Analysis (Un)conscious Bias and the Brownlow - What Changed in 2010?

Remove this Banner Ad

Feb 17, 2010
6,517
12,671
adelaide
AFL Club
Port Adelaide
As humans, we all display bias on a daily basis whether we are aware of it in our decision making or not. Conscious and unconscious bias affects every decision we make and that extends to the Brownlow and how umpires interpret the game and votes.

Over the past number of years, the perception has been that the Brownlow has become a midfielders award and that the number of votes received by the winners has been continually increasing, as evident by the record highest votes achieved twice in the last 4 years (36).

Obviously the voting system has not changed over this period so is there a reason why we are seeing the vote tally change and higher votes are being received by the winner?

One possibility is that this is due to unconscious bias on the umpires behalf due to the increased media attention in the game. The AFL has become a 24/7 media giant in recent years and it is impossible for the umpires to ignore the outside noise of who is considered the best players in the game and the favorites even if they are not consciously aware of this while casting their votes.

Where some players are constantly in the media spot light, is this creating bias in the voting system in which these players 'stand out' more when they are on the field and in the eyes of the umpires and in turn they receive a greater number of best on grounds?

As the voting system has not changed, i would argue that the number 1 reason for the higher votes has been an increase in 'Best on Ground' 3 votes received by the media spotlight players. Where as in the past, some of these games were only worth a 1 vote or 0 and the overall votes were spread out amongst more players.

The below stats are from the top 10 vote winners each year since 2000. There is a clear change in the votes received by the winner, the top 5 players and there are less key position players which are placing in the top 10.

This change seems to have occurred around 2010, so what happened in 2010 and since then which has caused this change in voting?

Graph 1: Votes received by the winner. As we can see, since 2000 there has been a clear increase in the votes received by the winner. From 2010 onwards, in only 3 seasons did the winner have less than 30 votes. Compared to 2000-2010, where in only 3 seasons did the winner receive 30 votes (including 2010). So what is the reason for this change in the last 10 years where the outcome is inverse in comparison to the previous decade?

1632096412577.png

Graph 2: Average votes received by the top 5 per year. Again we can see, from 2010 onwards there has been an increase in the votes received by the top 5 overall vote placers. With 2021 being the highest average votes at 31.2 for the top 5. From 2000 - 2010, the top 5 overall placed players did not average over 25 votes amongst them once and from 2011-2021, the top 5 averaged over 25 in 8 seasons.

1632096660564.png

Graph 3: Average votes received by the 6th-10th placed players. Again, from 2010 onwards, we can see that the players that place 6th-10th in the overall voting are also receiving more votes. Before 2010, the average votes over this group did not reach 20 once over that decade. From 2011 to 2021, in 7 years they averaged above 20 votes.

1632096815959.png

Graph 4: The number of key position players in the top 10. Prior to 2010, there was 8 seasons in the data pool which included more than 2 key position players in the top 10. From 2010 onwards, in only 2 season was there at least 2 in the top 10 (last one being Grundy in 2019 with 23 votes). What is the reason for this change which adds to the perception it has become a midfielders award? Is it simply that pre 2010, the game included players like Pav, Brown, Riewoldt, Goodes (included as a KPP), Chad Cornes and Franklin who placed a number of times in the top 10. From 2010 onwards, we had Franklin again, Gawn, Grundy, Sandilands and Goldstein. Is it simply that the game does not have any 'old fashion' CHFs in the game anymore? As in recent years, the majority of key position players that place in the top 10 are ruckman.

1632097079035.png

So the question is, since 2010, has the influence of the increased media in the AFL caused a change in how the umpires cast their votes and the results are a result of unconscious bias?

Or alternatively, is this change from 2010 onwards the result of the expansion teams with Gold Coast joining in 2011 and GWS in 2012? As the talent pool was further spread amongst a greater number of teams, did this highlight the 'stars' of the game more and therefor resulted in those players receiving more votes?

Or is there another reason for this change?
 
Last edited:
I was about to post in the Brownlow thread asking why the winners seem to be getting more and more votes every year, despite the voting system not changing. I was starting to think maybe the best players are literally getting better in comparison to their peers. You've put it much better than me and come up with a much better theory. Surely it's hype and bias - you can't get away from media praise of the top players if you're an AFL umpire.
 

Log in to remove this ad.

