Northstar Learning
Original notes on model design, validation, data integrity and the limits of systematic evidence. Articles disclose relevant sources, automation and commercial conflicts.
Research note 04 · Live market anatomy
A violent price move can contain both new information and mechanical position closure. The research problem is not to label every sell-off a liquidation cascade, but to test whether leverage, liquidity and price changed together in a way that alters continuation risk.
Research note 03 · Quant method explained
A short-term reversal can look predictable while remaining impossible to capture. Before improving the forecast, investigate what the recorded price actually represents.
Research note 02 · Research validation
A rigorous model-selection process records every attempted strategy—not just the winner—because repeated searching can turn ordinary noise into an impressive result.
Research note 01 · Model engineering
A practical look at how a trading idea becomes a controlled, testable and continuously monitored decision system.
General educational information only. Not investment advice. Some notes use an automated publication workflow; see the editorial and corrections policy.