Analysis
How to evaluate an alternative data vendor: a working checklist
Point-in-time integrity, provenance, history depth and decay tests — the questions that separate a signal from a spreadsheet.
Analysis: What the Numbers Mean
Analysis coverage from UZU NEWS.
Analysis
What regime-detection models can and cannot do
Hidden Markov and clustering models summarize states that already happened — timely labeling of live regimes is the hard, partly unsolved part.
Analysis
Feature importance is not causation: the mixer problem
Importance scores measure a model's reliance on inputs under correlations — a variable your model ranks first may only be standing in for the real driver.
Analysis
What does it mean for a probability forecast to be calibrated?
Calibration is the match between stated probabilities and observed frequencies — measurable with Brier scores and reliability diagrams, and distinct from…
Analysis
What transaction costs do to a backtest's promised returns
Spread, impact and timing drag compound with turnover — and the strategies they kill are usually the high-Sharape, high-frequency ones.