Two Aite Group industry analysts document how U.S. market structure went electronic between 1997 and 2010 and how high frequency shops came to dominate it. Read it for the mechanics of colocation, sponsored access, and order types, not for a strategy, because the book never hands you one.
The one idea
This is a consultant's field report on an industry, not a trading manual. Zubulake and Lee, both analysts at the research firm Aite Group, trace how U.S. equity trading moved from floor-based specialists to electronic communication networks (ECNs) after the Order Handling Rules of 1997, through Regulation NMS in 2005, to a market where high frequency firms account for roughly 60 percent of average daily equity volume by the book's telling. The lens throughout is market structure and infrastructure: who the players are, what the venues charge, and how the plumbing works.
The strategy content that exists is a taxonomy, not a playbook. Chapter 5 names market making, statistical arbitrage, correlation trading, basis trading, momentum, and news-driven trading as the main HFT strategy families, gives each a paragraph, and moves on. There is no entry rule, no exit rule, no risk sizing, and no backtest anywhere in the book. What the authors do deliver in detail is the cost and latency of the infrastructure underneath those strategies: colocation running $1,500 to over $10,000 a month per rack, sponsored access latency broken down to the microsecond, and the maker-taker rebate structure that shaped how liquidity gets priced.
The book's real value is as a snapshot of a specific and consequential moment: the year after the May 6, 2010 Flash Crash, when regulators, politicians, and the public were still deciding what to make of automated trading. Read that way, as a primary source on how the industry explained itself to itself in 2011, it earns its place. Read as a way to learn to trade, it has nothing to offer.
What it actually teaches
The book has no sequential trading method. What it does offer are four distinct components of how HFT actually operates, each documented with real figures.
READ THE BOOK, NOT JUST THE TOP PRICE
Chapter 5 opens with the mechanics every HFT strategy sits on top of: the top of book (current best bid and offer) and the depth of book (size available below that). Their worked example: a market at 10.50 bid / 10.51 offer with 1,000 shares on the bid and only 10 on the offer suggests the next trade is more likely at 10.51, but the authors immediately note this can be a false signal if a prior trade already cleared the real size at that price. This is the only piece of the book that resembles a tradable observation, and it is not developed further.
THE STRATEGY FAMILIES, NAMED BUT NOT SPECIFIED
Market making (buy the bid, sell the offer, adjust constantly for supply and demand), statistical arbitrage (trading the gap between an index and its underlying names as one leg moves before the others), correlation trading (one contract leads, another follows, such as a Canadian index reacting to a U.S. index move), basis trading (the spread between a commodity's spot price and its futures price), momentum trading (a chartist's breakout style adapted to machines, with no fundamental basis), and news-event trading (algorithms scanning releases for keywords that might move price). Each gets a paragraph. None gets a rule.
THE EXECUTION LAYER: ALGORITHMS AND ORDER TYPES
The book is precise here because this part is documentation, not strategy. VWAP slices an order to match the volume-weighted average price over a window; TWAP targets the time-weighted average from order submission to completion; a pegged order tracks the national best offer or bid; arrival price targets the bid/ask midpoint at submission, adjusted for a risk-aversion setting and a target percent of average daily volume. Table 5.1 lists the order types beneath those algorithms: IOC (immediate or cancel), FOK (fill or kill), iceberg (hidden size that refills as it fills), trailing stop, MIT (market if touched), MOO, and MOC.
THE INFRASTRUCTURE THAT MAKES ANY OF IT POSSIBLE
Chapter 3 puts real numbers on the plumbing. Colocation runs $1,500 to over $10,000 a month per rack depending on service level, before power costs. Sponsored access latency breaks down to roughly 450 microseconds for network connectivity, 85 for the exchange gateway, and 125 for pre-trade risk checks; total latency ranges from 4 to 8 milliseconds for standard direct market access down to 250 to 350 microseconds for unfiltered (naked) sponsored access. The book estimates sponsored access accounted for about 50 percent of average daily U.S. equities volume by 2009, split roughly 12 percent filtered and 38 percent unfiltered.
What it looks like Monday morning
There is no trade to place differently after this book. Its use is as a reference for the vocabulary and cost structure of the market a retail or prop trader is placing orders into: what a maker-taker rebate is doing to the quote you see, why an iceberg order can make displayed size misleading, and why a market order in a thin book is a different risk than a limit order chasing a fast-moving price.
