Various Stats From the First Round of the Anime Music Tournament

azumanga_daioh_band

I won’t lie; the first round of the Animusic Tourament took a long time to complete and my interest started to flag about three quarters of the way through. Now that we’ve come back around for the second round I’m feeling re-energized and, beyond voting for the second round (don’t forget to go vote here), I wanted to channel this energy into something related. I decided that I’d look into how much was gained by stretching the pool of participants out to 256.

Once I decided on doing some number-crunching, I realized there were several other stats I was interested in finding out and this ballooned the information I needed to gather greatly but I’m happy with the results, which I’ll summarize below.

Before I get further, I wanted to mention that I tried to minimize mistakes on my part but I’m sure there’s probably a couple.

The first step was to create a clean list of all 256 songs in the tournament and I finally hit upon importing the playlist file I created for the downloaded folder of 256 songs into Excel and tidying it up a bit (getting rid of the “.mp3” from each song title, for example). Next, I recorded which songs won their first round matches and double checked that, since I really didn’t want to get this part wrong. Next and finally, I went to ANN and, for each song, I marked if it was an opening/closing/insert song, the year of the anime’s start, and which animation studio created it. (I realize there are a few songs that using the year of an anime’s start won’t be the same as the year an anime song is from but going that next step to figure those out is too detailed for what I need.)

Once I got all that entered into Excel, it was time to start the number-crunching.

Age

Average Age of All Songs –

2005.97

~7.52 years old

Average Age of Round 1 Winners –

2006.29

~7.21 years old

Even a cursory glance at the songs that made it into the Animusic Tournament reveals how few older songs made it, not that that’s really surprising given the turn-over rate of anime fandom. This trend strengthened after the first round with the average age going from 7.52 to 7.21 years old, as measured from the halfway point of 2013. It’ll take a couple weeks to find out if the second round winners continue the trend or not, but I’m guessing they will.

Song Type

Type

# at start

# Win 1st Rnd

Percent at start

Percent after 1st Rnd

Percent Win 1st Rnd

Opening

152

75

59.4

58.6

49.3

Ending

75

39

29.3

30.5

52

Insert

29

14

11.3

10.9

48.3

I know there seemed many opening songs but I didn’t realize it was so heavily weighted that way. The most interesting number I see here is that the ending songs were able to do better at winning in this first round then either opening or insert songs. With the highest seeded insert song only at #15, I wonder how many more rounds will be needed before there are no more insert songs in the tournament.

By Decade and Year

Decade

# at Start

# Win 1st Rnd

% Win 1st Rnd

1970’s

1

1

100

1980’s

8

1

12.5

1990’s

28

15

53.6

2000’s

125

65

52

2010’s

94

46

48.9

The opening to Space Battleship Yamato (helped, no doubt from the new remake) keeps the 1970’s in the fight 🙂 . For all the obvious bias towards newer songs I find it very interesting that songs from the 2010’s won in the first round at a lower rate then either the 1990’s or 2000’s. I wonder if a yearly breakdown will shed some light on things.

Year

# at Start

# Win 1st Rnd

% Win 1st Rnd

1974

1

1

100

1982

1

0

0

1983

1

0

0

1984

2

1

50

1985

2

0

0

1986

1

0

0

1987

1

0

0

1990

1

1

100

1992

1

0

0

1994

2

1

50

1995

3

2

66.7

1996

1

0

0

1997

5

3

60

1998

10

6

60

1999

5

2

40

2000

5

2

40

2001

4

0

0

2002

9

5

55.6

2003

6

4

66.7

2004

18

11

61.1

2005

12

6

50

2006

16

7

43.8

2007

17

10

58.8

2008

18

9

50

2009

20

11

55

2010

23

11

47.8

2011

31

15

48.4

2012

33

18

54.5

2013

7

2

28.6

When it gets broken down by year a very interesting picture takes form. The new anime blog buzzword that all the cool anime blogs are using – recency bias – might be the explanation for 2012 seeing a relatively high percent of winners in the first round but why do all the years around it 2010, 2011, and 2013 show weaker win rates? Plain old nostalgia might explain why the 1990’s and 2000’s generally performed so well. If it was possible to poll the voters I wonder if we would have discovered that the relative high win rates of 1998, 2004, and 2007 corresponded to influxes of new anime fans who are now looking back to these years as the “best years”.

