Home bias in AFL umpiring: the 10 year tables
Friends, for your viewing pleasure – my opus is finally complete. The last 10 years of Home town umpiring looked at and analysed. Please note my caution that I do not guarantee my stats 100% – even going through the final compile I found 3 errors – but 96% yes. Sorry it is the best I can do with my workforce (One 63-year-old amateur statistician) – and my budget. ($0). I think there are some interesting take-aways – 1) umpires are biased 2) only years 2010-2011 get close to being a level playing field …. and … no, I’ll let you read it is more fun that way Cheers DW (perhaps Adam Spencer could turn his hand to this)
TEN YEAR FREE KICK TABLES
I have now had the opportunity to construct free kick tables going back 10 years looking at two particular statistics over the qualifying rounds (ie: not including finals). The stats are:
1) total free kicks paid to Home and Away Teams
2) games where one team receives 56% or more of frees paid. These are what I term Games of Large Differential. (G.O.L.D.)
These two stats are examined under the sub-categories
HOME VIC v AWAY VIC,
HOME NON-VIC v AWAY VIC,
HOME NON-VIC v AWAY NON-VIC,
HOME VIC v AWAY NON-VIC.
The first thing to note is the obvious – what every football fan knows and what the AFL and its umpiring department refuse to acknowledge: there is a bias to Home Teams.
UMPIRES ALWAYS FAVOUR THE HOME TEAM
This was shown across every category. Of the 80 possible annual entries over the last 10 years the Away Team fared better only three times. In 2007 in Home Vic v Away Vic games, the Away Teams actually got more frees over the year by a very slim margin, 50.18 to 49.82 percent. In 2013 in Large Differential Games between Vic Home v Vic Away, the Away teams got the nod 16-14. In 2012 in Large Differential Games between Vic Home Teams and Non Vic Away Teams, the Non-Vic Away Teams came out the better 12-8.
Two of the 80 entries showed a 50-50 split. In 2010 Non Vic Away Teams did get 50% of GOLD games in their favour v Vic Home Teams (8-8), and in 2007 in Home Vic v Away Vic games, the Away Teams came out 16-16.
So the 80 entries are split 75 to the Home Team 3 to the Away Team, 2 even.
How can anybody possibly argue that umpiring is neutral?
Sometimes the difference is huge.
It is not too bad for Away Victorian Teams playing Home Victorian Teams. The worst they have fared in total percentage frees over the course of a season is 47.7%. But those same teams playing away to Non-Vic Home sides have slipped as low as 46.2% (2008).
Mind you that is still a fair bit better than Away Non Victorian teams. Playing Victorian Home Teams they slipped to a low of 44.4 in the 2016 season just completed. This is the worst stat recorded in any individual entry. Nor is there any relief for Non-Vic Away Teams when they are playing Non-Vic Home Teams, in fact overall it is probably worse. Five years out of possible 10, they don’t crack 47% of the annual free kicks paid in those games. When you are talking 53-47, that’s a 6 percent differential and enough to swing any election anywhere.
DO ALL AWAY TEAMS SUFFER EVENLY?
This critical question pre-occupies the mind of the footy fan and given the 2016 Grand Final umpiring display becomes paramount when talking of a level playing field in a National Competition.
The stats over the 10 years indicate that Victorian Away Teams do not fare as badly as Non-Victorian Away Teams.
Firstly, in All-Victorian games, the Away sides, while clearly disadvantaged, over the decade average 48.7 percent of frees. The G.O.L.D. stat also goes against them in this category. Away Victorian teams against Home Victorian teams get only about 40% of G.O.L.D. in their favour.
That’s not good, but compared to Non-Vic sides playing Away to other Non-Vic sides it’s brilliant. Non-Vic teams average only 47.2 percent of frees over the decade when playing other Non-Vic sides. Worse, only 26.7 percent of G.O.L.D. against other Non-Vic Teams go in their favour. That’s about 1 in every 4.
