We need to change the way we calculate percentage.

[SuperVolley11]

Wow. This thread. Wow. It’s easier to get % as the winner in a lower scoring game. I’m not sure how it can be seen in any other way.

examples:
50 to 40 w % 125 L % 80
60 to 50 w % 120 L % 83
70 to 60 w % 116 L % 85
80 to 70 w % 114 L % 87
90 to 80 w % 112 L % 88
100 to 90 w % 111 L % 90

so therefore in the current system, it’s actually better to play in a low scoring game which may be the cause of weather or more realistically coaches who understand simple mathematics.

 

[El_Scorcho]

I think I get what the OP is saying despite the thread being torpedoed by his maths. +/- is the worst possible choice. Percentage as it stands is fine, but if you want to make a better system for more defensive scoring teams, just use the SANFL system. I don't think it will actually change the ladder across a season at all but it makes more sense mathematically.

The AFL does: For/Against*100

The SANFL does: For/(For+Against)*100.

So the maximum percentage you can get in the SANFL is 100 percent if you win every game in a shutout. Your percentage is your proportion of the total score in your games.

 

[MowShow]

Whilst I don't think it needs changing, it's pretty entertaining to see that this guy clearly understands mathematics/percentage better than a lot of the people telling him he doesn't understand it.

Edit: And yes, his suggested alternative is also almost certainly worse and more flawed.

 

[BF Tiger]

To demonstrate his point the OP could do his sums as applied to the 1983 season. For the exercise we can ignore W-L.

Three teams at the top of the ladder within 2% of each other, with fourth within 6%.
Fifth and sixth separated by 2%.
Seventh to tenth separated by 2%.
Eleventh and twelfth separated by 1.

Or he could look at 1984 Fitzroy, Geelong and Footscray. All went 11-11.
Fitzroy went 2405/2345 for 102.6%
Geelong went 2112/2239 for 94.3%
Footscray went 1992/2123 for 93.8%.

1993 would be a good season to analyse.

1994 even better, with 3rd to 6th separated by % and 7th to 9th separated by %.

We all look forward to your analysis Seeds .

 

[Seeds]

Wow. This thread. Wow. It’s easier to get % as the winner in a lower scoring game. I’m not sure how it can be seen in any other way.

examples:
50 to 40 w % 125 L % 80
60 to 50 w % 120 L % 83
70 to 60 w % 116 L % 85
80 to 70 w % 114 L % 87
90 to 80 w % 112 L % 88
100 to 90 w % 111 L % 90

so therefore in the current system, it’s actually better to play in a low scoring game which may be the cause of weather or more realistically coaches who understand simple mathematics.

You need to add those one off games to a whole season of results. Thén you see they are significantly underweighted.

 

[Seeds]

Whilst I don't think it needs changing, it's pretty entertaining to see that this guy clearly understands mathematics/percentage better than a lot of the people telling him he doesn't understand it.

Edit: And yes, his suggested alternative is also almost certainly worse and more flawed.

Why? if there is a flaw that makes it worse off I would like to hear it? I have been wrong before. Including in this thread twice already.

 

[sherb]

I liked the good old days in high school footy when we were playing a really weak side and had held them scoreless into the last quarter.

"We've got to let them score to get percentage for this game" was the cry from some in the team.

I didn't try to explain it to them........

 

[Seeds]

I liked the good old days in high school footy when we were playing a really weak side and had held them scoreless into the last quarter.

"We've got to let them score to get percentage for this game" was the cry from some in the team.

I didn't try to explain it to them........

Dang you have found a weakness with me new method. If there is a game where the opposition doesn’t score then you get an NA or 0 for percentage despite winning and scoring yourself.

i think I will put in a clause to rectify this by saying if a team doesnt score the winner gets 200 and the loser gets 0 for the match.

 

[Seeds]

I think I get what the OP is saying despite the thread being torpedoed by his maths. +/- is the worst possible choice. Percentage as it stands is fine, but if you want to make a better system for more defensive scoring teams, just use the SANFL system. I don't think it will actually change the ladder across a season at all but it makes more sense mathematically.

The AFL does: For/Against*100

The SANFL does: For/(For+Against)*100.

So the maximum percentage you can get in the SANFL is 100 percent if you win every game in a shutout. Your percentage is your proportion of the total score in your games.

F** me thats so simple. Why didn’t I think of that.

ok rather then my ad hoc scaling approach so that each game has 200 percent to hand out we just use this above approach for each game and then give each game an equal weight to derive the season overall percentage.

 

[Seeds]

Here is the formula for a seasons percentage

season has n games
i denotes a specific game


season percentage = 100*
[For/(For+Against) i=1 + For/(For+Against) i=2 ... + For/(For+Against) i=n]/n

 

[jim boy]

The fairest system would be to tally up the score in a match them then each team gets a percentage based on that. Then a teams ongoing percentge would be average of their percentage from each week played.
But that won t happen because:
a) far too complex
b) everybody knows the rules and can play in way that boosts their percentage if they so wish and only have themselves to blame if they miss out
c) percentages hardly matter - there would be very few, if any, instances where the choice of end-of-season percentage calculation methodology leads to teams changing positions resulting in a material difference (i.e 8th/9th, 4th/5th)

 

[SuperVolley11]

You need to add those one off games to a whole season of results. Thén you see they are significantly underweighted.

