Animusic Tourney Nominations – The Numbers

Trumpet

The nominations are in for the Animusic Tourney being hosted by our own zzeroparticle! Over the last few weeks, anime fans have been nominating up to 15 of their favourite anime songs to compete in the tournament. Once it begins, these entries will be whittled down to the very best over several rounds of voting. While we eagerly await the first round of competition, the nominations themselves provide a wealth of data that reveal some interesting tidbits about those who threw their favourite songs into the ring.

In all, 211 people put in 2751 nominations which resulted in 1189 songs getting a nod. 151 nominators (71.6%) chose to submit the maximum of 15 entries, and for some (like me) even that was not enough! Of the 1189 songs put forward, 797 (67%) received only a single nomination. Only 249 songs received three or more nominations and only 133 received at least 5 nods, so while our tastes are diverse (more than 650 different anime had songs nominated), a select group of songs drew most of the attention.

Number of Songs vs Number of Nominations

Number of Songs vs Number of Nominations

The top 100 songs received 1057 of the 2751 total nominations (38.4%) while the top 10 songs pulled in a combined 255 nominations (9.3%). This means that 8.4% of the songs got 38.4% of the nominations while the top 0.84% of the songs received 9.3% of nods.

But enough about nominations. Not all nominations are created equal. Top noms were worth a stunning 35 points while the bottom 5 a piddly 5 points each. Ranking matters. The song that got the second most nominations only placed 7th overall and the song that placed 2nd was 7th in nominations. In all, 40 931 points were allocated. The top 100 songs (8.4%) grabbed a whopping 42.8% of the points with the top 10 songs pulling in a collective 11.2% of points.

The average song may have received 34.4 points, but the median value of 17 points and mode of 5, demonstrates the top-heaviness of the graph.

Distribution of points between nominated songs (red line indicates the top 100 songs)

Distribution of points between nominated songs
(red line indicates the top 100 songs)

So what can we learn about the top 10 songs that pulled in the most nominations and points? Almost all are from recent series with only two coming from before 2000. The average year for these top songs is 2007.2, showing a clear bias for new shows. In fact, 5 of the 10 are more recent than 2010. All 10 are from TV series rather than movies or OVAs. Six are opening themes and the other four are ending themes, meaning no insert songs cracked the top ten in points.

Top 10 Songs:

  • Bakemonogatari – Kimi no Shiranai Monogatari – 663 points (2009)
  • Cowboy Bebop – Tank! – 543 points (1998)
  • Clannad After Story – Toki wo Kizamu Uta – 541 points (2007)
  • Steins;Gate – Hacking to the Gate – 488 points (2011)
  • Puella Magi Madoka Magica – Magia – 466 points (2011)
  • Shinsekai Yori – Wareta Ringo – 414 points (2012)
  • Ano Hi Mita Hana no Namae wo Bokutachi wa Mada Shiranai – secret base ~Kimi ga Kureta Mono~ (10 years after ver.) – 397 points (2011)
  • Neon Genesis Evangelion – A Cruel Angel’s Thesis – 377 points (1995)
  • Angel Beats – My Soul, Your Beats – 354 points (2010)
  • Spice and Wolf – Tabi no Tochuu – 344 points (2008)

It is possible, however, that these powerhouse songs are outliers. In identifying larger trends, a greater sample size is needed. The top 100 songs in terms of points may give a better picture of the tournament as a whole.

These top 100 songs represent 83 different anime, but the recency bias is still clearly present, even in the larger sample. The average year of these 83 anime is 2006.2 with the top 50 songs running an average year of 2007.0. Other than a few select tracks that are able to withstand the test of time, the more recent an anime is certainly helped its songs get more points. Of the 83 series that produced the top 100 songs, only 12 are from pre-2000 and only 4 from pre-1997. The overall standard deviation is just 6.43 years.

