The House Edge

Appearance

Chapter 2 Spec — Where Liquidity Lives ​

Status: STRAWMAN (drafted by agent, 2026-05-13 — Nick edits before Phase 1) Production order: Second chapter to draft, after Chapter 4

⚠️ This SPEC is a strawman. Nick wrote the Ch 4 SPEC; this one was drafted by the agent off OUTLINE.md and the Ch 4 template, then handed back for editing. The fields most worth Nick's attention are Worked example candidates (the Ch 4 SPEC included Nick's preference; this one flags the recommended option but leaves the call open), Tone notes specific to this chapter, and Open questions for the agent's Phase 1 research (these become the SPEC's research targets).

Learning objective ​

By the end of this chapter, the reader can:

  1. Distinguish the AMM and CLOB architectures and articulate which problems each one solves and which it doesn't.
  2. Name the economic actors who provide liquidity in each architecture — passive LPs in AMMs, resting market makers in CLOBs — and describe each one's revenue source and risk profile.
  3. Explain why aggregators (Jupiter, 1inch, CoW) and RFQ have become the dominant retail layer on top of either architecture.
  4. Identify the chain-specific reasons Solana converged on AMMs plus aggregators, Hyperliquid built a native CLOB, and Ethereum supports both.

Why this chapter ​

Chapter 2 establishes the foundational vocabulary the rest of the book needs.

Chapter 4 (the voice-calibration chapter, drafted first) already mentioned LPs as silent counterparties twice — in §5a (CEX–DEX arbitrage as a tax on the on-chain pool) and §7 (passive LPs as one of the named losers). Those references work in Ch 4 because the reader needs just enough to follow the argument. Chapter 2 makes the LP role explicit and develops the MM-vs-LP distinction that every later actor chapter leans on.

This chapter is not adversarial in the way Chapter 4 was. Its job is to build the board. The adversarial thesis returns in Chapter 3 (mempool dynamics) and is in full force by Chapters 5–7.

Key questions answered ​

  • What is liquidity, in plain terms, and why doesn't a market "just work" without it?
  • What is an AMM, and how does it differ economically from a CLOB?
  • What's the difference between being a passive LP and being a market maker, and why does it matter to a business reader?
  • Why have aggregators become the dominant retail interface on Solana and Ethereum, and what value does an aggregator capture for itself versus pass through to the trader?
  • Why did Hyperliquid build a native on-chain CLOB when every prior on-chain CLOB attempt has failed?

Characters introduced ​

  • The Market Maker (full sidebar — first appearance in the book)
  • The Passive LP (the silent counterparty Chapter 4 referenced; explicit treatment here, no full sidebar but a half-page "How an LP earns and loses" treatment)
  • The Aggregator (Jupiter as the named exemplar; cameo in Ch 4 as a routing layer, full treatment here)
  • The Solver (cameo only — full treatment in Chapter 7 when exclusive flow is the topic)

Worked example candidates ​

The agent should consider all three and pick one to anchor:

  1. A market maker on a Solana on-chain CLOB posts a two-sided quote for SOL/USDC and gets adversely selected when a CEX prints first. The chapter walks through the MM's P&L on a single tick: the spread they collected, the position they're now stuck with, the hedge they need to put on. Pros: vivid, lets the chapter develop the adverse-selection mechanic concretely, sets up the "informed flow vs. uninformed flow" thread that runs through later chapters. Cons: forces a Solana frame early and may be too inside-baseball for the reader's first encounter with market making.

  2. A passive LP deposits $50,000 into a Uniswap V3 ETH/USDC pool and tracks P&L over a month — fees earned, impermanent loss, the cost of staying. Pros: relatable, dollar-anchored in the Book Bible's preferred style, lets the chapter develop the LP role in detail and plant the "passive LP as silent counterparty" thread for Chapters 4 (already drafted) and 11. Cons: less adversarial than the rest of the book; needs to be paired with the explicit observation that someone is on the other side of every LP fill.

