Reading

i. Foundations of uncertainty

• Frank Knight — Risk, Uncertainty and Profit (1921). Draws the line between risk (known distribution) and uncertainty (unknown). Profit accrues to bearing the second kind. If you trade systematically, the distinction matters.
• Ian Hacking — The Emergence of Probability. Probability dates to about 1654. The founding documents conflate degree-of-belief and relative-frequency. Frequentists and Bayesians are still arguing about which is correct.
• Lara Buchak — Risk and Rationality. Risk-weighted expected utility is not the only rational decision rule. The strongest philosophical case for half-Kelly behavior.

ii. Markets and methodology

• Marcos López de Prado — Advances in Financial Machine Learning. Chapters 11 and 14 (PBO, Deflated Sharpe) are why trading-algo reports a p-value next to its Sharpe.
• Andrew Lo — Adaptive Markets. The cleanest synthesis of efficient-markets and behavioral results. Biology-of-belief framing is generative.
• Nassim Taleb — Fooled by Randomness. Good on the failure modes, oversells the solutions. Skim the second half.

iii. Inference and agents

• Judea Pearl — The Book of Why. Causation isn't statistics; a model that mixes them up mistakes correlation for handle. I rewrote trading-algo's logging after this — separating observed correlations from claimed causes.
• Imre Lakatos — Proofs and Refutations. Dialogue on how mathematicians reason from a wrong proof to a less wrong one. The closest thing to a manual for debugging your own model.
• Daniel Kahneman — Thinking, Fast and Slow. Half of it is the canonical reference for what it claims. The other half failed to replicate — check which is which before citing anything.

iv. Political economy & institutions

• John Rawls — A Theory of Justice. Read while writing the Saoirse PPE chapter. The veil of ignorance is the cleanest tool for institutional design when you don't know which seat you'll occupy.
• Robert Nozick — Anarchy, State, and Utopia. The strongest case against Rawls. Read alongside him.
• Hélène Landemore — Open Democracy. Where the Saoirse executive's sortition seat came from. Argues that a diverse deliberative body outperforms a body of experts on knowledge-production grounds.

v. On deck

Ramsey, Truth and Probability. De Finetti, Theory of Probability. Savage, Foundations of Statistics. Bayesian foundations stack, in that order.

mahimn · reading · rev may 2026