Enhanced AFL ladder
At the time of publication, Essendon are clearly the best side in the AFL. (Perhaps I am displaying a tiny little bit of bias here.) They are the best side for two reasons:
- Only recent form matters (and Essendon have been doing much better recently), and
- The AFL fixture is not fair - each team plays 12 teams once, and only 5 teams twice, and some sets of 5 teams are easier to beat than others. i.e. it’s easier to look good with a soft draw (I’m looking at you, Collingwood).
This code can be used to calculate a version of the ladder which aims takes these effects into account. It does this by allowing you to exclude all but the most recent rounds (the parameter rnd_cnt), and scaling the points you get for a win by the quality of your opposition, using the opposing team’s percentage as a proxy for quality (the parameter scale_power).
The points you get for a win are \( 4 r^p \) for a win, and \( 2 r^p \) for a draw, where \( r \) is the opposition’s percentage (as calculated on a standard ladder as a number between 0 and 1) and \( p \) is a parameter used to adjust the impact of the opposition’s percentage on the points awarded.
Setting \( p \) = 0 corresponds to the standard ladder (i.e. \( r^p \) = 1), setting \( p \) = 1 means that beating two teams with \( r \) = 50% is equivalent to beating one team with \( r \) = 100%. I think \( p \) = 2 is a good value, it means that you need to beat 2 teams with \( r \) = 71% (or 4 teams with \( r \) = 50%) to get the points equivalent to beating one team with \( r \) = 100%.
Anyway, below are the results (at the end of round 19, 2018).
Here is the standard ladder:
> super_ladder() # standard ladder
TEAM POINTS SCORE OPPONENT_SCORE PERCENTAGE
0 Richmond 56.0 1754 1273 137.784760
1 West Coast 52.0 1619 1335 121.273408
2 Collingwood 48.0 1680 1422 118.143460
3 Port Adelaide 48.0 1476 1270 116.220472
4 GWS Giants 46.0 1524 1341 113.646532
5 Melbourne 44.0 1868 1467 127.334697
6 Hawthorn 44.0 1631 1329 122.723853
7 Geelong 44.0 1603 1333 120.255064
8 Sydney 44.0 1508 1372 109.912536
9 North Melbourne 40.0 1564 1414 110.608204
10 Essendon 40.0 1523 1482 102.766532
11 Adelaide 36.0 1502 1528 98.298429
12 Fremantle 28.0 1279 1593 80.288763
13 Western Bulldogs 20.0 1219 1737 70.178469
14 St Kilda 18.0 1289 1703 75.689959
15 Brisbane 16.0 1498 1666 89.915966
16 Gold Coast 16.0 1096 1694 64.698937
17 Carlton 8.0 1125 1799 62.534742
With opposition quality taken into account:
> super_ladder(scale_power=2) # ladder allowing for opposition quality
TEAM POINTS SCORE OPPONENT_SCORE PERCENTAGE
0 Richmond 59.236194 1754 1273 137.784760
1 West Coast 52.770018 1619 1335 121.273408
2 Geelong 47.613679 1603 1333 120.255064
3 Sydney 45.272944 1508 1372 109.912536
4 GWS Giants 44.597565 1524 1341 113.646532
5 Essendon 43.314357 1523 1482 102.766532
6 Port Adelaide 42.953202 1476 1270 116.220472
7 Hawthorn 41.424336 1631 1329 122.723853
8 Collingwood 38.570038 1680 1422 118.143460
9 North Melbourne 35.218554 1564 1414 110.608204
10 Adelaide 34.827736 1502 1528 98.298429
11 Melbourne 32.130806 1868 1467 127.334697
12 Fremantle 20.992509 1279 1593 80.288763
13 Western Bulldogs 16.481467 1219 1737 70.178469
14 Brisbane 16.191667 1498 1666 89.915966
15 St Kilda 15.541328 1289 1703 75.689959
16 Gold Coast 14.524166 1096 1694 64.698937
17 Carlton 5.898765 1125 1799 62.534742
Note that Collingwood move from 3rd on the ladder to 9th - hard evidence that Collingwood suck. Melbourne also take a hit, and Essendon do quite well moving up to 6th.
Much has been made of Essendon’s slow start to the season, here’s what the ladder looks like when you only consider rounds 9 to 19 inclusive (i.e. 11 rounds or 10 games when you allow for the byes):
> super_ladder(rnd_cnt=11) # ladder for rounds 9 to 19
TEAM POINTS SCORE OPPONENT_SCORE PERCENTAGE
0 Collingwood 32.0 991 747 132.663989
1 Essendon 32.0 876 711 123.206751
2 Port Adelaide 28.0 784 596 131.543624
3 Richmond 28.0 932 723 128.907331
4 GWS Giants 28.0 884 762 116.010499
5 Melbourne 24.0 1087 777 139.897040
6 Hawthorn 24.0 888 697 127.403156
7 Geelong 24.0 894 757 118.097754
8 Sydney 24.0 827 733 112.824011
9 West Coast 24.0 822 740 111.081081
10 North Melbourne 24.0 921 843 109.252669
11 Brisbane 16.0 912 871 104.707233
12 Adelaide 16.0 725 884 82.013575
13 St Kilda 12.0 775 962 80.561331
14 Fremantle 12.0 655 931 70.354458
15 Western Bulldogs 4.0 593 982 60.386965
16 Carlton 4.0 568 999 56.856857
17 Gold Coast 4.0 526 945 55.661376
I like this even better, Essendon move to 2nd. But what about Collingwood? Surely nobody can deserve to be top of the ladder with their easy draw. (Why rounds 9 to 19? Well may you ask. It could be argued that this is the choice that most favours Essendon … never claimed to be unbiased!)
Finally, let’s include both effects (recent form and opposition quality):
> super_ladder(scale_power=2, rnd_cnt=11)
TEAM POINTS SCORE OPPONENT_SCORE PERCENTAGE
0 Essendon 33.368587 876 711 123.206751
1 Richmond 31.664983 932 723 128.907331
2 GWS Giants 30.972215 884 762 116.010499
3 West Coast 29.617419 822 740 111.081081
4 Collingwood 27.241834 991 747 132.663989
5 Geelong 24.612432 894 757 118.097754
6 Port Adelaide 22.520989 784 596 131.543624
7 Sydney 19.964535 827 733 112.824011
8 North Melbourne 19.382228 921 843 109.252669
9 Adelaide 16.358511 725 884 82.013575
10 Brisbane 16.258232 912 871 104.707233
11 Hawthorn 15.582871 888 697 127.403156
12 Melbourne 11.571230 1087 777 139.897040
13 Fremantle 10.905061 655 931 70.354458
14 St Kilda 10.360829 775 962 80.561331
15 Western Bulldogs 5.578832 593 982 60.386965
16 Gold Coast 5.091703 526 945 55.661376
17 Carlton 1.239275 568 999 56.856857
That’s the result I wanted! Essendon on top where they belong.
If this is a fair indication of the current form in the AFL, Essendon may end up rueing their slow start to the season. Pretty tough to be spectators in a finals series when you look like you could have had an impact.
BTW, there’s not much original thought here … there are online interactive sites that allow calculation of form ladders (although I didn’t see any that allow for opposition quality). This one even allows you to pile together multiple seasons into a single ladder (very interesting).