How come we use percentage instead of point difference?

[Bardo State]

I think percentage makes more sense in the AFL. What if you have a weird season where Collingwood play three quarters of their home games in the rain while Fremantle only have one with a shower? Wet weather probably limits the scoring and favours the team in an area with nicer weather. Ground dimensions are also slightly different. So if one team plays games at a smaller ground that results in slightly more congested football, why should that be a measure of which team is better? Minor quarrels but the system makes sense. The positive of a points difference system is that it encourages the best teams to score heavily and try to blow out their opponents.

 

[Beutbrute]

Percentage is better. A points difference would advantage higher scoring teams... There would be an intrinsic advantage to teams that play on fast tracks ie Etihad or mainly dry weather ie Domain etc

 

[Bonkers000]

A bit of (related) history.

Scottish Football used Goal average (same formula as our percentage) as their 'tiebreaker) up until 1971-72. In all those seasons, there was only one time when the champion would have been different had they used Goal Difference. In 1965, Kilmarnock defeated Hearts on Goal Average. If It had been Goal Difference, Hearts would have been champions.

In 1971-72, they changed to Goal Difference. Since then, in all those seasons, there was only one time when the champion would have been different had they used Goal Average. In 1986, Celtic won the Championship on Goal Difference. If they had used Goal Average, the Champions would have been....................Hearts.

Hearts last won the Championship in 1960. If you ever want to get beaten to death in an Edinburgh pub, you know what conversation to start.

Interesting post mate.
I wonder how many times the top 4 or top 8 would of changed had we used the other formula.

 

[blue harvest]

Percentage is better. A points difference would advantage higher scoring teams... There would be an intrinsic advantage to teams that play on fast tracks ie Etihad or mainly dry weather ie Domain etc

This is true to a certain extent, but remember on a good track their opposition is probably scoring more as well.

PS: Perth actually receives more rainfall and more days of rain than Melbourne over the footy season

 

[basashi]

I thought that the theory was that if you used points difference then teams with a lower scoring style or a home ground with wetter conditions would have an advantage over teams that tended to play in more high scoring affairs.

 

[falcons2]

Interesting post mate.
I wonder how many times the top 4 or top 8 would of changed had we used the other formula.

I just had a quick look and the only change in the past 5 years would be in 2014. Sydney and Hawthorn finished 1 & 2. Sydney the greater percentage of 2.1 but Hawthorn the greater points spread +74 on Sydney. Both won 17 games. On the current ladder GWS +2 would replace the Bulldogs. The more games that are played it levels out. So in summary its really a case of half a dozen of one and six of the other. Those looking for complex answers are wasting their time.

 

[Walter H White]

Technically points difference is fairer - when Geelong beat GC on the weekend, the commentators didn't say Geelong beat GC by (168/48*100)=350%, they said Geelong beat GC by 120 points. Surely winning by more, regardless of whether the team is more attacking or defensively orientated is more important than the % of the win - although, if you are a good team, the difference should generally be negligible (this isn't a big issue).

I think the reason the AFL uses %, if not for tradition, is likely to be due to the high scoring nature of the game. Football uses goal difference but the margins are far smaller in nominal terms so it makes logical sense to use goal difference, whereas if a team has +/- 1500 odd points difference in the AFL, it flows easier off the tongue and for our brains to read 150% or 62.5% etc. Either way, not much has changed though.

 

[jaymasha]

Can anyone be bothered to see how the ladder would look right now with points difference instead of %? Would it have significantly impacted any seasons historically?

Yes, I can. Will do last years ladder and the first six rounds of this year

 

[Ichabod Noodle]

In 1965, Kilmarnock defeated Hearts on Goal Average. If It had been Goal Difference, Hearts would have been champions.

You're telling me that somebody other than Rangers or Celtic once won the Scottish Football Comp?

 

[gothecrows]

Technically points difference is fairer - when Geelong beat GC on the weekend, the commentators didn't say Geelong beat GC by (168/48*100)=350%, they said Geelong beat GC by 120 points. Surely winning by more, regardless of whether the team is more attacking or defensively orientated is more important than the % of the win - although, if you are a good team, the difference should generally be negligible (this isn't a big issue).

I'm not sure how it's "technically" fairer at all. It's just a different system. Both take points for and against into account.
People are often saying things like "a team that wins 100-50 should be ahead of someone that wins 60-20", "higher scoring should be rewarded", etc. but that's purely opinion - why shouldn't defense be rewarded? There's no real justification in rewarding attack over defense, other than a personal desire to see more attacking football.

