Farm Fresh Insights

Sign in Subscribe

Latest — 26 Mar 2026

No Logbook, No Problem: Recovering LATE When You Can't Observe Compliance

Suppose you run a randomized experiment with imperfect compliance. You assigned treatment, measured outcomes, and now you want the LATE. The Wald estimator says divide the ITT by the compliance rate, $\pi = E[D_i \mid Z_i = 1] - E[D_i \mid Z_i = 0]$, and you'

More issues

Epic Children, Mostly Sons

Across 450 parent-child entries drawn from 32 epics and 23 civilizations, 85% of children are sons. 63% of couples have zero daughters, while only 6% have zero sons. I classified each entry by how well-documented the family actually is: mythological (gods like Zeus and Brahma), legendary (human figures with no

Regression to the Mean Model

Shuffle your XGBoost training data and you can get a different model. Not different hyperparameters or different features, just a different row order. With tree_method='hist' and multiple threads, the parallel quantile sketch that determines bin boundaries depends on how rows are chunked across workers (see stableboost

The Other Gender Gap

In 152 of 203 countries with available data, more women than men are enrolled in tertiary education. The median Gender Parity Index, the ratio of female to male gross enrollment, is 1.22. The interquartile range is [1.00, 1.35]. Only 51 countries remain below parity. The population-weighted mean

The Wrong Pairameter

In a trust game, the first mover has an endowment and chooses how much to send to the second mover. Whatever is sent gets multiplied. The second mover then decides what fraction of the multiplied amount to return. Sending more creates more total surplus. Returning more determines how that surplus

I-Ran the Numbers

How has Iran done on HDI, education, and health since 1990, relative to countries that started at roughly the same level? Iran's 1990 HDI was 0.626. To build a peer group, we take every country whose 1990 HDI falls within ±0.020 of that value. That gives

Comparative Advantage: Calibrated Pairwise Win Rates from LLM Judge Scores

Causal Judge Evaluation (CJE; Landesberg, 2025) formalizes a workflow that is already widespread in practice: score every response with a cheap LLM judge, label a small slice with a trusted oracle, learn the judge-to-oracle mapping, and evaluate at scale with calibrated uncertainty. However, calibration is only as informative as the

Liberation and Immanence

In the Indian religions of liberation, the highest good is release. Moksha means release from samsara, the cycle of death and rebirth. The Muktika Upanishad distinguishes several grades of liberation and places one above the rest: videhamukti, disembodied liberation. Buddhism reaches a similar conclusion through different metaphysics. Nirvana is the

Simulating and Evaluating Agentic Systems

Most teams building agentic systems know they need some way to test multi-turn conversations. The agent interprets ambiguous input, picks tools, maintains state, and has to land in the right place at the end. You can't eyeball that. So people reach for what they already know. The most