Methods and provenance

This site is the interactive companion to a full research manuscript. The motivating claim comes from Foolish Bailey’s Foolish Baseball video “The Brewers Aren’t Built for October” and the MLB Tonight segment that summarized it. The site name is a nod to the source, not a verdict on it. The video raises real questions; its approach-based arguments about swing rates and velocity are genuinely open, and they are the subject of a planned follow-up study. The split statistic it leads with, however, does not survive scrutiny, and that is what this site demonstrates.

The analysis covers every Statcast pitch from 2015 through the August 20, 2026 data freeze (8.13 million) and Retrosheet play-by-play from 1997 through 2025 (5.3 million plate appearances), with schedules from the MLB Stats API, identifiers from the Chadwick Bureau register, and FanGraphs leaderboard and pre-season ZiPS exports. Pitcher quality is measured ex ante with a Marcel-style projection; the core model is a plate-appearance-level regression with team random slopes, estimated in two stages and confirmed against a directly-fit mixed model; reliability is assessed by split-half and year-over-year correlation with cluster-bootstrap intervals. Every gate of the analysis committed its validation evidence before the next began.

Every number on this site regenerates from the pipeline in the repository: raw pulls are cached, all randomness is seeded, and the web app’s data files are baked from the model outputs by a single export script. The 2026 pre-registration is frozen at a recorded commit.

Play-by-play information used here was obtained free of charge from, and is copyrighted by, Retrosheet (retrosheet.org). Statcast data courtesy of Baseball Savant. Projections and leaderboard statistics courtesy of FanGraphs. This is independent research and is not affiliated with MLB, MLB Network, or Foolish Baseball.

A NERVANALYTICA project. Contact and other work: nerv-analytic.ai.