The House Edge

Appearance

Chapter 1 — Review Notes (Phase 3 self-check) ​

Status: REVIEW NOTES (Phase 3 — written by agent, accompanying DRAFT.md) Date: 2026-05-13 Working chapter: 01 — What Is a Trade, On-Chain Word count: ≈3,380 words including footnotes and the two inline diagrams after the voice-cleanup pass — deliberately under the Book Bible's 4,000–6,000 target, per Q7 of Phase 1.

Revision note, 2026-05-25: DRAFT.md received a prose cleanup pass to reduce agentic scaffolding, cut abstract phrasing, and make the chapter sound more like a clear human narrator while preserving the existing structure, examples, numbers, and footnotes.

The eight required questions ​

  1. Could a smart business reader with zero crypto background follow this chapter on first read?Yes — this is the chapter most explicitly designed for that reader. Every term (intent, settlement, finality, gas, transaction, block, wallet, mempool) is defined inline before it carries load. The Coinbase parallel runs through every section. The technical detail is bounded: the chapter explains what each phase does and what it costs, not how the consensus mechanism works under the hood. If the reader is going to find a hard moment, it is probably the Finality section's L2 fraud-proof-window paragraph — it is the densest part of the chapter and unavoidably so.

  2. Is every actor named, and is it clear how each one makes money?Yes. Alice (the trader — paying fees), Coinbase (charging the taker fee; revenue is itemised), Jupiter (routing layer, monetises slippage protection — cameo with forward to Ch 2 and Ch 6), the slot leader / validator (cameo with forward to Ch 5), the prop-AMM operators (one-sentence cameo with forward to Ch 2). No "Meet the actor" sidebar in this chapter — that was an explicit SPEC decision, since Ch 1 is setup-only.

  3. Is there a worked example with specific dollar amounts threaded through the chapter?Yes. Alice's $10,000 USDC→SOL swap, run twice — once on Coinbase Advanced Trade, once on Solana via Jupiter. Threaded through six explicit beats (Setup plant, Worked Example introduction with Mermaid sequence, Intent on both sides, Settlement on both sides with the cost-breakdown table, Finality on both sides, Verdict). The total bill comes out at ~$25 on Coinbase vs ~$50 on Solana — same order of magnitude, different line items, which is the chapter's structural point.

  4. Does the chapter end with a clear "who wins, who loses" verdict?Yes, intentionally light-touch. The SPEC was explicit that this chapter is structural rather than adversarial, and the verdict reads as "traders who match the venue to the job win; traders who route to the wrong venue pay more than they needed to; this is not bad, just market structure." The closing line ("Chapter 1's verdict is restraint: do not draw conclusions until the actors are on the board") earns its position because the next several chapters are about the actors and the privileges.

  5. Are all numbers sourced in footnotes?Yes. 17 footnotes. Every dollar figure, percentage, time-to-finality figure, gas price, and named incident is cited with a URL and access date. The Boulton/Shohfi/Walz DERA Working Paper (2025) is the primary source for the PFOF aside — added late after Nick shared the PDF directly. Alpenglow is cited via the CoinDesk piece dated today (2026-05-13).

  6. Does the chain comparison box exist and contain real differences (not "Solana is faster")?Yes. Three paragraphs, each making structurally different points: Solana (slot timing + Alpenglow shipping; gas costs; failed-tx framing with the 92%-bot caveat; no global mempool); Hyperliquid (200ms median end-to-end; no public mempool; specific fee tiers; perpetuals-first); Ethereum and L2s (12.8min L1 finality; L1 vs L2 cost split; the 7-day fraud-proof window for optimistic rollup canonical settlement). No throughput-comparison sentences.

  7. Did I avoid every banned move from the Book Bible?Yes, with one flagged near-miss.

    • No hype words. No doom words. No tribal chain endorsements. No comparing crypto to "Web 1.0." No "It's important to note that…"
    • Near-miss: The PFOF aside cites concentration numbers ("Citadel Securities and Virtu Financial alone account for 60–70% of equity wholesale flow") which could read as setting up a future adversarial argument. I think it's load-bearing — the chapter explicitly forward-links to Chapter 7 on exclusive order flow, and the SEC DERA paper's 45×-equity figure is one of the chapter's most striking numbers — but it does push the structural-not-adversarial framing.

  8. Would the Goldman MD finish this chapter without checking her phone?Probably. I read it aloud once. The T+1 cold open is the right hook for this reader (Goldman, NYSE, DTCC are all familiar). The Alice / Coinbase / Solana parallel keeps the chapter concrete. The PFOF aside is the chapter's most novel piece for a TradFi-fluent reader who already knew T+1 and gas — a useful surprise. The Finality section is the chapter's most technical stretch but the dollar table at the end of Settlement breaks the prose load. The 3,800-word total is short enough that she can finish it in a sitting.

What changed between phases — and what's load-bearing ​

Claims dropped from RESEARCH.md ​

  • The Helius "Coinbase doesn't use PFOF" inferential note (RESEARCH.md original claim 13). The chapter still makes the claim, but with a more careful framing — Coinbase Advanced Trade's spot retail flow doesn't use PFOF, but the chapter is precise about what that means and what it doesn't mean. The DERA paper is the primary citation for the PFOF mechanism description; Coinbase's own help pages are the citation for its specific revenue model.
  • Most of the academic-failure-analysis paper detail (RESEARCH.md N14). Used one line of it — the 92%-bot framing — and cited the paper. The chapter doesn't need the full breakdown of failed-tx categorisation.