Your assuming that the team has a 1 off low scoring game. That all other 21 games for the season will be 120 to 110 scores.

 

[BF Tiger]

Your assuming that the team has a 1 off low scoring game. That all other 21 games for the season will be 120 to 110 scores.

Correct. Some are close, some are 5-6 goal margins and some will be 10+ goals. A mix of wins and losses. A mix of high scoring close games and low scoring blowouts.

 

[SuperVolley11]

I’d like to show a example of where a more defensive team actually has the advantage of getting a higher %. In 2019, the dogs won 12 games. F - 1941 A - 1810 % 107.2
The dogs play a lot of games indoors at docklands therefore you’d expect them to score higher and be in generally higher scoring games according to you. Also in 2019 the hawks won 11 games. F - 1742 (199 less) A - 1602 (208 less) % 108.7.
The hawks play a majority of their games at the mcg and 4? In Tasmania which I’m assuming would have windy and wet matches more often then not. Pretty clear example of the team being involved with lower scoring having a higher % with a very comparable differential between F and A.

 

[NewMelbournian]

I want a bonus point for getting to 100. We need to do more to reward attack.

 

[Vaccination]

I want a bonus point for getting to 100. We need to do more to reward attack.

The last thing the AFL needs is gimmick bonus points

 

[Seeds]

I want a bonus point for getting to 100. We need to do more to reward attack.

Twenty twenty cricket sucks and has largely destroyed the game of cricket. Think about that.

 

[Seeds]

Your assuming that the team has a 1 off low scoring game. That all other 21 games for the season will be 120 to 110 scores.

1, 4, 8 ,15.

it doesn’t matter as long as it’s not 22.

teams that perform better in wet weather (or when playing against defensive opponents) should not be penalised as they currently are.

#wetweathergamesmattertoo

 

[SuperVolley11]

1, 4, 8 ,15.

it doesn’t matter as long as it’s not 22.

teams that perform better in wet weather (or when playing against defensive opponents) should not be penalised as they currently are.

#wetweathergamesmattertoo

Yet they are not. See my example of hawks and dogs in 2019.

 

[Seeds]

The fairest system would be to tally up the score in a match them then each team gets a percentage based on that. Then a teams ongoing percentge would be average of their percentage from each week played.
But that won t happen because:
a) far too complex
b) everybody knows the rules and can play in way that boosts their percentage if they so wish and only have themselves to blame if they miss out
c) percentages hardly matter - there would be very few, if any, instances where the choice of end-of-season percentage calculation methodology leads to teams changing positions resulting in a material difference (i.e 8th/9th, 4th/5th)

a) not even remotely complex. I just don’t get why you don’t include the opposition score? Is winning 120 to 116 really worth more percentage then winning 105 to 25?

b) no they can’t. You can’t play a style that manipulâtes the problem with percentage because it biases against wins in low scoring games and biases against loses in high scoring games. You don’t know pre game whether you win or lose.

c) that’s just an assumption. We aren’t talking small changes here. Good wet weather sides have probably lost 10-20 percent in some seasons. Likewise bad wet weather sides have probably gained 5-10 percent.

 

[wilsonmatic]

This thread sums up why I don't venture to the main board very often.

 

[Seeds]

I’d like to show a example of where a more defensive team actually has the advantage of getting a higher %. In 2019, the dogs won 12 games. F - 1941 A - 1810 % 107.2
The dogs play a lot of games indoors at docklands therefore you’d expect them to score higher and be in generally higher scoring games according to you. Also in 2019 the hawks won 11 games. F - 1742 (199 less) A - 1602 (208 less) % 108.7.
The hawks play a majority of their games at the mcg and 4? In Tasmania which I’m assuming would have windy and wet matches more often then not. Pretty clear example of the team being involved with lower scoring having a higher % with a very comparable differential between F and A.

i Think you missed the point. This doesnt tell me anything about whether dogs or hawks are better wet weather/defensive game teams. thats what matters. Not the fact the one team plays mostly in dry whilst another teams plays mostly wet. That’s irrelevant to overall percentage.


i.e the current system doesnt pénalisé teams that play in lots of wet weather/ defensive games. It penalises teams that are relatively good at winning wet weather/defensive games and over rewards teams that are relatively better at winning dry/attacking games.

 

[jim boy]

c) that’s just an assumption. We aren’t talking small changes here. Good wet weather sides have probably lost 10-20 percent in some seasons. Likewise bad wet weather sides have probably gained 5-10 percent.

Not really, I once looked at percentage for the cats over a 20 or 30 year period. Only once did it ever really make a substantial difference to ladder position, and then it was quite significant and unlikely to be changed by such a recalculation - percentage is way overrated

 

[BF Tiger]

Still waiting on real-life data.

 

[estibador]

I think we should have a wildcard game for the two teams with best percentage from 8-18 for the last spot in the final 8. Could be 8v9. Could be 11v 13.
Encourages team to not give up in a game and get blown away

Interesting concept. I don't hate it.

I went and looked as far back as 2015 - there's only twice in that period it wouldn't have been 8v9:

2016 would have been 8v10, and
2019 would have been 9v10 and Essendon in 8th would have missed out.

In fact in 2019 you could have invited the best 5 percentages from 8-18 to the wildcards and Essendon still would have missed out.