That said, it should be that as long as Non-Vic Teams have an equal amount of Home and Away games against Other Non-Vic Teams, things even out. That is, they are heavily favoured at Home and heavily penalised when Away.
In terms of fairness on a National level then, the absolutely critical testing ground is that comparing Victorian and Non-Victorian Teams when they play one another.
VICTORIAN v NON-VICTORIAN
The umpiring bias between these sides seems to ebb and flow a little. There is not a single year where Away sides, Non-Vic or Vic have received more frees than the Home sides.
When Games of Large Differential (G.O.L.D.) are combined the Home advantage should balance out and leave two fairly even columns. When combined by individual year and examined we get this:
TOTAL GAMES OF G.O.L.D. BY YEAR | NON-VIC GREATER | VIC GREATER
2007 37 | 17 | 20
2008 41 | 26 | 15
2009 42 | 19 | 23
2010 41 | 24 | 17
2011 45 | 26 | 19
2012 44 | 28 | 16
2013 47 | 18 | 29
2014 47 | 20 | 27
2015 42 | 16 | 26
2016 39 | 13 | 26
TOTALS | 207 | 218
The table indicates that there are two trends, the first five years favour Non-Vic, the last 5 years favour Vic teams. Overall the Vics are moderately advantaged over Non Vic Teams. Six out 10 years fall in their favour. Alarmingly for Non-Vic Teams the last four years all fall drastically in favour of the Victorians.
Appearances can be deceptive however. To fully investigate whether Victorian sides or Non Victorian sides have a strong umpiring advantage we need to take into account THE EAGLES FACTOR.
When Victorian fans or coaches moan that their teams fare just as badly away as Non-Vic Teams, they have in their murky mind’s eye, games at Subiaco playing the Eagles. And with good cause.
In the last 10 years the Eagles are 40-2 in G.O.L.D. playing at Home against Victorian sides. Even when playing away they are running 2-1 in their favour (20-10). This extraordinary statistic distorts what happens in general between NON-VIC and VIC teams. (for example Sydney are 19-12 in GOLD games at Home and 7-19 in GOLD games away)
Looking at the previous table of G.O.L.D. excluding the Eagles games, we get this –
TOTAL GAMES OF G.O.L.D. BY YEAR | NON-VIC GREATER | VIC GREATER
2007 28 | 11 | 17
2008 33 | 19 | 14
2009 37 | 16 | 21
2010 33 | 16 | 17
2011 36 | 19 | 17
2012 36 | 20 | 16
2013 41 | 12 | 29
2014 41 | 16 | 25
2015 34 | 10 | 24
2016 33 | 7 | 26
TOTALS | 146 | 206
With Eagles games taken out of the equation the split is 7 years -3 years in favour of the Vic teams but the differential blows out. With Eagles games excluded G.O.L.D. fall nearly 60% in favour of the Victorian teams. And while the Non-Vic sides fail to exceed 20 in any year, the Vics sides exceed 20, five times and every year for the last four.
So there are at least two different trends working simultaneously: The Eagles are overwhelmingly favoured when they play Victorian sides, while Non-Victorian sides (EXCLUDING THE EAGLES) are at a consistent and considerable disadvantage against Victorian sides over the course of the season.
Further we can see this difference is growing and in 2016 it became well-nigh impossible for a Non-Victorian Team to be on the right side of a G.O.L.D. in Victoria (or Tasmania). Prior to the Grand Final G.O.L.D. between Non-Vic Away Teams and Vic Home Teams went 19-2 in favour of Vic Home Teams.
After the grand final this rose to 20-2.
Umpiring, far from being neutral, shows an endemic bias towards Home teams in every situation.
In general when Victorian sides play away they face less bias than when Non-Victorian sides play away.
Total games between Victorian and Non-Victorian teams show a slight positive bias towards Victorian teams, however when games involving the West Coast Eagles are excluded this bias becomes substantial.
The Eagles games reveal substantial positive bias whether they play home or away.
NB: I am not a professional statistician and working alone and in haste am more than capable of the odd error but even with my failings I’d guess my figures are 96% plus correct.