An independent research practice, not a trading business.

The premise

Most retail-facing material about markets skips the machinery entirely. It goes straight to patterns, indicators and conviction, and leaves the participant to discover spread, slippage, assignment, decay and position sizing the expensive way — in a live account, with real money, one surprise at a time.

The machinery is not difficult. It is just rarely taught, because it is unglamorous and because explaining it honestly makes the activity sound harder than a sales page would like. EdgeQuant writes it down anyway.

What the work looks like

The subject is price formation: how a price comes to be what it is, what the process of getting there leaves behind, and how much of the conventional account of that survives being checked properly. The work is empirical, it is done on data rather than on opinion, and it is adversarial toward its own results by design.

Beyond that level of description, active lines of enquiry are not discussed publicly — not the questions, the data, the techniques, or the instruments and horizons in scope. That is a settled policy rather than an omission, and the reasoning is set out on the research page. Finished work is published in full with its method attached; unfinished work is not previewed.

Principles

• Adversarial by default. The first job with a promising result is to try to kill it. Permutation nulls, resampled baselines and out-of-sample holdouts come before, not after, the writing-up.
• Costs are part of the result. A gross edge is not an edge. Spread, slippage, financing, borrow and tax are subtracted before anything is called a finding.
• Null results are published. A field that only reports what worked will eventually believe a great many things that are not true.
• Illustrative data is labelled as illustrative. Every figure in the published guides is generated sample data, and says so on the figure itself.
• Regulatory limits are treated as limits. Where rules constrain distribution of trading systems or the giving of advice, the constraint shapes what gets built — it is not something to route around.

The tooling

The research runs on software built in-house rather than bought, because the questions being asked do not fit what is available off the shelf. These are instruments for doing the work, in the way a laboratory builds its own apparatus. They are not products, they are not for sale or licence, they are not described in detail anywhere public, and they do not place orders on anyone else's behalf.

Independence and disclosure

EdgeQuant takes no payment for coverage, runs no affiliate links, and accepts no sponsorship from brokers, exchanges, data vendors or platform providers. If that ever changes it will be disclosed prominently on the page in question rather than in a footer.