FULL COURT BUCKETS: 2026 WNBA FINALS DATA
Data through Game 4, October 24, 2026

FILES AND RECORDS
finals-team-games.csv: 8 records, one team in one game.
game4-halves.csv: 4 records, one team in one half of Game 4.
reese-howard-finals.csv: 8 records, Angel Reese and Rhyne Howard in Games 1–4.
series-derived.csv: 4 records, winner/loser comparisons for each game.
finals-games-1-4.json: the same normalized records, metadata and series totals/splits.
sources/all_summaries.json: official box-score statistical snapshots, preserving original field names.
sources/official_halves.json: counts derived from official Game 4 play-by-play, preserving original field names.
calculate-data.mjs: Node.js reproduction script. Run node calculate-data.mjs in any directory. It reads sources beside the script and rewrites its derived CSV/JSON files beside the script. No third-party packages or network access required for reproduction. Optional --source-dir PATH and --output-dir PATH select other locations. The --artifact flag is only for the original authoring environment.

SOURCES
https://www.wnba.com/game/atl-vs-gsv-1042600301/box-score
https://www.wnba.com/game/atl-vs-gsv-1042600302/box-score
https://www.wnba.com/game/gsv-vs-atl-1042600303/box-score
https://www.wnba.com/game/gsv-vs-atl-1042600304/box-score
https://www.wnba.com/game/gsv-vs-atl-1042600304/play-by-play

IDENTIFIERS AND UNITS
game_number: Finals game number, 1–4.
game_id: official WNBA game identifier, retained as a text identifier in JSON.
team/opponent: ATL = Atlanta Dream; GSV = Golden State Valkyries.
result/team_result: W or L for the team’s full-game result.
half: 1 = periods 1–2; 2 = periods 3–4. Half points are points within that half, not cumulative score.
data_cutoff: 2026-10-24. This is the dataset cutoff date, not each game’s date.
minutes: official player playing time in minutes:seconds.
All counts are whole numbers. Points columns measure scoreboard points; attempt columns measure attempts, not scoring possessions or trips to the line.
All percentage columns ending _pct use a 0–100 scale, rounded to six decimal places. Display 42.105263 as 42.1%, not 4,210.5%.
Blank CSV / null JSON percentage: denominator is zero. For example, Howard had zero free-throw attempts in Game 2; the free-throw percentage is undefined, not an observed 0% success rate.

OFFICIAL TEAM FIELD MAPPINGS
In sources/all_summaries.json, each game contains teams.ATL and teams.GSV.
points <- team.score (also reconciled to team.stats.points).
turnovers_player <- team.stats.turnovers (sum of player turnovers).
turnovers_team <- team.postgame.turnoversTeam (turnovers charged to the team rather than a named player).
turnovers_total <- team.postgame.turnoversTotal = turnovers_player + turnovers_team.
points_off_turnovers <- team.postgame.pointsFromTurnovers.
field_goals_made <- team.stats.fieldGoalsMade.
field_goals_attempted <- team.stats.fieldGoalsAttempted.
three_pointers_made <- team.stats.threePointersMade.
three_pointers_attempted <- team.stats.threePointersAttempted.
free_throws_made <- team.stats.freeThrowsMade.
free_throws_attempted <- team.stats.freeThrowsAttempted.
offensive_rebounds <- team.stats.reboundsOffensive.
defensive_rebounds <- team.stats.reboundsDefensive.
rebounds <- team.stats.reboundsTotal.
assists <- team.stats.assists.
steals <- team.stats.steals.
blocks <- team.stats.blocks.
personal_fouls <- team.stats.foulsPersonal.
points_in_paint <- team.postgame.pointsInThePaint.
second_chance_points <- team.postgame.pointsSecondChance.
fast_break_points <- team.postgame.pointsFastBreak.
bench_points <- team.postgame.benchPoints.
Rebounds are the official box-score player-credit totals. They exclude the separate postgame.reboundsTeam counter, which is preserved in the source snapshot but is not added to box-score rebounds.
Player rows use the same statistical mappings under player.statistics; turnovers is that player’s statistics.turnovers and plus_minus is statistics.plusMinusPoints. No team turnovers are assigned to individual players.

