs V1; live Core markets still tied to V1 narrative at the time.

  4. Tweet 4 — First live test

    15m HYPE market: YES at 0.50 for 15 USDH — link to testnet trade UI.

  5. Tweet 5 — VoteGlobalAction discovery(thread · see t4)

    Settlement on L1 VoteGlobalAction; no Merkle claim; example settlement tx in thread.

  6. Tweet 6 — First winning prediction

    BTC above threshold YES won; instant payout; settler wallet rotation vs Polymarket/Kalshi.

  7. Tweet 7 — Creation tx found(thread · see t6)

    registerTokensAndStandaloneOutcome for recurring priceBinary BTC; full L1 create → trade → settle.

  8. Tweet 8 — Pair minting mechanics

    BUY YES mirrors NO; YES+NO=1.00; native Core matching vs Polymarket CTF.

  9. Tweet 9 — Minimum order (EnigmaValidator)

    size × min(markPx, 1−markPx) ≥ $10; API vs UI; Liquidiction — author profile (post URL may vary).

  10. Tweet 10 — Correction + new findings

    EVM contracts likely third-party; official HIP-4 = L1 only; multi-outcome + HIP-3 feed angle.