Sources · methodology · data quality
Methodology
How the sources are parsed, how players and games are matched, and how totals are computed.
- Each source is parsed into a normalized schema (14 statistical categories). For each season one source is designated primary (StatCrew reports through 2019, box scores from 2021), and the others are used to validate it.
- Players are matched across seasons by name, with abbreviations, nicknames and last-name-only entries resolved against rosters and the record book. 73 such resolutions were applied; every one is listed under data quality.
- Career totals are sums of season rows; rate statistics (averages, percentages, passer rating) are recomputed from the totals rather than averaged.
- Games are matched across sources by season, date (±3 days) and score; conflicting dates and scores are recorded as notes rather than silently overwritten.
Known gaps
- Individual statistics begin in 1999; earlier careers appear only in the official record book.
- The game log is missing a handful of early-era games against opponents not in the athletics database; seasons whose log does not match the official record are flagged "partial log".
- Wheaton did not field a team in 1918 and the 2020 season was canceled.