  3. A trader on Hyperliquid lifts an offer on the BTC perpetual; the depth that filled them was provided by HLP plus three named market makers; the spread that closed was 8 bps. Pros: showcases CLOB mechanics, concrete, lets HLP appear early (it's load-bearing for Chapter 9). Cons: Hyperliquid has less broad reader familiarity at this stage of the book; the chapter would need to define perpetuals briefly even though they don't recur much in the book.

Agent's recommendation (Nick edits): Option 2 (the LP P&L), with Option 1 as a 1–2 paragraph cameo in the MM section and Option 3 as a paragraph cameo in the chain-comparison box. Reasons: Option 2 is the most accessible for a business reader on the chapter's first material encounter with liquidity, it carries the Bible's "show the dollar" mandate clean through, and it gives the chapter its strongest forward link — once the reader has watched a $50K LP position lose $X to impermanent loss and earn $Y in fees, the "passive LP as chronic loser" thread for Chapter 11 is already half-built.

Glossary terms this chapter introduces ​

The chapter will introduce a lot of terms — most belong in cameo form (one sentence inline plus glossary entry); only a handful warrant full treatment in the prose.

Full prose treatment (definitions earn their length in the chapter):

  • Liquidity
  • AMM (automated market maker)
  • CLOB (central limit order book)
  • Liquidity provider (LP) — distinguishing passive LP and active LP
  • Market maker (MM) — distinguishing on-chain MM and off-chain-MM-quoting-on-chain
  • Spread
  • Slippage
  • Impermanent loss / divergence loss
  • Aggregator
  • RFQ (request for quote)

Cameo only (one inline sentence, full glossary entry):

  • Concentrated liquidity (Uniswap V3-style)
  • Constant-product curve (xy = k)
  • Resting order, marketable order
  • Taker / Maker
  • HLP (Hyperliquid LP vault) — cameo here, full treatment in Chapter 9
  • Solver — cameo here, full treatment in Chapter 7

Diagrams needed ​

2–4 diagrams, per the Bible's per-chapter budget.

  1. The AMM curve. A constant-product bonding curve showing the price-impact-vs-trade-size relationship. Annotated: where the curve sits, what a swap does to it, what an LP deposit does. The reader should be able to point at the curve and explain why a $10K swap on a $1M pool moves the price more than a $10K swap on a $100M pool.

  2. The CLOB order book. A snapshot of bid/ask depth — say five levels each side, with named market makers attributed to each level. The reader should see the spread, see the depth, see who's quoting where.

  3. Aggregator routing. A Sankey-style diagram showing Alice's $10,000 swap from the prologue (recall: the same Alice from Chapter 4) being routed across three pools to minimise slippage — for instance, 60% through one Raydium pool, 25% through Orca, 15% through a Phoenix CLOB if relevant. Each leg annotated with the price it took.

  4. (Optional) MM vs LP P&L profile. A two-column table comparing the income sources, risk drivers, and time horizons of an MM versus a passive LP. Useful if the prose runs short; potentially redundant if the worked example does the job.

Backward:

  • Chapter 1 (What Is a Trade, On-Chain). Not yet drafted. Chapter 1 will define the three phases (intent, settlement, finality) the reader needs before Chapter 2's mechanics land. If Chapter 1 is not yet drafted when Chapter 2 enters Phase 3, this chapter must briefly establish "what a trade is on-chain" inline (one or two sentences) before §3 begins.
  • Prologue (Alice's $10K swap). Not yet drafted. The aggregator routing diagram should plausibly continue the prologue's setup.

Forward:

  • Chapter 3 (The Mempool and What Replaces It). The pre-trade visibility surface that creates the sandwich opportunity Chapter 4 already developed.
  • Chapter 4 (The Searcher). Already drafted. The passive-LP thread planted in Ch 2 is referenced in Ch 4 §5a and §7 — Ch 2 should make explicit that the LP role is the one Chapter 4 referred to without fully introducing.
  • Chapter 6 (Infrastructure Layer). Aggregators get richer treatment as routing infrastructure once the actor structure is in place.
  • Chapters 8, 9, 10 (Solana, Hyperliquid, Ethereum). The chain-specific convergence story this chapter argues for is the foundation those chapters develop in depth.
  • Chapter 11 (Who's at a Disadvantage). The "passive LP as chronic loser" thread Ch 2 sets up is one of the four loser archetypes Chapter 11 revisits.

