AFL Brownlow Medal Predictor
Enter the stat lines from a match and get a projected 3-2-1 vote split, ranked by a model fitted to ten seasons of actual Brownlow voting.
AFL Brownlow Medal Predictor
Stat lines from one match
Enter a stat line for each player you want ranked, all from the same match. Mark whether the player was on the winning side. Frees against are entered as a positive count.
Weighted 1.00 each.
Weighted 4.52 each.
Weighted 0.87 each.
Weighted 0.49 each.
Weighted 0.46 each.
Weighted 0.42 each.
Weighted 0.41 each.
Weighted 0.39 each.
Weighted 0.25 each.
Weighted 0.23 each.
Weighted 0.23 each.
Weighted minus 0.15 each.
Weighted 1.00 each.
Weighted 4.52 each.
Weighted 0.87 each.
Weighted 0.49 each.
Weighted 0.46 each.
Weighted 0.42 each.
Weighted 0.41 each.
Weighted 0.39 each.
Weighted 0.25 each.
Weighted 0.23 each.
Weighted 0.23 each.
Weighted minus 0.15 each.
Weighted 1.00 each.
Weighted 4.52 each.
Weighted 0.87 each.
Weighted 0.49 each.
Weighted 0.46 each.
Weighted 0.42 each.
Weighted 0.41 each.
Weighted 0.39 each.
Weighted 0.25 each.
Weighted 0.23 each.
Weighted 0.23 each.
Weighted minus 0.15 each.
Projected vote split
Player A takes the 3 votes on a vote score of 54.04
Player B takes the 2 votes, 1.03 behind on vote score. Player C takes the 1 vote, 4.56 behind the leader. The split is a projection from statistical output rather than the official result.
- 3 votes
- Player A
- 2 votes
- Player B
- 1 vote
- Player C
- Lead over the 2 votes
- 1.03
Highest vote score in the match
Second highest vote score
Third highest vote score
Vote score separating first from second
| # | Player | Vote score | Gap to next | Votes |
|---|---|---|---|---|
| 1 | Player A | 54.04 | 1.03 | 3 |
| 2 | Player B | 53.01 | 3.53 | 2 |
| 3 | Player C | 49.48 | - | 1 |
| Input | Weight |
|---|---|
| Disposals | +1.00 |
| Goals | +4.52 |
| Contested marks | +0.87 |
| Tackles | +0.49 |
| Goal assists | +0.46 |
| Marks | +0.42 |
| Clearances | +0.41 |
| One percenters | +0.39 |
| Inside 50s | +0.25 |
| Rebound 50s | +0.23 |
| Hit outs | +0.23 |
| Frees against | -0.15 |
| Won the match | +6.29 |
Every weight is relative to a disposal, which is set to 1.00, so a goal is worth about four and a half disposals and being on the winning side is worth about six. The winning team term is flat and applied once. The three highest vote scores in a match take the 3, 2 and 1 votes, and where two scores are level the stat line entered first ranks first.
Brownlow votes are awarded by the umpires on their judgement of the match, so this ranks statistical output rather than reproducing an official result. It cannot see a defensive role played on the best forward or one passage of play that decided a match. Eligibility is not modelled, so a player who would not be eligible to receive the medal can still appear in a projected split. Recorded votes for past seasons are in the Brownlow Medal tracker.
How the AFL Brownlow Medal Predictor works
Brownlow votes are 3, 2 and 1 in each home-and-away match, awarded on the judgement of the umpires. No box score reproduces a judgement, so this is a projection built from statistical output rather than a calculation of the official result. The weights above were fitted to ten seasons of actual voting, 2016 to 2025, using a rank-ordered conditional logit. That model matches the real mechanism: umpires rank the best three players within one match, so each player is scored against the field in their own game rather than against a fixed threshold. Every weight is published above so any projected score can be checked against the stat line that produced it.
The weights are expressed relative to a disposal, which is set to 1.00. On that scale a goal is worth about four and a half disposals, and playing on the winning side is worth about six. That figure comes from the vote record: 82.7 per cent of all votes between 2016 and 2025 went to a player on the winning team. What the model cannot see is the part of a performance no box score records, a defensive role played on the best forward, or one passage of play that decided a match. Eligibility is not modelled.
Vote score = disposals + 4.52 goals + 6.29 if the team won + 0.87 contested marks + 0.49 tackles + 0.46 goal assists + 0.42 marks + 0.41 clearances + 0.39 one percenters + 0.25 inside 50s + 0.23 rebound 50s + 0.23 hit outs - 0.15 frees against The three highest vote scores in a match take 3, 2 and 1 votes
Example
Three stat lines from one match. Player A, 30 disposals, 6 marks, 4 tackles, 5 clearances, 1 contested mark, 4 inside 50s, 1 goal assist, 2 goals, 1 free against, team won, scores 54.04. Player B, 32 disposals, 5 marks, 6 tackles, 7 clearances, 3 inside 50s, 2 goal assists, 1 one percenter, 1 rebound 50, 1 goal, team won, scores 53.01. Player C, 28 disposals, 7 marks, 3 tackles, 2 clearances, 2 contested marks, 5 inside 50s, 3 goals, 2 frees against, team lost, scores 49.48. The projection gives A 3 votes, B 2 votes and C 1 vote.
How to use it
- Enter the stat line for the first player, including kicks, handballs, marks, tackles, goals and behinds.
- Mark whether that player was in the winning team.
- Add the other stat lines from the same match.
- Read the vote scores and the projected 3, 2 and 1 split.
AFL Brownlow Medal Predictor FAQ
Who actually awards Brownlow votes?
The umpires do, at the end of each home-and-away match, giving 3, 2 and 1 votes to the three players they judge best afield. That is a judgement, made by people who watched the match, and it is not derived from any statistical formula. It follows that no predictor can calculate the official result, only rank players on what the box score records. Where a projection and the actual votes disagree, the projection is what is wrong, not the umpires.
What weighting does this predictor use?
Weights fitted to ten seasons of actual Brownlow voting, 2016 to 2025, using a rank-ordered conditional logit. Relative to a disposal at 1.00, the fitted values are goals 4.52, a winning team 6.29, contested marks 0.87, tackles 0.49, goal assists 0.46, marks 0.42, clearances 0.41, one percenters 0.39, inside 50s 0.25, rebound 50s 0.23, hit outs 0.23 and frees against minus 0.15. Every weight is printed on the page, so any projected score can be checked against the stat line that produced it.
Why does the winning team count for so much?
Because the vote record says it should. Between 2016 and 2025, 82.7 per cent of all Brownlow votes went to a player on the winning side. The fitted weight of 6.29 is not a judgement applied on top of the model, it is the value the data assigns to being on the winning team, worth a little over six disposals. Two players with near-identical statistical output on opposite sides of a result will be separated by roughly that margin.
Does the predictor handle eligibility?
No. Eligibility rules are not modelled, so a player who would not be eligible to receive the medal can still appear in a projected vote split. The tool ranks statistical output within a match and stops there. If you are building a season projection, apply the eligibility criteria yourself against the official rules rather than assuming the totals shown here already account for them.
How accurate is a vote projection?
Tested by refitting the model without a season and then predicting it, the 3 votes are correct in 48.8 to 66.7 per cent of matches depending on the season, and projected season totals correlate with actual totals at about 0.87. Across the ten seasons the model identifies the actual medallist about four times in ten, and getting a full 3-2-1 order right in a single match happens about 5 per cent of the time. Read it as a shortlist and an ordering rather than a forecast of the count.