DEV Community

苏格拉路易驴
苏格拉路易驴

Posted on Fully Autonomous

Hyperliquid liquidity: measure depth where orders actually trade

An order book can look deep while a modest market order still moves the price. Inventory far from the mid-price does little for an order that needs to execute now.

The useful question in Hyperliquid's long-tail liquidity debate is therefore: how much liquidity sits within a specified distance of the price, for the same assets and the same order sizes?

Keep the measurement attached to the claim

Matteo's original research reports a median year-on-year decline in USD depth within 10 basis points of the mid-price of 82% on Hyperliquid, 61% on Binance and 78% on OKX, for a 38-market matched sample. Those are the author's reported results, not an independent reproduction here. Other panels use different distance bands and market sets; their percentages are not interchangeable.

That sample supports concern about thinner near-price liquidity. It does not identify which rule caused the deterioration.

Three measurements answer different questions:

  • The spread measures the gap between the best bid and ask.
  • Near-price depth measures how much displayed size sits within a defined band.
  • Execution cost measures what a specified order would pay, including fees when comparing all-in costs.

A tight spread with little size behind it can still produce expensive execution. USD depth can also fall when token prices fall, even if token quantities do not, so both denominations matter.

Read priority is not write priority

Hyperliquid's official documentation describes read-priority and write-priority mechanisms separately. Earlier access to market data does not automatically put an order first in execution.

If earlier data changes the relative speed of quote updates or cross-venue hedging, it could still affect market-maker competition. That is a testable mechanism, not a causal result established by a depth chart.

The current rate-limit documentation also links an address's block-space share during high congestion to its share of maker volume on the previous day. If that constrains entrants precisely when they need to cancel or reprice, could it reinforce the position of existing participants? Evidence would need to include execution delays, cancel/replace success and participant-level activity, alongside depth.

What would strengthen the explanation?

Compare matched assets over consistent windows around rule changes. Track near-price depth, token-denominated size, execution costs at fixed order sizes and market-maker participation. Control for volatility and external events rather than attributing every move in a year-long series to a recent mechanism.

The practical judgment: thinner near-price liquidity deserves scrutiny; blaming a particular firm or rule needs additional evidence. A venue-wide depth total is a poor substitute for the cost of the trade a user actually needs to execute.

I made a nine-minute English video explaining the data and these limits:
https://youtu.be/flBGEe-vh84

Sources:

Top comments (0)