Skip to main content

​
Reading the order book

The order book is execution context, not certainty.

​
V2: on-chain BST depth

In QuantDesk V2, displayed depth comes from crankless BST slabs — bid and ask red-black trees stored inside each MarketV2 account. Matches happen atomically in place_order_v2 without waiting for an off-chain keeper. Integrators can fetch aggregated depth via:
See V2 API endpoints for response fields and cache behavior (redis vs database fallback).

​
V2 crankless BST matching

V1 relied on hybrid models where off-chain keepers could delay or batch matching. V2 moves matching on-chain into a bounded Binary Search Tree (BST) slab per market side (bids and asks).

​
V1 vs V2 execution

​
How a trade flows

Key properties:
  • No crank for core fills — takers match against resting liquidity in the same transaction they submit.
  • Bounded slabs — each side caps at 256 nodes so compute stays within Solana CU limits.
  • Seat gate — makers claim a MarketSeat PDA before posting limit orders (spam protection).
For the integrator tutorial, see Building on QuantDesk. For portfolio state layout, see Unified portfolio architecture.

​
What to watch

  • Spread: tighter is usually easier execution
  • Near depth: more depth can reduce slippage
  • Imbalance: useful context, not a standalone signal

​
Practical habits

  1. Check spread before urgent orders
  2. Compare size against near depth
  3. Prefer staged entries in thin liquidity
  4. Re-evaluate after partial fills