The more useful habit it teaches is skepticism about round numbers. Nearly every statistic in the book, HFT's share of volume, colocation costs, FX adoption rates, is either an Aite Group estimate or a forward projection labeled e2010 through e2015 in the book's own charts. A reader should treat every figure here as directional, from a single research shop, at a single point in time, not as an audited fact to build a rule on.
| Order type | What it does |
|---|---|
| IOC | Immediate or cancel: fills what it can, cancels the rest |
| FOK | Fill or kill: must fill completely or the whole order cancels |
| Iceberg | Hidden size order: adds more size as it fills, same amount or randomized |
| Trailing Stop | Stop price adjusts as the market moves |
| MIT | Market if touched: converts to a market order once a price point is hit |
| MOO | Market on open: executes at market price when the market opens |
| MOC | Market on close: executes at market price within one minute of the close |
This table, not the strategy taxonomy, is the most concretely useful page in the book.
The goal should be to be the smartest as opposed to the fastest.Paul Zubulake and Sang Lee, The High Frequency Game Changer
Where it fails
- The catalog title and the book's own title page do not match. The PDF under review opens with the title page "The High Frequency Game Changer: How Automated Trading Strategies Have Revolutionized the Markets" by Paul Zubulake and Sang Lee (Aite Group, Wiley, 2011). It is not the widely referenced academic text sometimes called "The High Frequency Trading Book." The file this page was built from is what a reader following the buy link needs to be able to identify, and right now the shelf entry name does not match what is inside the file.
- There is no method to test, backtest, or fail. The brief for this site asks what a reader does differently after reading. There is no entry rule, exit rule, position size, or risk parameter anywhere in the 165 pages. The closest thing to a strategy is the six-item taxonomy in chapter 5 (market making, stat arb, correlation, basis, momentum, news-driven), and each entry is one paragraph of description with no numbers a trader could act on. A reader hoping to walk away with something to code or paper-trade gets nothing.
- Every statistic is a single research shop's estimate, and many are explicit projections. Figures like 1.2, 6.1, 6.3, and 6.11 are labeled with years such as e2010 through e2015, meaning the authors' own firm, Aite Group, is projecting forward from data collected around 2009. The HFT share-of-volume figures, colocation pricing, and FX adoption percentages all come from the same single source with no independent verification cited. None of these numbers should be treated as historical fact; they are one consultancy's estimate from 2010.
- The regulatory and infrastructure content is now a historical snapshot, not current information. Chapter 8's account of Dodd-Frank is written in the future tense, describing rules "expected" to take effect, because the book went to press before most implementing regulations were finalized. Reg SCI (2014), the 2015 SEC market structure reforms, the shift to zero-commission retail trading, and the payment-for-order-flow scrutiny of the 2020s are all outside the book's window. The maker-taker economics, colocation pricing, and sponsored access latency figures in chapter 3 describe a 2009-2010 cost structure that has since compressed significantly as colocation and low-latency connectivity commoditized.
- The Flash Crash account was written before the full official explanation existed. The book describes May 6, 2010 and blames it broadly on "fundamental decisions by individual money managers," without naming the trigger the SEC and CFTC joint report identified later that year: a single large sell algorithm executing a $4.1 billion E-mini S&P order without regard to price or time. Because the authors wrote before that report's full findings were digested, the chapter reads as a defense of the automated trading community's role rather than an account of what specifically happened.
- It is written for someone starting a proprietary trading firm, not someone trading their own account. The closing section of chapter 1, "Building a High Frequency Trading Team," assumes the reader is choosing a chief technology officer, negotiating broker and exchange relationships, and picking an asset class to build a firm around. Nothing in the book addresses position sizing, risk management, or strategy selection at the scale of an individual trader or small account. The audience is implicitly an institutional builder or an industry watcher, not a retail reader looking for an edge.
Who it is for
Buy it if
You want a plain-English primer on how U.S. and global market structure works, what colocation and sponsored access actually cost, and how the HFT industry explained itself to regulators and the public right after the 2010 Flash Crash. It is a fast, well-organized read for that narrow purpose.
Skip it if
You are looking for a trading strategy, a backtest, or anything you could code and run. Skip it too if you need current numbers; the statistics and regulatory picture are frozen at 2009-2010 and the book projects forward from there rather than reporting what actually happened after publication.
As an Amazon Associate I earn from qualifying purchases. It costs you nothing extra, and it does not change what gets recommended or what the verdict says.