Research
The method, written down
Everything here is free to read and shows its work: assumptions, data, and where the approach fails.
Free research. No email required.
Why most backtests are measuring their own assumptions
Look-ahead, survivorship and selection compound quietly, and none of them announce themselves in the equity curve. A worked example of an edge disappearing under bookkeeping.
PJ Landells · 2026-08-29 · 2 min
Read the article
- Prop Firm Intelligence2 min
How prop firms make money
In the evaluation-based prop-firm model the customer is the trader. Understanding the three flows of money — evaluation fees, account fees and the profit split — explains most of the rules, and shows why company-reported payout totals must be treated as claims until verified.
Kernwell Research Desk · 2026-09-17
- Market Structure2 min
How market makers provide liquidity
A market maker quotes both sides and takes inventory risk to earn the spread. Adverse selection, hedging and fragmentation explain why liquidity is never a fixed quantity — and what a widening spread is actually saying.
Kernwell Research Desk · 2026-09-17
- Systems & Engineering5 min
How to automate a trading strategy backtest
A backtest is a measurement of a rule under stated assumptions. Automating it means automating the assumptions — data hygiene, costs, look-ahead, and held-out data — so that every run can be reproduced and every result decomposed.
Kernwell Research Desk · 2026-09-17
- Strategy Development3 min
Solving a problem you don't have
A better trend estimate is worth nothing if no rule consumes a trend estimate. The question that kills most method searches is not whether the method is good, but which decision it changes.
PJ Landells · 2026-09-14
- Systems & Engineering3 min
A number that looks authoritative
We built a market-regime classifier and then refused to put its confidence score on the dashboard. The decision to display a number is separate from the decision to compute one, and it needs its own evidence.
PJ Landells · 2026-09-07
- Market Structure2 min
Reading balance and imbalance in overnight inventory
Overnight inventory constrains what an open can plausibly do. It is a lens, not an indicator, and the difference matters more than the concept.
PJ Landells · 2026-08-29
Stay in the loop
One email a week
What we published, what we’re testing next, and what the result was. Some weeks the finding is that an idea didn’t work — those go out too. Double opt-in, and one click to leave from any email.