I think the reason the AFL uses %, if not for tradition, is likely to be due to the high scoring nature of the game. Football uses goal difference but the margins are far smaller in nominal terms so it makes logical sense to use goal difference, whereas if a team has +/- 1500 odd points difference in the AFL, it flows easier off the tongue and for our brains to read 150% or 62.5% etc. Either way, not much has changed though.

This bit I agree with - percentage is just easier for a high-scoring game.

 

[mcgarnacle]

considering that the League fixture is uneven, percentage or net pts diff should really be relegated as a tie-breaker. teams on the same points may not play the same teams twice over the year which is an inequality of the single ladder comp points system for finals qualification. using percentage compounds this inequality.

the best tie-breaker is a head-to-head tiebreaker. For the sake of an example, if Melbourne and Richmond are equal on comp points, the tiebreaker is dependent on the outcome of their match(es) during the season. 1 game is easy. 2 games is easy too if they were won by the same team. if they split their two matches, then you can aggregate the scores. if aggregate scores are the same, then go to the next tiebreaker such as percentage.

if Melbourne, Richmond and Collingwood are on equal comp points, the tiebreaker is dependent on the head-to-head results between these three clubs and determine the tie-break via a process of elimination.

competition points has determined a team's standing to the rest of the comp. teams on equal points should be separated based on games against each other.

 

[jaymasha]

Yes, I can. Will do last years ladder and the first six rounds of this year

2016 ladder: currently, the only change is GWS up to 4th, and the bulldogs down to 5th.


NM 134

Gee 284

Syd 197

Wb 192(-1)

Gws 194(+1)

Wce 142

Ade 124

Haw -50

Mel 13

Gc -30

Port -63

Stk -72

Coll -136

Carl -140

Rich -164

Bris -210

Ess-223

Fre-192

 

[gothecrows]

considering that the League fixture is uneven
...
the best tie-breaker is a head-to-head tiebreaker. For the sake of an example, if Melbourne and Richmond are equal on comp points, the tiebreaker is dependent on the outcome of their match(es) during the season. 1 game is easy.

Well, no - your first point about the uneven fixture (which was correct), means that using head-to-head as a tiebreaker is a really bad idea.
Adelaide play Hawthorn once this season, at the MCG - a game Hawthorn won by 3 points.
If that game was played at Adelaide Oval, then it is more likely that Adelaide would have one (by no means a guarantee, but certainly more likely).

In the event that they finish equal on points, why would it be fairer to reward Hawthorn for a 3-point win on its home ground, rather than Adelaide, who has a vastly superior percentage (currently)?

 

[jaymasha]

2015 ladder:
Even with some large point differences, there are no positional changes from last years ladder.

fre 293

wce 758

haw 904

syd 428

rich 362

wb 276

ade 286

nm125

port 128

gee 20

gws -19

coll 116

mel-471

stk-467

ess-554

gc-607

bris-749

carl-829

 

[The Swert]

I know they give the same result, but how come we use percentage while almost every other sport uses point difference?

We probably started using percentage before other sports started using point difference.

It's usage dates back to at least the 1890s. The 1896 VFA Season saw Melbourne and Collingwood equal on points with no formal tie-breaking system in place. An additional play-off match was then needed to settle the premiership. The probably led to the adoption of percentage by the breakaway VFL the next year.

 

[Doc Hattori]

I know they give the same result, but how come we use percentage while almost every other sport uses point difference?

You obviously missed every day at school after grade 2

 

[Collins-Langford-Ayres]

I'm glad a Hawk supporter didn't post this...

 

[mcgarnacle]

Well, no - your first point about the uneven fixture (which was correct), means that using head-to-head as a tiebreaker is a really bad idea.
Adelaide play Hawthorn once this season, at the MCG - a game Hawthorn won by 3 points.
If that game was played at Adelaide Oval, then it is more likely that Adelaide would have one (by no means a guarantee, but certainly more likely).

In the event that they finish equal on points, why would it be fairer to reward Hawthorn for a 3-point win on its home ground, rather than Adelaide, who has a vastly superior percentage (currently)?

because that's the directly comparable data that exists between the two teams in question and removes all the variables contributed by every other team.

Does a Hawthorn vs Adelaide result at the MCG have any less significance if it was a final or Grand Final?

