Monday, June 25, 2012

Are The Hawks Doing Better Than The Pies?

Yesterday I mentioned the AFL Power Rankings over at the Big Footy message boards and how they seemed to cop a lot of abuse; I had a look at the updated ratings today to see if the level of vitriol had subsided this week. It hadn’t; the main point of contention this week being that Hawthorn are ranked #1 and Collingwood are ranked #2, despite the Pies being on top of the AFL ladder, and the Hawks sitting in 6th. The rankings on this website would be even more inflammatory to Collingwood boosters: the Magpies sit down in 5th! (And over at Footy Maths, while the Pies sit on top of the current rankings under a proposed alternative ranking system the Hawks would be on top.)

How can this be? I suspect the reasons are fairly similar for all of the ranking systems, but since I can only speak with authority about my own, I’ll look at what is happening there. Ranking points on this site are determined solely by margin of victory or loss, home ground advantage, strength of opposition, and how recent the match was (more recent matches receive higher weighting). At the start of the year, the Pies started slightly ahead of the Hawks.  These are the ranking points they have currently gained from each of the games so far this season:   


V Coll: 1.93

V Geel: 1.46

V Adel: 2.31

V WC: 1.73

V Syd: -1.38

V St.K: 3.65

V Melb: 1.86

V Fre: 1.96

V Rich: -5.10

V NM: 9.22

V PA: 2.20

V BL: 3.74

Total: 23.58


V Haw: 0.42

V Rich: 0.72

V Carl: -2.25

V PA: -1.08

V Ess: 0.45

V WB: 0.95

V BL: 4.03

V Geel: 3.19

V Adel: 2.92

V GC: 2.60

V Melb: 0.30

V WC: 1.32

Total: 13.57

The Hawks have gained 10 more ranking points from this year's matches than the Pies. Almost that entire difference came from Hawthorn’s 115-point demolition of North Melbourne; otherwise the two teams’ performances have been fairly similar. Of the teams that both the Hawks and Pies have played so far this season, the Pies currently have more points from their performances against Geelong, Adelaide, Richmond, and Brisbane, while the Hawks have more points from their performances against Port Adelaide, Melbourne, and West Coast. Yes, the Hawks are considered to have performed better against West Coast even though they lost, because they lost by only 5 points in Perth, while the Pies won by only 3 points in Melbourne (small comfort to Hawks fans probably). The Hawks also picked up significant points from beating St. Kilda and Collingwood itself, while the Pies picked up significant points from beating Gold Coast but lost them from being thrashed by Carlton.  

One thing to consider in assessing the difference between the two teams is that one more goal by Hawthorn in each of its games against West Coast and Geelong and it would currently sit atop the AFL ladder (of course, they didn’t kick those goals, and that is why they are sixth, but it suggests that they were not far away at all from being the most successful team so far in terms of wins). So to answer the question: yes, I do think Hawthorn is doing better than Collingwood despite the Pies being on top of the ladder. But in the end, if Bucks and Maxwell end up lifting the premiership cup it doesn’t really matter what any ranking systems say, does it?      

Sunday, June 24, 2012

AFL Power Rankings: Round 13 2012

The top five stay the same this week, the margins being close enough in the Swans v Cats and the Pies v Eagles matches that none of the teams switched about. Really the only major movers are the Lions and Dogs, who swap places after the former's big win against the latter in Melbourne.

Another 'AFL Power Rankings' system is one that is posted each week over at the Big Footy website. Those rankings are a little different to here, but holy crap, do they get a lot of abuse! Kind of makes me glad that these rankings have relative anonymity. One thing that seemed to be bothering people is that the various ranking systems rate Adelaide well below where they are on the AFL ladder. There probably won't be so many of those comments this week though.

1 (1) Hawthorn 29.4 (29.2)
2 (2) Sydney 22.4 (20.7)
3 (3) West Coast 21.8 (20.1)
4 (4) Geelong 18.9 (20.0)
5 (5) Collingwood 18.3 (18.5)
6 (6) St. Kilda 14.9 (15.0)
7 (8) Carlton 10.1 (10.1)
8 (7) Adelaide 9.2 (11.0)
9 (9) Essendon 8.0 (6.6)
10 (10) Richmond 6.5 (6.2)
11 (11) North Melbourne -5.1 (-7.4)
12 (14) Brisbane -6.2 (-13.4)
13 (13) Fremantle -12.0 (-10.8)
14 (12) Western Bulldogs -16.5 (-10.2)
15 (15) Port Adelaide -20.0 (-19.9)
16 (16) Melbourne -29.9 (-34.1)
17 (18) Gold Coast -53.6 (-53.9)
18 (17) Greater Western Sydney -55.5 (-51.5)

Saturday, June 23, 2012

How Diverse Is Kew Becoming?