New claims added in the draft (and where they came from) ​

  • The full PFOF aside in §5 Settlement (~150 words). Triggered by Nick sharing the SEC DERA paper mid-draft. The 45×-equity, 4.5×-options, ~$4.8M/day-impact figures all come from the abstract and Section 5 of the paper. The wholesaler-concentration data (three firms = 70–82% of equity PFOF; Citadel and Virtu = 60–70% of equity wholesale flow) comes from the paper's §2.2 citing Hu and Murphy (2024) and Bryzgalova et al. (2023).
  • The chapter's structural claim that "finality is a curve, not a number" in §5 Finality. Synthesis, supported by the cited finality numbers across chains, but not directly stated in any single source.
  • The two-venue dollar-cost table at the bottom of Settlement. Constructed from the cited fees + estimated pool slippage. The Coinbase 25-bps figure is approximate (depends on volume tier); the Solana $50-slippage figure is illustrative for a clean major-pair routed swap. Both are flagged in footnotes 4 and 8.

Things I'm uncertain about ​

  1. The Alpenglow framing. The chapter says Alpenglow is "shipping" — live on testnet today (2026-05-13), mainnet expected late Q3 / early Q4 2026. If Alpenglow slips, the chapter still reads correctly because the 12.8s production number is the anchor. If Alpenglow ships on schedule, the chapter still reads correctly because the forward sentence is conditional. Higher-risk scenario: Alpenglow ships faster than planned (mainnet in Q2 2026), in which case the chapter's "shipping later this year" line dates faster than ideal. Manageable.
  2. The PFOF aside's location and weight. I put it inside §5 Settlement as a styled ::: info aside. It could equally have been a longer footnote (cleaner reading flow) or a separate sub-subsection (more weight). The styled aside is the middle path. If the reader's pace breaks at the aside, it should drop to a footnote.
  3. Hyperliquid spread/lead-lag still not cited in this chapter. The chapter cites Hyperliquid's fee numbers (from primary docs) but not the disclosed spreads (Ch 2 cites those, with the operator-disclosed hedge). The decision was deliberate: Ch 1 is setup-only and doesn't need the spread detail. Flagged for cross-chapter consistency only.
  4. The Coinbase taker fee number (25 bps). Coinbase's actual fee schedule depends on the trader's volume tier; 25 bps is approximately the base-tier taker rate for a moderate-volume retail customer. The chapter says "approximately" and footnotes the variability. If you want a tighter number, I could cite Coinbase Pro's current spot taker (typically 4–10 bps for higher tiers) but the higher-tier rate isn't what Alice would actually pay.
  5. The Verdict's lightness. The "is this bad? no" answer is correct for the chapter but feels less satisfying than Ch 4's clinical verdict or Ch 2's "the role has been redefined under the LP's feet" finish. Setup chapters earn lighter verdicts. Flagging because if you want a stronger close, the simplest fix is to tee up Chapter 3 more sharply in the closing.

Places where the prose got technical and might lose the reader ​

  • §5 Finality, the L2 fraud-proof window paragraph. Four new concepts in three sentences (sequencer soft confirmation, fraud-proof window, canonical settlement, native vs liquidity bridges). I kept it as a paragraph rather than breaking it out because the structural point — finality is a curve — earns the density. If a reader stops here, the fix is to drop the bridges sentence and let the canonical-vs-soft framing carry it alone.
  • The PFOF aside. Cites Citadel, Virtu, Hu and Murphy, Bryzgalova et al. — three named institutions and two academic sources in a single block. The aside is fenced in a styled callout so the reader can skip it if they want, but if they read it, the density is real.

Files written/modified in Phase 3 (this chapter) ​

  • book/chapters/01_trade/DRAFT.md — new, ≈3,800 words (17 footnotes; inline Mermaid sequence diagram + inline markdown cost-breakdown table)
  • book/glossary/GLOSSARY.md — appended 8 entries (Block, Finality, Gas, Intent, Settlement, Trade, Transaction, Wallet); now 38 total entries
  • book/OUTLINE.md — Chapter 1 entry updated with subtitle and final section headings

Chapter 2 back-edits made during Phase 3 ​

Two corrections to book/chapters/02_liquidity/DRAFT.md and to Ch 2's RESEARCH/OUTLINE planning docs, surfaced by Ch 1's Phase 1 research:

  • Hyperliquid finality: "~70-millisecond intervals" → "~200ms median end-to-end latency (~900ms p99)" — corrects against the canonical Hyperliquid Documentation.
  • Hyperliquid fee: "3.2 bps taker" (in Ch 2 planning docs) → "4.5 bps base / 2.4 bps top-tier on perpetuals" — corrects against the current Hyperliquid Fees page.

Both back-edits are documented in Ch 2 REVIEW_NOTES.md rev 2.

Phase 3 is complete. The chapter is now in Nick's review queue.