The 80% power lie

2026-02-19

This came up in class today and I wanted to repost it: To get NIH funding, you need to demonstrate (that is, convincingly claim) that your study has 80% power. I hate the term “power” as it’s all tied into … Continue reading → …

What do Socrates, Herb Simon, and Andrew Gelman have in common? A love of the oral tradition

2026-02-19

This post is from Bob. With all the hand-wringing over generative AI ruining people’s ability to think, I thought I’d turn the clock back a few millennia, a few decades, then a few minutes. Socrates Let’s start by turning the … Continue reading → …

Quantitative Horse Racing with R: Calibration, Backtesting, and Deployment

2026-02-19

R DuckDB Parquet Calibration Ranking Bayesian Odds TS Backtesting Racing analytics as an inference-and-decision system Thoroughbred flat racing is not a binary classification problem. It is a multi-competitor outcome process with hierarchy (horse / trainer / jockey / track), time dependence (form cycles, market moves), and decision layers (how you act on probabilities). …

Continue reading: Quantitative Horse Racing with R: Calibration, Backtesting, …

You can do more for neural networks in R with {kindling}

2026-02-19

This post has been written in collaboration with Joshua Marie.

Why this post matters
Neural networks in R are no longer niche. Today, we can choose among:

{nnet} for classic, small-scale neural nets,
{neuralnet} another classic neural nets packag…

Continue reading: You can do more …