Tone notes specific to this chapter ​

  • The chapter is not adversarial. It is structural. Resist the urge to score points against any particular architecture; the chapter's job is to describe the board. The verdict on each architecture lives in Chapter 11.
  • Diagrams replace prose, they don't supplement it. This chapter introduces three or four mechanics that lend themselves to visual explanation — the AMM curve, the order book, the routing graph. The text should say what the diagram means, not narrate what the diagram shows.
  • "Meet the Market Maker" is the book's first actor sidebar — set the format. A half-page side-bar in a different visual style; the format established here is what subsequent chapters will mirror (Meet the Searcher already exists in Ch 4 and uses VitePress's ::: info container — the format is portable to print via a pandoc filter at MANUSCRIPT.md assembly time).
  • Do not preview Chapter 4. A reader who has not yet read Chapter 4 should be able to finish Chapter 2 without needing it. A reader who has read Chapter 4 should find one or two satisfying back-references (the LP role; the aggregator).
  • The chain-specific convergence story is the chapter's argument. Solana → AMMs because high throughput, low fees, and composability favour pool-based liquidity. Hyperliquid → CLOB because the design target was professional traders willing to pay for tight spreads and the custom-consensus stack made on-chain matching feasible. Ethereum → both because the platform is permissionless and the gas-cost regime favours different architectures at different size bands. This is not a value judgment; it is a structural observation.

Open questions for the agent's Phase 1 research ​

Number these in the research note. Treat them as research targets, not as exhaustive.

  • LP P&L on Uniswap V3 in 2026. The Topaze Labs / Bancor-style "passive LPs lose to informed flow" measurement series — what's the most recent published instalment? Has the 2025 IRA Sigmadex / Crocswap work been updated for 2026 conditions?
  • HLP returns, Q1/Q2 2026. Hyperliquid publishes the HLP vault's return profile. What does it look like in 2026? Specifically: is HLP earning enough to compensate for the adverse selection it absorbs?
  • On-chain CLOB activity on Solana. Phoenix, Open Book, dexalot — what's the actual volume picture in 2026? Are any of these above noise, or has the AMM-plus-aggregator stack won definitively?
  • Aggregator market share by chain in 2026. Jupiter dominates Solana; on Ethereum the 1inch / CoW / 0x / Paraswap split is more even. Latest numbers.
  • RFQ retail flow share, Ethereum 2026. CoW Swap publishes a "share of retail flow filled by solvers vs. AMM pools" metric. What is it in 2026?
  • Concentrated liquidity utilisation. What fraction of Uniswap V3 TVL is in active (in-range) positions versus out-of-range? This number is the strongest single signal for "concentrated liquidity has not solved passive LP risk."
  • Adverse selection economics on Hyperliquid. Is there public data on how Hyperliquid MMs price the risk of CEX-first prints? HLP's existence is a partial answer; the active MMs' behaviour is the rest.
  • The 2025–2026 SVM order-book attempts. dFlow, Manifest, and similar — any of them gaining traction, or all stuck below the AMM-aggregator equilibrium?

Out of scope for this chapter ​

  • Specific MEV attacks against LPs and MMs (Chapter 4, already drafted)
  • Mempool dynamics and pre-trade visibility (Chapter 3)
  • Validator and builder economics (Chapter 5)
  • Cross-chain liquidity bridges and the L2-fragmentation thesis (Chapter 10)
  • Specific memecoin launchpads, token-launch mechanics, or Pump.fun-style primary issuance (footnote only if at all; not the chapter's subject)
  • The full mechanics of perpetuals or futures funding rates (Chapter 9 touches them in the Hyperliquid context; not here)

Keep this chapter focused on liquidity architectures and the actors who provide liquidity. Everything else is a cameo or a forward link.