The blog
Notes from the library.
Essays, case studies and practical guides on the statistical illusions that turn correct numbers into confident mistakes — every post drawn out in figures.
It takes two: when the real cause is a combination
A smooth dose-response curve can point to a culprit that does nothing on its own. Alcohol and tobacco, asbestos and smoking — and why re-running the correlation never settles it.
Read →Why every backtest works
Every fund track record beats the market and every backtest is a winner — because the losers were deleted. Survivorship bias in fund databases, stock indices and strategy backtests.
Read →Does bacon really cause cancer?
How "processed meat causes cancer" became "bacon is as deadly as smoking" — hazard vs. risk, and the missing baseline.
Read →Survivorship bias: the evidence is in the graveyard
Wald's bombers, fund graveyards, the falling-cats study and founder folklore — how filters write datasets.
Read →Five statistical paradoxes that change how you read the news
Honest numbers, dishonest conclusions — and the five questions that catch them. With five figures.
Read →Simpson's paradox in the wild: three examples that fooled the experts
Berkeley admissions, COVID vaccine tables and a famous baseball season — each reversal drawn out in full.
Read →What a positive test result actually means
Your test is 90% accurate and came back positive. The honest answer is about 9% — counted out in pictures.
Read →