Think you've touched on most of the points I would put it down to without thinking about it too much:

- 2010 as an inflection point lines up with the introduction of expansion teams leading to a wider distribution of top-end talent
- Change in defensive structures (especially reduction in taggers) allowing players to consistently rack up huge possession counts (16 of the top 20 career average disposals (min. 50 games) played in the last decade, and the individual game/season records have almost all been set recently too)
- Increased media coverage and narrative forming on certain players
 
Is there a correlation (Ron The Bear ) between the average number of disposals per match, and the votes of the winners/top 5 etc.

The reason I ask is that teams/ players are now trying to control movement of the ball - flicking it around and then distributing it forward. Is there a relationship between the number of disposals across a match and the voting ?? As it would suggest as teams do this, then the time a midfielder spends in the umpires spotlight would increase and then their chances of getting votes does too.
 
Is there a correlation (Ron The Bear ) between the average number of disposals per match, and the votes of the winners/top 5 etc.

Year3 votes2 votes1 vote
1984232221
1985242221
1986232221
1987232221
1988242322
1989252322
1990242322
1991252422
1992252322
1993252321
1994242121
1995242220
1996252222
1997252321
1998242321
1999252321
2000252322
2001242221
2002242121
2003242321
2004242221
2005252322
2006252323
2007262423
2008272422
2009292623
2010282725
2011292624
2012292725
2013292625
2014292726
2015292725
2016302726
2017292726
2018292726
2019302826
2020242220
2021302725

TBH I've forgotten the statistics methods needed to summarise that.
 
I was about to post in the Brownlow thread asking why the winners seem to be getting more and more votes every year, despite the voting system not changing. I was starting to think maybe the best players are literally getting better in comparison to their peers. You've put it much better than me and come up with a much better theory. Surely it's hype and bias - you can't get away from media praise of the top players if you're an AFL umpire.

It’s impossible for umpires to be truely unbiased I think and this isn’t having a dig at them as it’s just human nature.

Even as supporters I think we all find that when watching the “big game” type players that we notice them more than we will less known players even if they have a greater influence on a game. Those type of players trigger emotions and memories within us which influence our views. I think this is where the bias comes in and the difference between perhaps a 1-2 or 2-3 vote game as the more “unknown” player does not trigger those same reactions and therefor may not get the higher vote.

Which in turn, results in some players finishing with those extra 6/7/8 votes compared to what they used too.
 
Year3 votes2 votes1 vote
1984232221
1985242221
1986232221
1987232221
1988242322
1989252322
1990242322
1991252422
1992252322
1993252321
1994242121
1995242220
1996252222
1997252321
1998242321
1999252321
2000252322
2001242221
2002242121
2003242321
2004242221
2005252322
2006252323
2007262423
2008272422
2009292623
2010282725
2011292624
2012292725
2013292625
2014292726
2015292725
2016302726
2017292726
2018292726
2019302826
2020242220
2021302725

TBH I've forgotten the statistics methods needed to summarise that.

cheers mate. I think the 2020 numbers need to be factored up slightly, due to shorter game time - it's somewhere between 15 - 25% increase from memory.

So the trend seems to have started around the 2007/8 period - where increasing disposals are needed to get in the votes....
 
cheers mate. I think the 2020 numbers need to be factored up slightly, due to shorter game time - it's somewhere between 15 - 25% increase from memory.

So the trend seems to have started around the 2007/8 period - where increasing disposals are needed to get in the votes....

Interesting to see that in recent years to even receive one vote you generally need more disposals than what awarded you 3 votes back in the 80s/90s
 
Is there a correlation (Ron The Bear ) between the average number of disposals per match, and the votes of the winners/top 5 etc.

The reason I ask is that teams/ players are now trying to control movement of the ball - flicking it around and then distributing it forward. Is there a relationship between the number of disposals across a match and the voting ?? As it would suggest as teams do this, then the time a midfielder spends in the umpires spotlight would increase and then their chances of getting votes does too.

You'd need to think of a way to normalise for the increase in overall disposals (assuming there has been one) - i.e. is there a change in the proportion of disposals. If a player in 2005 was getting 10% of his teams disposals at 25 a game, and is still getting 10% of his teams disposals at 35 a game, does that change anything?
 
You'd need to think of a way to normalise for the increase in overall disposals (assuming there has been one) - i.e. is there a change in the proportion of disposals. If a player in 2005 was getting 10% of his teams disposals at 25 a game, and is still getting 10% of his teams disposals at 35 a game, does that change anything?

Agreed - i was starting from a very simplistic question, it needs to have more data and analysis built around it....
 
