Where the Game Logs and Box Scores Come From

Season totals say what a player did over a year; game logs say how he got there. Baseball Data Hub's game-level pages, including player game logs, daily game lists and box scores, are built from the game records published by Retrosheet, the volunteer project that has spent decades reconstructing major league games from newspaper accounts, scorebooks and league records.

What Retrosheet provides

The site uses Retrosheet's game-level files: one record per game, with date, teams, score and site; one batting line and one pitching line per player per game; and team totals for each game. It does not currently use Retrosheet's play-by-play event files, so the pages show what each player did in a game but not the sequence of plate appearances.

The game records on the site run from 1898 through the 2025 season, more than 222,000 games in all. The current season's figures come from the MLB Stats API instead, and move into the game-level pages once Retrosheet publishes that season.

Player game logs

Every player with at least ten games in the Retrosheet records has a game log, linked from his profile. There are more than 25,000 such pages. For example, Babe Ruth's game log lists every game on record, one row per game, with the date, opponent, team, result and the player's batting or pitching line.

• Batting and pitching games are listed together, most recent first, 500 games to a page. Longer careers run to several pages, each labelled with its page number and the years it covers.
• Postseason games are listed separately, on the first page, with their own totals. They are not included in the regular-season career totals.

A game log and a profile can differ slightly. A profile's season totals come from the Lahman Baseball Database, which records official season statistics, while the game log adds up individual games as Retrosheet has reconstructed them. For most seasons the two agree exactly. For some early seasons, where Retrosheet has not been able to find every game or has corrected an official figure from the original box scores, they can be a hit or a game apart.

Reading a game-log row

A batting row gives the date, the opponent, the player's team and the game result, followed by the player's at-bats, runs, hits, doubles, triples, home runs, runs batted in, walks, strikeouts and stolen bases. The last column, AVG, is not the batting average for that game. It is his running career average through that game, so reading down the column shows how the career figure moved over time.

A pitching row has the same first columns, then the decision (W for a win, L for a loss, SV for a save, a dash if he had none), innings pitched, hits, runs, earned runs, walks, strikeouts and home runs allowed. The final ERA column is again a running career figure through that game.

Daily game lists

Each date with a major league game has its own page, such as October 8, 1956. There are more than 23,000 of them. A date page lists every game played that day with its final score and links to the box score where one exists, with links to the previous and next dates that had games.

Box scores

Full box scores are published for postseason games: the World Series and every playoff round, and the Negro World Series games in the records. There are about 2,100 of them. The October 8, 1956 date page, for example, links to the box score of Don Larsen's perfect game in Game 5 of the World Series. A box score shows the line score by inning, each team's batting table with totals, and each team's pitching lines.

Regular-season games are listed on the date pages with their final scores but do not have box-score pages of their own.

Finding games

The Advanced Game Finder searches the same game records by season, team, player and type of game, such as no-hitters, cycles, shutouts and extra-inning games. It links each result to its date page and, for postseason games, its box score. The This Day in Baseball History pages cover the same calendar from the side of the players born and the players who died on each day.

Gaps to be aware of

• Coverage of the earliest seasons depends on what Retrosheet has been able to reconstruct from surviving sources.
• Game logs exist only for players with at least ten recorded games.
• The current season appears in profiles and leaderboards before it appears in game logs.

Credit

Retrosheet asks that every use of its data carry this notice, which also appears on the game log and date pages: The information used here was obtained free of charge from and is copyrighted by Retrosheet. Interested parties may contact Retrosheet at 20 Sunset Rd., Newark, DE 19711. The data sources page lists every source the site uses.