F-0006 · first recorded 2026-08-20

Binance order-book behaviour under stress transfers to Hyperliquid

ASSUMED

Observationnone — nothing has been measured on Hyperliquid
Samplenone
Methodn/a
What is not establisheduntested and load-bearing; hl2 began recording Hyperliquid l2Book at nSigFigs=3 on 2026-08-20 to settle it
Who else publishes thisn/a
Evidenceresearch/ASSUMPTION-binance-physics-may-not-transfer.md

The most dangerous category in the registry. The cascade model calibrates depth evaporation on three years of Binance data and applies it to Hyperliquid, which has HLP backstopping liquidations, a different maker population, an order of magnitude less open interest, and block-paced matching.

If Hyperliquid’s book holds up better, our cascade depths are too pessimistic. If worse, too optimistic — the dangerous direction for anyone acting on a forecast.

No Hyperliquid cascade figure may be published without this caveat on its face until F-0006 moves to MEASURED or REFUTED.

← Every claim and its status · This finding as JSON →