The talent is more spread out than ever, so the stars will stand out more. Every team has a few and the rest become role players.
 

(Log in to remove this ad.)

Massive umpire bias and vote giving should be taken away from them.

Happens at all levels too.

This season a friend AVG over 35ppg, lead his league in their data points ratings, was selected as rover for the league l team of the season and even got a call up to a VFL club based on that form later in the season.

League B&F comes along and he got 3 votes for the whole season.

Many umpires just don't have a clue about football.
 
Having a further look into how much harder it is for key forwards to gain votes, I've had a look at a couple comparisons between how many goals are required on average for a key forward to get a vote compared to a midfielder needs disposals.

Although this isn’t an overly comprehensive analysis but we can somewhat determine how many goals is roughly equivalent to how many disposals in the view of the umpires.

Too simplify this, I’m only looking at the key forwards, how many votes they received and how many goals they kicked (Stats are only from the H&A season).

Taylor Walker - 9 Votes - 48 goals / 17 games
- Averaged 5.3 goals per vote

Daniher - 4 Votes - 45 goals / 22 games
- Averaged 11.25 goals per vote

Hipwood - 1 Vote - 26 goals / 16 games
- Averaged 26 goals per vote

H McKay - 8 Votes - 58 goals / 19 games
- Averaged 7.25 goals per vote

Peter Wright - 5 Votes - 29 goals / 20 games
- Averaged 5.8 goals per vote

Hawkins - 6 Votes - 54 goals / 22 games
- Averaged 9 goals per vote

Cameron - 6 Votes - 34 goals / 12 games
- Averaged 5.66 goals per vote

B King - 3 Votes - 47 goals / 22 games
- Averaged 15.6 goals per vote

Tom McDonald - 8 Votes - 30 goals / 20 games -
- Averaged 3.75 goals per vote

Larkey - 3 Votes - 42 goals / 22
- Averaged 14 goals per vote

Dixon - 3 Votes - 46 goals / 22 games
- Averaged 15.3 goals per vote

J Riewoldt - 8 Votes - 51 goals / 22 games
- Averaged 6.375 goals per vote

Franklin - 8 Votes - 48 goals / 17 games
- Averaged 6 goals per vote

O Allen - 6 Votes - 28 goals / 21 games
- Averaged 4.66 goals per vote

Darling - 4 Votes - 42 goals / 22 games
- Averaged 10.5 goals per vote

J Bruce - 5 Votes - 48 goals / 20 games
- Averaged 9.6 goals per vote


Average goals per vote amongst key forwards = 9.75 goals per 1 vote.

I am genuinlly surprised that out of all those key forwards, Tom McDonald and Oscar Allen actually average the most votes per goal, while Hipwood, Dixon, Larkey and Ben King are by far the most harshly judged in their performances.

Now when we compared this with how many disposals are needed from midfielders per vote (Only a small sample size from the 4 leading mids in the Brownlow).

Wines - 706 Disposals for 36 votes, or a vote every 19.6 disposals.
Bont - 598 Disposals for 33 votes, or a vote every 18.1 disposals
Oliver - 696 Disposals for 31 votes, or a vote every 22.45 disposals
Walsh - 656 Disposals for 30 votes, or a vote every 21.86 disposals

The average between the top 4 mids is 2656 disposals for 130 votes, or a vote every 20.4 disposals.

So all a midfielder needs to do to get a vote on average is gather 20.4 disposals, while a key forward needs to kick 9.75 goals.

Based on this data, for Harry Mckay to have won the Brownlow he would have had to have kicked 351 goals to have got 36 votes and won the Brownlow.

Now obviously the above data is skewed by certain players (eg, Hipwood ave 26 goals per vote, Ben king ave 15.6 goals per vote and McDonald 3.75), however even if we base it off Mckay's stats alone in which he averaged 7.25 goals per vote as the Coleman Medalist, he would still have needed to kick 261 goals to get him 36 votes.
 
Having a further look into how much harder it is for key forwards to gain votes, I've had a look at a couple comparisons between how many goals are required on average for a key forward to get a vote compared to a midfielder needs disposals.

Although this isn’t an overly comprehensive analysis but we can somewhat determine how many goals is roughly equivalent to how many disposals in the view of the umpires.

Too simplify this, I’m only looking at the key forwards, how many votes they received and how many goals they kicked (Stats are only from the H&A season).