For those who are not as nerdly as the Wheatley household, the Australian Bureau of Statistics Census Data was released this week. I asked Ms Wheatley today what she would be interested in me finding out with the data, and she said finding out if Kew, Victoria (the suburb we are currently situated in) is becoming more diverse.

 I had a look at two data series: the percentage of persons in Kew who were born outside Australia, and the percentage of persons in Kew who spoke a language at home other than English. By both measures, Kew has become more diverse over the past decade, as I suspect most Australian suburbs have. Driving this trend is the doubling in Kew's population that were born in China. I haven't yet noticed this trend becoming apparent at the Hotel Kew though.

2001 2006 2011
% born outside Australia 26.5 26.8 28.8
% language spoken at home other than English 19.8 21.4 23.9

Sunday, June 17, 2012

AFL Power Rankings: Round 12 2012

Very little movement this week; Hawks extend their margin at the top and Swans move up to 2nd after the Eagles struggle to beat the Blues at home.

1 (1) Hawthorn 29.2 (28.6)
2 (3) Sydney 20.7 (20.5)
3 (2) West Coast 20.1 (21.0)
4 (4) Geelong 20.0 (20.2)
5 (5) Collingwood 18.5 (18.4)
6 (6) St. Kilda 15.0 (16.1)
7 (7) Adelaide 11.0 (9.7)
8 (8) Carlton 9.5 (8.7)
9 (10) Essendon 6.6 (7.0)
10 (9) Richmond 6.2 (8.6)
11 (11) North Melbourne -7.4 (-4.5)
12 (12) Western Bulldogs -10.2 (-10.1)
13 (13) Fremantle -10.8 (-10.6)
14 (14) Brisbane -13.4 (-12.6)
15 (15) Port Adelaide -19.9 (-19.7)
16 (16) Melbourne -34.1 (-33.8)
17 (17) Greater Western Sydney -51.5 (-53.3)
18 (18) Gold Coast -53.9 (-57.2)

The Home Town Free Kick Bias In The AFL

A suggestion that came into my feed on Twitter was that there is a greater "home ground" bias in relation to free kicks at Subiaco than at other grounds. I thought this would be easy enough to investigate - turns out though that there is no handy compilation of free kicks data to be found on the net (at least not that I'm aware of). Nevertheless I was able to compile some free kicks data for last season to see how the "home ground" bias varied across teams. The method I have used is to compare, for each team, the ratio of free kicks for to free kicks against when playing at home against interstate sides with the ratio of free kicks for to free kicks against when playing away against interstate sides. I used this difference to measure the "home ground" bias because some teams are good at picking up free kicks both home and away (see West Coast Eagles).

The results are in the table below. The first thing to note is that teams playing at home against interstate sides were far more successful at getting free kicks than teams playing away against interstate sides in 2011. But in terms of which teams were more successful at getting free kicks at home compared to on the road there doesn't seem to be much of a pattern. Keep in mind these are small samples, particularly for the Victorian teams. But for 2011 at least, it seems that the Subiaco co-tenants were good at picking up free kicks wherever they played.

Ratio of free kicks for to free kicks against when playing at home against interstate sides (2011) Ratio of free kicks for to free kicks against when playing away against interstate sides (2011) Difference
Carl 1.65 0.87 0.77
Haw 1.31 0.72 0.59
NM 1.17 0.65 0.52
Rich 1.18 0.70 0.48
GC 1.18 0.79 0.40
PA 1.35 0.96 0.39
Geel 1.17 0.87 0.30
Adel 1.10 0.84 0.27
Syd 1.09 0.84 0.25
Melb 0.94 0.70 0.24
Fre 1.24 0.99 0.24
Bris 1.08 0.85 0.23
WB 1.21 0.99 0.22
WC 1.29 1.09 0.19
StK 0.91 0.73 0.18
Ess 0.87 0.87 0.00
Coll 1.08 1.26 -0.18
Total 1.16 0.86 0.30

Thursday, June 14, 2012

Mid-Year AFL All-Australian Team 2012

B: Beau Waters, Darren Glass, Heath Shaw
HB: Brendon Goddard, Ben Rutten, Brett Deledio
C: Brent Stanton, Scott Pendlebury, Jobe Watson
HF: Steele Sidebottom, Lance Franklin, Steve Johnson
F: Taylor Walker, Jack Riewoldt, Lewis Jetta
R: Dean Cox, Scott Thompson, Gary Ablett
I: Ivan Maric, Joel Selwood, Patrick Dangerfield, Josh Kennedy

Wednesday, June 13, 2012

The Finger Points Outwards - No. 42: Does Distracting Free Throw Shooters Work?

Watching the NBA Finals today, and seeing the Thunder fans hold up "Welcome to Brick Town" signs as the Miami Heat players aimed to make free throws, I wondered if players do actually shoot worse from the free throw line when they play away compared to when they play at home.

My suspicion was that they do not, or at least not by much. This article (though it's a few years old now) supports my suspicion.