Animation Company

Company

# at Start

# Win 1st Rnd

% Win 1st Rnd

A-1 Pictures

9

6

66.7

AIC

8

4.5

56.3

Bones

15

10

66.7

Brains Base

8

5

62.5

Gainax

8.5

5.5

64.7

Gonzo

8

3

37.5

J.C. Staff

19.5

11

56.4

KyoAni

13

8

61.5

Madhouse

12

9

75

Production IG

17

10.5

61.8

Satelight

10

5

50

Shaft

14

8

57.1

Sunrise

23

6

26.0

xebec

5.5

0

0

Toei

5

0

0

The top part of this list is all the animation companies that placed at least 8 songs into the tournament. I chose 8 because I wanted to see how well the various animation companies are at picking their music over multiple shows and any lower would see a company make the list for a single franchise. The bottom part of the list is the two companies that placed the most songs who failed to have one advance to the second round. Probably the biggest surprise here is that Sunrise placed the most songs into the Animusic Tournament then saw so few advance to the second round. It’s winning rate was less then half that of the other animation companies. The only other one that even remotely came close was Gonzo’s 37.5% rate but, with it’s much lower number of entries, if it had won one more match it would have risen to 50%. I wonder if Sunrise’s poor performance stems from having a small, loyal fan base that had enough votes to get their songs nominated but then not enough votes to get them past the first round. On the other hand,  Madhouse (75%), A-1 Pictures (66.7%), and Bones (66.7%) definitely have the fans to support their songs. It will be interesting to see if these three will continue to perform well in the second round. And while J.C. Staff and Production IG don’t have the highest win rate in the first round they do now have the highest number of songs entering the second round, 11 and 10.5 respectively (or ~1/6 of all the remaining songs between the two).

By Rank

# Win 1st Rnd

% Win 1st Rnd

Top 10

10

100

Top 20

18

90

Top 128

104

81.3

Bottom 128

24

18.8

1 – 50

42

84

51 – 100

41

82

101 – 150

27

54

151 – 200

10

20

201 – 250

8

16

251 – 256

0

0

101 – 128

21

75

129 – 150

6

27.3

I like how the actual rank of the songs where left out in the individual polls because there’s no question now that the seeding system got it right – the stronger songs got seeded high and mostly won their matches. There were upsets; most notably the two top 20 songs that didn’t win in the first round – #17 Roundabout from JoJo’s Bizarre Adventure and #19 Vidro Moyou from Ano Natsu de Matteru. However, the strength of being the higher seed in the individual matches can be seen when breaking down the #101 – 150 section because those ranked #101 – 128 obliterated their lower ranked opponents (those ranked #129 – 156, which basically matches up with the #129 – 150 section) even though their actual ranks were similar.

To answer my original question, seeing as how there were 24 songs from the bottom 128 that made it (including 8 from the lowly ranked #201 – 250 section), I think it was a good idea to stretch the Animusic Tournament out to 256 songs. Though, next time, I’d really be in favor of a format that runs through the bottom 128 quicker.

By Franchise

Franchise

# at Start

# Win 1st Rnd

% Win 1st Rnd

Angel Beats

4

2

50

Aquarion

6

3

50

Aria

7

4

57.1

Monogatari

5

4

80

Cowboy Bebop

6

3

50

Evangelion

5

4

80

Fate/Zero

4

1

25

Full Metal Alchemist

4

3

75

Ghost in the Shell

4

4

100

Macross

6

3

50

Mawaru Penguindrum

4

2

50

Gundam

5

3

60

Railgun/Index

4

2

50

Psycho-pass

3

3

100

Puella Magi Madoka Magica

3

3

100

Steins;Gate

3

3

100

The Melancholy of Haruhi Suzumiya

3

3

100

Last, but not least, I updated my listing of the top franchises. The top part lists all the franchises that had 4 or more songs going into tournament. The Aria franchise placed the most songs – 7 – into the tournament and did better then average in the first round with 4 songs making it to the second round. Also placing 4 songs in the second round is the Monogatari (4 of 5), Evangelion (4 of 5), and Ghost in the Shell (4 of 4) franchises. In fact, the Ghost in the Shell franchise is the only franchise, amongst the 4 and more group, that placed all their songs into the second round. While compiling this list I realized that there were several series that placed 3 songs into the tournament that saw all 3 advance to the second round. Those franchises are listed on the bottom part of the above list. I’m thinking those franchises/series that did perfect, or near perfect, have the fan base to push their songs deep into the tournament. We shall see; expect an update after the second round.

I could go on with more stats; for example, I’m curious how gender of the singer affected which songs got in and then won as well as how solo/duo/group acts fared but that’s a lot more work and I’m not that familiar with the singers themselves behind the songs to easily do it.

I’ll close by mentioning that, after the first round, I’m currently ranked number 1 on Omo’s Fantasy Challonge Bracket for the Animusic Tournament which is cool but also embarrassing because it will be all the more obvious when my pickings fall apart and I end up near the bottom. And I will because my strategy for the first round was to pick the higher seed unless there was a glaring reason not too which worked well for that round but with each round I more and more depended on my gut feeling, which will turn out badly in the end.

Advertisements

2 thoughts on “Various Stats From the First Round of the Anime Music Tournament”

  1. The coolest part of your analysis was this:

    the stronger songs got seeded high and mostly won their matches

    I mean, how often can any tournament make this sort of claim? Looking at stuff like the BCS, World Cup FIFA rankings, and March Madness bracketology, most fail to get this good a level of predictability. So if nothing else, I’m pleased as punch that this system worked as well as it did 😀

    …my interest started to flag about three quarters of the way through.

    Now imagine how it feels to be the tournament organizer. 😉

    Like

  2. @zzeroparticle – If it helps, thank you for organizing this tournament. I’ve had a lot of fun so far and have been introduced to many great songs because of it.

    Like

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s