Taylor Walker - 9 Votes - 48 goals / 17 games
- Averaged 5.3 goals per vote

Daniher - 4 Votes - 45 goals / 22 games
- Averaged 11.25 goals per vote

Hipwood - 1 Vote - 26 goals / 16 games
- Averaged 26 goals per vote

H McKay - 8 Votes - 58 goals / 19 games
- Averaged 7.25 goals per vote

Peter Wright - 5 Votes - 29 goals / 20 games
- Averaged 5.8 goals per vote

Hawkins - 6 Votes - 54 goals / 22 games
- Averaged 9 goals per vote

Cameron - 6 Votes - 34 goals / 12 games
- Averaged 5.66 goals per vote

B King - 3 Votes - 47 goals / 22 games
- Averaged 15.6 goals per vote

Tom McDonald - 8 Votes - 30 goals / 20 games -
- Averaged 3.75 goals per vote

Larkey - 3 Votes - 42 goals / 22
- Averaged 14 goals per vote

Dixon - 3 Votes - 46 goals / 22 games
- Averaged 15.3 goals per vote

J Riewoldt - 8 Votes - 51 goals / 22 games
- Averaged 6.375 goals per vote

Franklin - 8 Votes - 48 goals / 17 games
- Averaged 6 goals per vote

O Allen - 6 Votes - 28 goals / 21 games
- Averaged 4.66 goals per vote

Darling - 4 Votes - 42 goals / 22 games
- Averaged 10.5 goals per vote

J Bruce - 5 Votes - 48 goals / 20 games
- Averaged 9.6 goals per vote


Average goals per vote amongst key forwards = 9.75 goals per 1 vote.

I am genuinlly surprised that out of all those key forwards, Tom McDonald and Oscar Allen actually average the most votes per goal, while Hipwood, Dixon, Larkey and Ben King are by far the most harshly judged in their performances.

Now when we compared this with how many disposals are needed from midfielders per vote (Only a small sample size from the 4 leading mids in the Brownlow).

Wines - 706 Disposals for 36 votes, or a vote every 19.6 disposals.
Bont - 598 Disposals for 33 votes, or a vote every 18.1 disposals
Oliver - 696 Disposals for 31 votes, or a vote every 22.45 disposals
Walsh - 656 Disposals for 30 votes, or a vote every 21.86 disposals

The average between the top 4 mids is 2656 disposals for 130 votes, or a vote every 20.4 disposals.

So all a midfielder needs to do to get a vote on average is gather 20.4 disposals, while a key forward needs to kick 9.75 goals.

Based on this data, for Harry Mckay to have won the Brownlow he would have had to have kicked 351 goals to have got 36 votes and won the Brownlow.

Now obviously the above data is skewed by certain players (eg, Hipwood ave 26 goals per vote, Ben king ave 15.6 goals per vote and McDonald 3.75), however even if we base it off Mckay's stats alone in which he averaged 7.25 goals per vote as the Coleman Medalist, he would still have needed to kick 261 goals to get him 36 votes.

I fully expect David King to be quoting these figures in a few days time and claiming them as his "own" research....

Nice work mate !!
 
I fully expect David King to be quoting these figures in a few days time and claiming them as his "own" research....

Nice work mate !!

Haha as long as he takes the hit for any miscalculations then it’s free game!

I certainly didn’t expect the numbers to be that far apart but the fact that key forwards will only get votes when they kick a bag shows it’s almost impossible for one to ever win it again unless something changes.
 
Having a further look into how much harder it is for key forwards to gain votes, I've had a look at a couple comparisons between how many goals are required on average for a key forward to get a vote compared to a midfielder needs disposals.

Although this isn’t an overly comprehensive analysis but we can somewhat determine how many goals is roughly equivalent to how many disposals in the view of the umpires.

Too simplify this, I’m only looking at the key forwards, how many votes they received and how many goals they kicked (Stats are only from the H&A season).