(If that really is the case, it would be interesting to see in which areas teams actually are worse in when they are away - obviously they must do worse in at least some aspects of the game, because teams win less on the road.)

Monday, June 11, 2012

AFL Power Rankings: Round 11 2012

The Swans move up to third this week - other than that there is not a lot of movement as six teams had the bye. However, even teams that had the bye change their ranking points this week as the value of the teams they recently played changes. For example, St. Kilda's mammoth win against the Gold Coast increases their ranking points and also gives a boost to the points of the teams they recently played, including the Eagles who had the bye (and the Swans).

1 (1) Hawthorn 28.6 (28.3)
2 (2) West Coast 21.0 (20.0)
3 (5) Sydney 20.5 (19.3)
4 (3) Geelong 20.2 (20.0)
5 (4) Collingwood 18.4 (19.9)
6 (6) St. Kilda 16.1 (12.4)
7 (8) Adelaide 9.7 (9.5)
8 (9) Carlton 8.7 (7.1)
9 (7) Richmond 8.6 (10.7)
10 (10) Essendon 7.0 (6.2)
11 (11) North Melbourne -4.5 (-5.2)
12 (12) Western Bulldogs -10.1 (-9.2)
13 (14) Fremantle -10.6 (-14.4)
14 (13) Brisbane -12.6 (-12.6)
15 (15) Port Adelaide -19.7 (-20.4)
16 (16) Melbourne -33.8 (-36.1)
17 (18) Greater Western Sydney -53.3 (-53.4)
18 (17) Gold Coast -57.2 (-54.8)

Sunday, June 3, 2012

AFL Power Rankings: Round 10 2012

After dropping to third last week Hawthorn are back at the top following a Buddy-led demolition of North Melbourne. The Hawks have opened up a bit of a gap from a batch of teams that are bunched together at second to fifth, including last week’s leader, West Coast, who were one of several teams to lose this week to much lower-ranked opponents.

There is some movement in the middle of the rankings as well, with Richmond and Adelaide vaulting into the “top eight”, while Carlton fall from fifth to ninth, and Essendon from eighth to tenth. Some might be wondering how the Bombers, at 8-2 for the season, can be so low. Given that the ranking system is based on “relative margins”, their weak performances against bad opposition cost them in terms of ranking points – if their matches this season against Gold Coast, Port, and Melbourne were removed they would be up in sixth spot. So if you think the Dons’ struggles against the bottom teams don’t really matter, just imagine them as the sixth-ranked team instead.

1 (3) Hawthorn 28.3 (19.6)
2 (1) West Coast 20.0 (23.1)
3 (2) Geelong 20.0 (22.2)
4 (4) Collingwood 19.9 (18.9)
5 (7) Sydney 19.3 (13.2)
6 (6) St. Kilda 12.4 (14.9)
7 (9) Richmond 10.7 (8.0)
8 (10) Adelaide 9.5 (6.7)
9 (5) Carlton 7.1 (15.6)
10 (8) Essendon 6.2 (10.8)
11 (11) North Melbourne -5.2 (2.4)
12 (12) Western Bulldogs -9.2 (-4.8)
13 (14) Brisbane -12.6 (-14.7)
14 (13) Fremantle -14.4 (-12.0)
15 (15) Port Adelaide -20.4 (-28.4)
16 (16) Melbourne -36.1 (-40.2)
17 (18) Greater Western Sydney -53.4 (-56.2)
18 (17) Gold Coast -54.8 (-54.4)

How Best-of-Five-Set Matches Benefit Dominant Male Tennis Players

This post was inspired by an article on The Economist, which went through some possible explanations for why, in recent years, the best male tennis players have dominated Grand Slams more than the best female tennis players. An addendum at the end of the article noted that part of the reason for this is that women’s matches in Grand Slams are three sets, while men’s matches are five sets, therefore making upsets more likely in the women’s game.
Not being able to calculate probabilities in my head, I was curious to see how the different match lengths affects the chances of the “better” player winning the match. Assume that a given male player and a given female player have the same probability of winning a particular set. The tables below show, for selected probabilities of winning a set that are greater than 0.5, the probability of those players winning a particular match, if their probabilities of winning a particular set remains constant. They also show the probability of those players winning a Grand slam tournament (although more caution should be taken with these probabilities, given that a player’s probability of winning a particular set is likely reduced as they progress through the tournament).

Probability of winning match
Probability of winning set

Probability of winning Grand Slam
Probability of winning set

So the fact that men play five sets does make a bit of difference to a dominant male player’s chances of winning a Grand Slam match (and the tournament) compared to a dominant female player’s. But it doesn’t look like it’s enough to primarily explain the greater recent dominance of Federer, Nadal and Djokovic compared to the top female players. Also, as noted in the article, there have been long periods when the top female players have been more dominant (Graf, Seles, Navratilova, Evert, etc), so the set structure just favours the top male players being more dominant, it by no means guarantees it.