Methodology
How Statgami turns source records into rarity counts, distributions and live findings.
Unit of analysis
A denominator follows the event being measured. Batter and pitcher feats generally use games with the required player line; team shapes use team-games; play events use games with usable play-by-play. The exact denominator is printed on each rarity page.
Coverage windows
Coverage is measured by source layer rather than assumed. A page may begin in 1897, 1908, 1988, 2000 or another year depending on the data it needs. Closed sources are shown as a year range, not as current coverage.
Current-season top-up
Completed historical records are joined to current-season MLB feeds. Current games can appear before the next historical rebuild; when that happens, a page labels its scope and the build timestamp rather than pretending every index refreshes at the same moment.
Rarity score
The score is a display ranking derived from observed frequency. It helps order findings; it is not a probability forecast and it does not account for every era, park or rule difference.
Quality controls
Build checks test well-known examples, impossible counts, coverage windows, denominators and page rendering. Known limitations are published in corrections and limitations.