AFL Player Match Statistics 2000-2023

AFL Player Match Statistics, 2000-2023 206,722 player-match records from 4,670 AFL games across 24 seasons, with match context, weather, attendance and the full advanced statistical suite (pressure acts, contested one-on-ones, metres gained, score involvements, ruck contests) alongside AFL Fantasy and SuperCoach scores. One row = one player in one match. AT A GLANCE Rows: 206,722Columns: 81Matches: 4,670Players: 2,599Venues: 27Seasons: 2000-2023Format: UTF-8 CSV, comma-delimited, RFC 4180 quotingSize: approximately 60 MBLicence: CC BY 4.0 WHY THIS ONE Most public AFL datasets stop at the basic box score: kicks, marks, handballs. This one carries the advanced tracking statistics that only became available from the mid-2010s, plus per-match weather and attendance, which makes it usable for venue, fatigue and conditions modelling rather than just counting stats. COLUMNS Match context: match_id, match_date, match_local_start, match_round, venue_name, match_home_team, match_away_team, match_attendance, per-team goals, behinds and score, match_margin, match_winner, match_weather_temp_c, match_weather_type Player: player_id, player_first_name, player_last_name, player_team, player_height_cm, player_weight_kg, player_position, guernsey_number, player_is_retired, subbed Core statistics: kicks, marks, handballs, disposals, effective disposals, disposal efficiency, goals, behinds, hitouts, tackles, rebounds, inside fifties, clearances, clangers, free kicks for and against, contested and uncontested possessions, contested marks, marks inside fifty, one percenters, bounces, goal assists, time on ground percentage, Brownlow votes Advanced: centre and stoppage clearances, score involvements, metres gained, turnovers, intercepts, tackles inside fifty, contested defensive and offensive one-on-ones, defensive-half pressure acts, effective kicks, forward-fifty and general ground ball gets, hitouts to advantage, hitout win percentage, intercept marks, marks on lead, pressure acts, rating points, ruck contests, score launches, shots at goal, spoils Fantasy: afl_fantasy_score, supercoach_score KNOWN LIMITATIONS - READ BEFORE MODELLING These are honest caveats, not defects to discover later. Weather covers 125,278 rows (60.6 per cent), from the 2010 season onward. The match_weather_type and match_weather_temp_c fields are empty for 2000 to 2009. Do not treat missing weather as fine conditions. Advanced statistics begin in 2012, covering approximately 108,000 rows (52 per cent). Pressure acts, metres gained, score involvements, centre and stoppage clearances and the contest one-on-one family are all empty before the 2012 season, when the AFL's tracking provider introduced them. Core statistics such as disposals, marks, tackles, goals and Brownlow votes cover all 206,722 rows from 2000. The match_local_start field has been reconstructed. The source stored a correct local time-of-day against a constant placeholder date; it has been recombined with the real match date. The time-of-day is as supplied, and no timezone conversion has been applied. Missing values are empty, not the string NA. The source used the literal string "NA"; 3.27 million such cells have been converted to true empties so numeric columns parse cleanly. The 2020 season was shortened and relocated. Match lengths and venues that year are not comparable to the rest of the range. Player names are not globally unique. Join on player_id. RESEARCH IDEAS Does rain measurably suppress disposal efficiency, and for which positions? Do interstate travelling teams lose more contested possessions late in matches? Which advanced statistics actually predict Brownlow votes once you control for team success? Venue effects on scoring once weather and attendance are held constant. Projecting AFL Fantasy and SuperCoach scores from prior-form windows. PROVENANCE AND LICENCE Derived from publicly published AFL match statistics, assembled via the fitzRoy R package (https://jimmyday12.github.io/fitzRoy/) and cleaned as described above. Statistics are factual records of public sporting events. Released under Creative Commons Attribution 4.0 International (https://creativecommons.org/licenses/by/4.0/), free to use including commercially, with attribution. Attribution: Bet Better - https://betbetter.world Maintained by Bet Better (https://betbetter.world), which also publishes a free sports model API of win probabilities across nine sports at https://betbetter.world/api/ This dataset contains no betting odds and no wagering data of any kind.