At the end of the day, premierships are determined by winning. They're not determined by winning well, or losing by little.

as i said, single ladder competition points is already flawed under the current fixturing method. percentage or net diff just compounds this using different parameters, but increasing the variables dramatically.

 

[mcgarnacle]

I think percentage makes more sense in the AFL. What if you have a weird season where Collingwood play three quarters of their home games in the rain while Fremantle only have one with a shower? Wet weather probably limits the scoring and favours the team in an area with nicer weather. Ground dimensions are also slightly different. So if one team plays games at a smaller ground that results in slightly more congested football, why should that be a measure of which team is better? Minor quarrels but the system makes sense. The positive of a points difference system is that it encourages the best teams to score heavily and try to blow out their opponents.

all those variables are why head-to-head would be a better tie breaker. you just limiting the variables to the performances of the two (or more) tied teams, against each other. League performance gets them to a tied position, head-to-head performance determines the rank. Winning games is how competition points are earned after all.

 

[Wallaby]

We probably started using percentage before other sports started using point difference.

It's usage dates back to at least the 1890s. The 1896 VFA Season saw Melbourne and Collingwood equal on points with no formal tie-breaking system in place. An additional play-off match was then needed to settle the premiership. The probably led to the adoption of percentage by the breakaway VFL the next year.

Almost all soccer competitions originally used Goal Average (same as our percentage). They changed (most changed) to Goal Difference because they were trying to encourage more attacking play.

The problem with using GD (or Points Difference) in AFL is, as quite a few have pointed out, there are significant advantages in regards to some grounds due to differing conditions. We do not have standardised grounds (far bigger variations than are allowed in soccer), and things like wind direction can play a major part in whether a game is high-scoring or not - again far more than soccer.

We no longer play at the same times - night games have been shown to be slightly less high-scoring than day games (lights? Dew?). Somtimes it is fine Friday night and wet on Saturday. And even if we do play at the same time, now we have indoor grounds. How would you feel if on the last round, you and another club had to play different matches, but at the same time. There was a howling gale, rain, mud etc. One game at the MCG - one game 5 Ks away at Docklands. You both win easily - but who is going to win by more?

I think, with the setup we have now - %age is clearly the best option. Back when we played 6 games on a Saturday afternoon on suburban grounds in Melbourne - Points Difference would have been fine.

 

[gothecrows]

because that's the directly comparable data that exists between the two teams in question and removes all the variables contributed by every other team.

Right... so you just ignore the point completely that even though it is a head-to-head comparison, it is not inherently "fair", since in the case of a single game, one team will potentially have a significant home ground advantage.

You're "removing variables", which makes sense if you're trying to compare two things without any relationship to external factors, but both teams do indeed have a relationship to the external factors (other teams in this case).
The AFL ladder is a ranking of teams as a whole - i.e. against every team, and not just the ones you are equal on points with.

League performance gets them to a tied position, head-to-head performance determines the rank. Winning games is how competition points are earned after all.

You even mention in a later post that league performance gets them to a tied position - ignoring your earlier stated fact the the fixture is uneven.
Two teams might only be tied because one played demonstrably weaker opponents. The result of a single game cannot be used to determine ranking because of this.

 

[Superstar swan]

2015 ladder:
Even with some large point differences, there are no positional changes from last years ladder.

fre 293

wce 758

haw 904

syd 428

rich 362

wb 276

ade 286

nm125

port 128

gee 20

gws -19

coll 116

mel-471

stk-467

ess-554

gc-607

bris-749

carl-829

Interesting that the top team only scored 7 more points than 7th place

 

[shintemaster]

This is true to a certain extent, but remember on a good track their opposition is probably scoring more as well.

True. That assumes however that teams that regularly play on fast, dry tracks (ie. Etihad) aren't inherently or structurally encouraged to be better on those tracks. ie. They are more likely or more capable of scoring big even in comparison to their opponents.

 

[Bardo State]

all those variables are why head-to-head would be a better tie breaker. you just limiting the variables to the performances of the two (or more) tied teams, against each other. League performance gets them to a tied position, head-to-head performance determines the rank. Winning games is how competition points are earned after all.

Then what if three teams are equal on point and Team A> Team B> Team C> Team A on head to head? Just seems a bit too convoluted.

 

[gothecrows]

Interesting that the top team only scored 7 more points than 7th place

No - they scored 250 points less than 7th place. They gave up 257 points less, so the differential was only 7.
Similarly though, the difference in % was only 3%
Hawthorn and West Coast had by far the best % (as could be assumed from the points differential)