Chapter 4: The Speed Game

Every month, some product promises retail investors "HFT-powered AI trading" or "institutional-speed execution." This chapter exists to kill that idea cleanly, with specifics, so it never costs you money.

What high-frequency trading actually is

High-frequency trading is not "trading fast" in the colloquial sense. It is a physical infrastructure business. The core strategies:

Market making. Firms like Jane Street, Citadel Securities, and Virtu simultaneously post buy and sell quotes for thousands of securities, profiting from the tiny spread between the bid and the ask. They don't bet on where prices are going. They sit in the middle and collect fractions of pennies, millions of times a day. Profitability comes from systems, speed, and structure, from being first to update quotes when conditions change so you're never the one holding a stale price.

Latency arbitrage. The same stock trades on multiple exchanges at microscopically different prices for microscopically brief moments. If you can see all venues simultaneously and react faster than everyone else, you can buy on one and sell on another before the prices converge. The discrepancies last milliseconds. Capturing them requires colocation, your servers physically inside the exchange's data center, plus direct exchange data feeds and hardware optimized down to the microsecond, sometimes FPGA chips programmed for a single task.

The infrastructure list, to make it concrete: colocated servers, proprietary fiber or microwave links between data centers, direct market-data feeds (not the consolidated feed your broker shows you), and order types engineered for speed. Millions to build, millions a year to maintain.

Why "retail HFT" is a contradiction

A broker API, a cloud server, and a webhook cannot compete with a colocated server on a direct feed. This isn't a matter of degree, not "they're faster, but I'm pretty fast." It's different physics. By the time your cloud-hosted bot sees a price, reacts, and routes an order through your broker, the colocated firm's FPGA has seen it, traded it, and moved on. Thousands of times over.

Any product marketing "HFT for retail" is mislabeling what it sells. Usually it's one of three things: a regular algorithmic strategy (fine, but not HFT), a signal service (Chapter 5), or hope with a server bill. The label is doing the selling, not the technology.

This matters because the speed game is where retail investors already win without trying. Every time your broker's smart order router finds you a slightly better price across exchanges, price improvement on your market order, you're benefiting from infrastructure the speed game built. You don't need to play it. You just need to not pay someone who claims they'll let you play it.

The regulatory frame

Because speed advantages are so powerful, they're heavily regulated. Under Europe's MiFID II, Article 17 requires high-frequency firms to store time-sequenced records of every order, and market makers must hold binding written agreements with trading venues, quoting continuously. Regulators know exactly how fast the machines are. They built the rules assuming it.

That regulatory attention is worth keeping in mind for Chapter 12. The rules weren't written for AI specifically. They were written for automated trading generally, and they apply to AI systems just the same. Technology-neutral regulation is the theme of the entire regulatory chapter.

The takeaway

The speed game teaches this book's central humility: the market's hardest edges are structural, not intellectual. They come down to where your servers sit, what data you're allowed to buy, and how many PhDs you employ, not who has the cleverer algorithm. A retail trader's clever algorithm, running on retail infrastructure with retail data, is bringing a chess engine to a fight that's actually about who owns the building the chessboard is in.

So where can the individual investor use AI for real? That's the next three chapters: the retail toolbox, the robo-advisors, and the build-it-yourself path. None of them will make you Jane Street. All of them can make you a better-informed investor, if you know what they're actually for.