Free course · With Dylan Siegel

Build your first trading algorithm.

Start with how markets work. Learn to test an idea, write the code with AI tools, and run it on paper. One lesson at a time, with a clear next step.

4 parts20 lessonsNo experience needed

Begin the course →Explore the lessons

Part 1 · 5 lessons

Foundations

Understand the pieces of a market and describe a trading rule in plain English.

  1. 1.1 Start here: how to use this site as a courseThe five-minute orientation. Read this, then part one. 3 min read
  2. 1.2 What a trading algorithm actually doesTurn a trading idea into a short set of decisions a computer can follow. Sample draft
  3. 1.3 Market microstructure, explained like you'll build something with itCharts hide the machine. Learn what's actually happening. 3 min read
  4. 1.4 Why your backtest lies to you (and how to make it stop)A good backtest is an argument, not a result. Here's how to make the argument honest. 4 min read
  5. 1.5 Latency isn't what you think it isThe distribution matters. The tail matters more. 4 min read

Part 2 · 5 lessons

Signals

Turn a market observation into a number you can calculate and inspect.

  1. 2.1 Turn an observation into a signalDefine a number you can calculate before deciding what it might mean. Sample draft
  2. 2.2 Build a simple moving-average signalUse a small price series to see what smoothing changes and what it leaves out. Sample draft
  3. 2.3 Queue imbalance — the cleanest microstructure signal you can buildOne of the simplest signals that still holds up in modern markets. 4 min read
  4. 2.4 Order Flow Imbalance — the signal everyone gets halfway rightOFI is useful. Simple formulas are not enough. 4 min read
  5. 2.5 Decide whether two signals add anythingCheck whether a second measurement brings new information or repeats the first. Sample draft

Part 3 · 5 lessons

Method

Design a fair test and decide what the result actually tells you.

  1. 3.1 Write a question you can actually testChoose one observation, one outcome, and a clear way to compare the two. Sample draft
  2. 3.2 Keep some data out of the design processSeparate developing an idea from checking how it behaves on unseen observations. Sample draft
  3. 3.3 Put costs into the testMake the difference between a price prediction and a hypothetical fill explicit. Sample draft
  4. 3.4 Read a backtest beyond the headline numberAsk where a result comes from and how dependent it is on a few choices. Sample draft
  5. 3.5 A research framework for order-by-order dataMost signals die in the gap between book data and live execution. This is the process that narrows it. 4 min read

Part 4 · 5 lessons

Building

Put the pieces into a small program and follow its decisions on paper.

  1. 4.1 Plan a small program you can understandSplit your first system into data, decisions, paper fills, and a readable log. Sample draft
  2. 4.2 Use AI tools to build one piece at a timeGive the coding tool a precise task and check the behavior it produces. Sample draft
  3. 4.3 Replay data before connecting a live feedRun the same sequence twice and check that it produces the same decisions. Sample draft
  4. 4.4 Follow the decisions in a paper runConnect the pieces while keeping fills and positions explicitly simulated. Sample draft
  5. 4.5 Make the program easy to stop and inspectFinish with clear status, useful logs, and a repeatable paper-run review. Sample draft

After the course

Keep asking better questions.

The course gives you a process. The research shows the questions I’m exploring with it.

Explore the research →