Champion Data leaderboard projecting tightest Brownlow Medal race in years

A year after short-priced favourite Tom Mitchell romped to a comfortable Brownlow Medal win, the AFL seems set for one of its tightest finishes in recent memory.

Champion Data's model predicts a tight Brownlow Medal race.

A year after short-priced favourite Tom Mitchell romped to a comfortable Brownlow Medal win, the AFL seems set for one of its tightest finishes in recent memory.

A bevy of big names, including former winners Patrick Dangerfield and Nat Fyfe, look set to battle it out over the final rounds of the home and away season as the race for ‘Charlie’ intensifies.

Geelong’s Tim Kelly, who leads the voting in the AFL Coaches Association award, is a $4 favourite with TAB.

CLICK HERE TO SUBSCRIBE TO THE SACKED PODCAST

But Champion Data figures reveal Fremantle star Fyfe, the 2015 winner, might have his nose in front in the 2019 medal count after Round 18, with two star Cats hot on his heels.

The projected leaderboard shows a log-jam of the game’s biggest names — all midfielders — among the top 10 contenders.

Collingwood ruckman Brodie Grundy, seventh in betting with TAB at $15, is in 15th place on Champion Data’s predictor.

Champion Data's model predicts a tight Brownlow Medal race.

Champion Data’s model has Dangerfield polling in each of the past four rounds, while the only game in that stretch that Kelly is likely to score votes is in Round 15 — when he had 33 disposals and two goals against Adelaide.

Patrick Cripps has also hit a slump, with Champion putting a line through his vote chances from Rounds 13-16.

RACE FOR THE BROWNLOW

Player (club)	Champion Data's predicted votes (after Round 18)	TAB odds
1. Nat Fyfe (Frem)	20.4	$7
2. Patrick Dangerfield (Geel)	19.9	$5
3. Tim Kelly (Geel)	18.8	$4
4. Travis Boak (PA)	17.2	$34
5. Lachie Neale (Bris)	17.2	$4.50
6. Adam Treloar (Coll)	16.3	$34
7. Patrick Cripps (Carl)	15.0	$8
8. Zach Merrett (Ess)	15.0	$34
9. Marcus Bontempelli (WB)	14.4	$13
10. Jack Macrae (WB)	14.4	$81

HOW IT WORKS

Champion Data’s prediction is done using six years of historical data, feeding player match stats into a neural network. Player involvements measured by stats such as SuperCoach points, disposals, hit-outs, clearances, inside-50s, tackles, and contested possessions are considered by the model, as well as the result and final margin for the player’s team. Adjustments are also made for each player’s polling history. Instead of assigning just a 3, 2, 1 for each game, the probability of each player polling in each spot is calculated to find each players’ expected votes for any given game.