Learn
Four lessons that take you from "I have never read a chart" to running your own measured reads through an AI agent. Written for beginners, free, and worth reading even if you never buy anything here.
The one idea underneath all of it: a small number you counted beats a big number you were told. Everything on this site is built around that, including the parts where our own results failed.
The path
Read them in order. Lesson 1 comes before any market content on purpose — without it, every figure in the other three is just a number to be impressed by.
- 1
How to read a trading probability 12 min
Base rates, calibration, sample size, p-values in plain words, and why 52–56% is what real edge looks like. Read this first — it is the lens for everything else.
- 2
Day types 10 min
What price action means, the five shapes a session takes, why a probability distribution beats a label, and a practice routine that actually builds the skill.
- 3
Volume-price analysis 12 min
Effort versus result, what Wyckoff and Weis really claimed, the five events — and the scoreboard where two confirmed and two measured reversed against the textbooks.
- 4
Set up your AI agent 8 min
Get a key, install a Skill or paste one prompt, ask a question, and read the answer critically. Every major agent, plus troubleshooting for every error code.
What you will be able to do at the end
- Hear "this setup works 85% of the time" and know the four questions that either defend it or dismantle it.
- Look at a session in progress and describe it as a distribution over five shapes instead of guessing a label.
- Read volume against price movement and know which of the classical readings survived measurement — including the two that measured backwards.
- Ask an AI agent a market question and correctly judge whether its answer came from data or from its own imagination.
A short glossary, so nothing below surprises you
| Term | Plain meaning |
|---|---|
| Base rate | How often something happens with no prediction at all. The score for knowing nothing. |
| Calibrated | When the model says 70%, the thing happens about 70% of the time. A testable property of the numbers themselves. |
| Held-out | Data the model never saw while being built. Scoring on anything else is grading your own homework. |
| Pre-registered | The test was written down before it was run. This is what separates a test from a search. |
| Zero-shot | Applied to an asset the model was never trained on, with nothing re-tuned. |
| Out of distribution | Being used on something unlike its training data. It still produces an answer; nobody has checked whether that answer is any good. |
The numbers you will meet
Every figure across these lessons is measured, on held-out data, and published with its sample size. Here are the headline ones so you know what "good" looks like around here:
That last tile is not an accident of layout. It belongs next to the others, and lesson 1 explains why a published failure is the most persuasive number on the page.
Then: the three products
The lessons stand on their own. If you want to run the measurements yourself, these are the tools they describe.
Brooks Daily Bias $30/mo
Calibrated day-type probabilities for the session in progress, the most similar historical sessions by shape, and the fact-checked claims database.
Weis Wave $30/mo
Volume grouped into waves, the five classical events, and each one's measured win rate and sample size instead of its reputation.
Max Pain free
Options max pain, open-interest walls, put/call ratio and gamma exposure. Public pages, no key, no account, no signup.
Both APIs together are $50/month. Billing is manual — you email, you get a link, a key is issued by hand. See pricing.
And the scaffolding
Quant Data is designed to be used through an AI agent, not by staring at JSON. The three Skills are installable files that teach your agent the endpoints, the symbol rules, and — the part that matters — how to report the answer without inventing confidence that is not in the data.
Start
Lesson 1 — reading a probability Or poke at something free
Educational analytics. Nothing on these pages is investment advice, a recommendation to buy or sell, or a price forecast. Trading involves risk of loss.