Taylor Walker - 9 Votes - 48 goals / 17 games
- Averaged 5.3 goals per vote

Daniher - 4 Votes - 45 goals / 22 games
- Averaged 11.25 goals per vote

Hipwood - 1 Vote - 26 goals / 16 games
- Averaged 26 goals per vote

H McKay - 8 Votes - 58 goals / 19 games
- Averaged 7.25 goals per vote

Peter Wright - 5 Votes - 29 goals / 20 games
- Averaged 5.8 goals per vote

Hawkins - 6 Votes - 54 goals / 22 games
- Averaged 9 goals per vote

Cameron - 6 Votes - 34 goals / 12 games
- Averaged 5.66 goals per vote

B King - 3 Votes - 47 goals / 22 games
- Averaged 15.6 goals per vote

Tom McDonald - 8 Votes - 30 goals / 20 games -
- Averaged 3.75 goals per vote

Larkey - 3 Votes - 42 goals / 22
- Averaged 14 goals per vote

Dixon - 3 Votes - 46 goals / 22 games
- Averaged 15.3 goals per vote

J Riewoldt - 8 Votes - 51 goals / 22 games
- Averaged 6.375 goals per vote

Franklin - 8 Votes - 48 goals / 17 games
- Averaged 6 goals per vote

O Allen - 6 Votes - 28 goals / 21 games
- Averaged 4.66 goals per vote

Darling - 4 Votes - 42 goals / 22 games
- Averaged 10.5 goals per vote

J Bruce - 5 Votes - 48 goals / 20 games
- Averaged 9.6 goals per vote


Average goals per vote amongst key forwards = 9.75 goals per 1 vote.

I am genuinlly surprised that out of all those key forwards, Tom McDonald and Oscar Allen actually average the most votes per goal, while Hipwood, Dixon, Larkey and Ben King are by far the most harshly judged in their performances.

Now when we compared this with how many disposals are needed from midfielders per vote (Only a small sample size from the 4 leading mids in the Brownlow).

Wines - 706 Disposals for 36 votes, or a vote every 19.6 disposals.
Bont - 598 Disposals for 33 votes, or a vote every 18.1 disposals
Oliver - 696 Disposals for 31 votes, or a vote every 22.45 disposals
Walsh - 656 Disposals for 30 votes, or a vote every 21.86 disposals

The average between the top 4 mids is 2656 disposals for 130 votes, or a vote every 20.4 disposals.

So all a midfielder needs to do to get a vote on average is gather 20.4 disposals, while a key forward needs to kick 9.75 goals.

Based on this data, for Harry Mckay to have won the Brownlow he would have had to have kicked 351 goals to have got 36 votes and won the Brownlow.

Now obviously the above data is skewed by certain players (eg, Hipwood ave 26 goals per vote, Ben king ave 15.6 goals per vote and McDonald 3.75), however even if we base it off Mckay's stats alone in which he averaged 7.25 goals per vote as the Coleman Medalist, he would still have needed to kick 261 goals to get him 36 votes.
A goal shouldn't be worth much more than an ordinary possession, because that's really all it is.
 
The data in this thread are really interesting and not unexpected. It offers a great opportunity to cut the Brownlow count from 3 hrs to 1 hr. Just count the votes for the top 10 midfielders and don’t worry about forwards, defenders etc. You can be sure the winner will be among those 10.
 
A goal shouldn't be worth much more than an ordinary possession, because that's really all it is.

Id say the opposite and that a goal is absolutely more valuable then an ordinary disposal when the whole point of the game is to outscore your opposition and to do that you need to kick goals.

Using McKay as the example again. He averaged a tick over 3 goals a game. Let’s say Carlton averaged 13 goals a game which I think would be close, that means McKay kicks 23% of their goals most weeks.

No midfielder is getting 23% of their teams disposals on their own. Most games are usually about the 350 disposals for a team, say Mitchell gets 35 disposals, that’s only 10% of the teams disposals.

If it was a choice between 23% of your teams score or 10% of their disposals, I think I’d rate the goal kicker as having the better game (unless the mid is also involved in a lot of goal assists or goals themselves as well).

Can’t really place a blanket statement on it either way but currently goal kickers are absolutely being ignored.
 
still same amount of players in each game and same amount of votes awarded.

your logic is flawed.

Yes but fewer players are likely to get votes. If there were less teams and all the good players played for the same teams, the votes would get spread around. With only a handful of guns at every club, the votes usually go the same players.
 
still same amount of players in each game and same amount of votes awarded.

your logic is flawed.

That data certainly supports the theory that there was a change in how votes were given around the 2010 timeframe which is when the expansion teams came in and diluted the talent pool across more teams.

If you don’t buy into this theory, do you have another view on the voting trends the past decade compared to the ones before?
 
That data certainly supports the theory that there was a change in how votes were given around the 2010 timeframe which is when the expansion teams came in and diluted the talent pool across more teams.

If you don’t buy into this theory, do you have another view on the voting trends the past decade compared to the ones before?
media hype
 

Remove this Banner Ad

Back
Top