DERIVED FIELDS AND EXACT FORMULAS
opponent_points: points in the opposing team’s matched game/half record.
other_points = points - points_off_turnovers.
two_pointers_made = field_goals_made - three_pointers_made.
two_pointers_attempted = field_goals_attempted - three_pointers_attempted.
two_point_points = 2 * two_pointers_made.
three_point_points = 3 * three_pointers_made.
points = two_point_points + three_point_points + free_throws_made.
scoring_margin = points - opponent_points.
turnover_advantage = opponent turnovers_total - own turnovers_total. Positive means the row’s team committed fewer turnovers.
points_off_turnovers_margin = own points_off_turnovers - opponent points_off_turnovers.
other_points_margin = own other_points - opponent other_points.
field_goal_percentage_pct = 100 * field_goals_made / field_goals_attempted.
two_point_percentage_pct = 100 * two_pointers_made / two_pointers_attempted.
three_point_percentage_pct = 100 * three_pointers_made / three_pointers_attempted.
free_throw_percentage_pct = 100 * free_throws_made / free_throws_attempted.
Series totals sum the appropriate records. Per-game averages divide group totals by games. Shooting splits divide total makes by total attempts, never average rounded game percentages.
series-derived.csv winner_* and loser_* fields select the corresponding team-game values by final result. winning_margin is winner_points - loser_points. Its two component margins are winner points-off-turnovers margin and winner other-points margin.

HALF COUNTS
Read official WNBA page __NEXT_DATA__.props.pageProps.playByPlay.actions.
Field-goal attempt: isFieldGoal = 1. Made field goal: additionally shotResult = Made. A three-pointer has shotValue = 3.
Free-throw attempt: actionType = Free Throw. A missed attempt has a description starting MISS; other free-throw attempts count as makes. Counts were reconciled to full-game official totals.
Turnover: actionType = Turnover. For the Atlanta shot-clock turnover, teamId = 0 and personId equals Atlanta’s team ID. Retain that event as a team turnover.
Do not sum pointsTotal, which is not a uniform per-event scoring value. Reconstruct points as 2 * FGM + FG3M + FTM, then reconcile to period-end scores.
Do not deduplicate all events by actionNumber: a turnover and its steal can share that number. actionId uniquely identifies the supplied Game 4 events.
Half offensive/defensive rebounds were counted from increments in player cumulative Off/Def rebound counts in rebound events. Full-game totals reconcile to the box score.
In sources/official_halves.json, zero-valued TOV_TEAM keys may be absent because the source is a counter. Exported half turnovers_team is exactly TOV_TOTAL - TOV_PLAYER.
Half-level points off turnovers were not retrieved. They are omitted, not imputed.

INTERPRETATION
Points off turnovers are already included in total points. Other points is a scoring decomposition, not a prediction of what the score would have been without turnovers. Do not add points off turnovers to total points.
Points in the paint, second-chance points and fast-break points can overlap each other and points off turnovers. Do not sum those categories to reconstruct a score.
No estimated possessions, pace, turnover rates, or FGA + 0.44 * FTA figures are included. All reported shot attempts are literal field-goal attempts or free-throw attempts.
Four games establish a descriptive pattern, not a causal or predictive rule.

REUSE AND CREDIT
Full Court Buckets makes this data compilation and analysis free to use and republish.
Suggested attribution: Full Court Buckets analysis of official WNBA data.
Please credit Full Court Buckets and link to its published data article. The permanent article URL should be supplied by the publisher when this package is posted.

DEEP-DIVE FILES (Games 1-4; source: Full Court Buckets game data, verified against official WNBA box scores and play-by-play)
Build: python3 data/build_halves.py && python3 data/deep/analyze.py && python3 data/deep/zones_crosscheck.py && python3 data/deep/summaries.py ; cross-check: python3 data/deep/crosscheck_deep.py, python3 data/crosscheck_wnbacom_halves.py
finals-halves.csv: first/second-half team stats, same columns as game4-halves.csv.
finals-turnovers.csv: one row per turnover. Columns: game, period, clock, team, player (TEAM = team turnover), type (bad pass, lost ball, offensive foul, travel, shot clock, ...), live_ball_steal (1 if a steal is credited), stolen_by, opp_points_next_24s (points the opponent scored in the 24 game-clock seconds after the turnover, same quarter), and the play text.
finals-turnovers-by-type.csv: turnovers aggregated by team/game/player/type, plus LIVE-BALL and DEAD-BALL rows. Includes opp_points_within_24s and opp_points_per_turnover.
finals-shots.csv: one row per field-goal attempt (blocked shots count as misses for the shooter's team). Zone: rim 0-4 ft or layup/dunk/tip/finger roll/putback; paint/short 5-14 ft or floater/hook; midrange 15+ ft two-pointer; three. zone_source = "api text" or "wnba.com shotDistance" (the 30 attempts with no distance or type in the API text).
finals-shot-zones.csv: FGM/FGA/FG% by team, game (or all) and half (or all) for each zone.
finals-runs.csv: unanswered scoring runs of 8+ points, with start and end clock, the score before and after (ATL-GSV), and the plays that started and ended each run.
finals-lead-tracker.csv: lead changes, ties (after the opening tip) and largest lead for each team in each game.
finals-clutch.csv: Game 1, the last 5:00 of the 4th quarter. The score was within 5 for all of it. Rows for each team and player. Game 4 was within 5 for only part of its last 1:40, so it isn't included.
finals-clutch-plays.csv: every play in the last 5:00 of a 4th quarter when the score was within 5 before the play.