Retrosheet was founded by David W. Smith in 1989 for the purpose of computerizing play-by-play accounts of as many pre-1984 major league games as possible ("Project Scoresheet" started by Bill James and others had started compiling such data beginning with the 1984 season, although the project collapsed a few years later). Retrosheet succeeded beyond the founder's wildest dreams, comprising over 120,000 game accounts by the end of the 2010s. The database has been a boon to statistical and historical researchers, and is the foundation for the game accounts found of Baseball-Reference.com. The availability of complete play-by-play data for the period covered by the site means that a number of heretofore unknown records were discovered, particularly those covering consecutive games, innings or plate appearances. Those records are often said to be "in the retrosheet era", as the data is not yet available to know whether anyone else had performed the feat in earlier years, for which thorough data is still being compiled - if it still exists. Other top-rank sabermetricians such as Dick Cramer and Tom Ruane are also active on the site.
One of the major drivers of Retrosheet's success was that its data was available for free, encouraging its use but also prompting volunteers to add to the database for the purpose of further disseminating information and prompting new research.
By 2012, Retrosheet offered play-by-play data from almost all major league games since 1918, with earlier years being regularly added. It is the premier open-source provider of major league play-by-play data on the Internet, an area that was once the exclusive purview of the Elias Sports Bureau, which was unwilling to share its proprietary data.
- Brian Murphy: "One site's mission to unearth baseball history", mlb.com, April 9, 2023.