Every concept you can poke. Every concept laddering up to how it makes money, in equities, crypto and prediction markets. Honest about what’s still alpha and what’s already dead.
HFT Book is a free, interactive reference on high-frequency and quantitative trading. Each idea (the order book, market making, latency arbitrage, optimal execution) comes with a live illustration you can manipulate until it clicks, links showing how it fits the wider machine, and a plain 2026 read on whether the edge still exists.
Most explanations of HFT are a paywalled paper, a 2011 textbook PDF, a dead blog, or a thin listicle. This is the alternative: one atlas where every concept is explained clearly, shown moving wherever it can be, connected to everything it touches, and dated honestly to 2026. Read it top to bottom, or drop in anywhere.
Every page ends with a “Where this fits” block: what it’s a building block of, what it composes with, where it makes money, and the tool you’d need. Follow the up-links and you climb the money ladder.
The concept map shows every idea and how it connects, with the money-ladder paths highlighted. Start there if you want the shape of the field before the detail.
Where a concept can be shown moving, there’s a live widget: trade against a book, blow up a market-making model, run a latency race. Understanding by manipulation.
Every node is a page; every edge is a link you can follow. The bright path is the money ladder: how a fair-value estimator becomes a market-making business.
The thing missing from every textbook: how does this actually make money? Here, every concept answers it. A fair-value tweak ladders up into a market-making strategy, ladders across into three markets, and ends in the data and code you'd need to run it.
Every page on the site carries its own rung of a ladder like this. Follow the up-arrows and you’re never lost.
Twenty-eight of these across the atlas. Here are three that span the field: a fair-value estimator, a market-making model you can blow up, and a latency race.
Lit venues, Reg NMS, maker-taker rebates: the classic arena the canon was written about.
CEX/DEX, 24/7 books, your own infrastructure, no prime broker: the open venue where a self-taught quant can actually start.
Bounded payoffs, event-driven, thin books: microstructure essentially nobody else has written up cleanly.
Most material on HFT is years out of date and quietly oversells edges that are long gone. We don't. Every technique carries a plain 2026 read (still alpha, commoditised to a utility, dead, or structural) plus what AI changes about it.
Once a concept clicks, the next thing you want is to run it: against real book data, in an honest backtest, with reference code that reproduces the maths on the page. We're building exactly that: clean L2/L3 order-book datasets, a backtest harness, and tested reference implementations. Nothing's for sale yet.