Number of Anime per year

Number of Anime per year

But the age of an anime is not the only factor in its songs being nominated. Popularity goes a long way as well. How many anime fans can say they don’t know Tank! or A Cruel Angel’s Thesis even though they are older than the other winners? To investigate the popularity of the anime that produced the top 100 songs, I headed over to MAL to gather some data on viewership (as an interesting side note, not one of the 83 series received a MAL score lower than 7 and the average was 8.19).

MAL is by no means a complete record of everyone who has seen a certain anime, but it gives a good sense of the relative popularity. Unlike the age results, the popularity results are less consistent, though they do show an overall trend of more popular shows having their songs ranked higher.

Number of viewers for the anime of the top 100 songs

Number of viewers for the anime of the top 100 songs
(red line is the trend line)

The average viewers who heard the top 100 songs was 76,899 and the average for the top 50 songs was 85,309. But while overall popularity did help, there were several less popular series that were able to launch highly ranked songs, demonstrating that the quality of the song itself trumps the popularity of the anime. Four anime with less than 10 000 views according to MAL got songs on this elite list and songs from some very popular series with over 200 000 viewers did not make the top 20 (Code Geass – Colors – 32, Fullmetal Alchemist – Brothers – 52, Elfen Lied – Lilium – 81).

Songs from anime with less than 10 000 views that made the top 100:

  • Space Battleship Yamato – Uchuu Senkan Yamato
  • Legend of Black Heaven – Cautionary Warning
  • Turn A Gundam – Tsuki no Mayu
  • Macross: Do you remember love – Ai Oboete Imasu ka

It’s interesting to note as well that the 12 series from pre-2000 that I mentioned before had a combined average of just 53 296 views with the lowest having only 2118 (Space Battleship Yamato, 1974).

A final factor to consider in the results is the identity of the 211 nominators. Though anonymous, the tournament was promoted mostly through the aniblogging community rather than throughout the larger anime fandom. Anibloggers tend to focus as a collective on currently airing and newer anime, leading to the recency bias, but many also take the time to watch classic and less well-known anime. I believe the tourney benefited from a group of nominators with more diverse anime experience.

As the voting begins it will be interesting to see if the results match up with the nominations. Regardless of which song ends up winning, the tournament will surely make for some good listening. I know for myself, as a fan with an already large collection of anime music and sheet music, this can only be a good thing. 

(Special thanks to zzeroparticle for providing the data and offering advice)

12 Responses to “Animusic Tourney Nominations – The Numbers”

  1. Couple things.

    > Songs from anime with less than 10 000 views that made the top 100:

    I think this is most noteworthy being that one out of the 4 did not belong–Macross, Gundam and Yamato are AAA IPs. Kachou Ouji? LOL. Maybe this is kind of like Elfen Lied in its western bias except isolated to the Sykes song.

    Second, recency bias re: Secret Base is a peculiar thing. While the aniblogging community generally are not well-versed in popular jpop songs, Secret Base is one of the few that does break out to the west in general, so I wonder if that had a major impact on its popularity. Of course it being recently used in a popular anime definitely helps, to reinforce that particular thesis.

  2. Remember that folks could submit multiple nomination ballots if they wanted to. I tested one with Polar Bear Cafe songs just to see if I could. It might be more worthwhile to count the total votes the preliminary round gets for the true number of followers Animusic Tourney has, since that is just a little harder to vote multiple times (just a little).

  3. Uhm, you say bloggers have a recency bias, on the other hand you say the aniblogging community has a more diverse anime experience as if that’s fact. I will go on a limb here and say anime bloggers inherently watch less anime because they tend to spend more time writing and maintaining their blogs and interacting with the community. I will agree that bloggers probably invest more time in comparing anime attributes, such as songs, so while I’m not sure there is a better suited community for this selection phase, there’s no ground to elevate your community above the rest.

  4. @cyth
    I think it’s fair to say that aniblogging types tend to watch more anime than…”the community.” Now the real question is what is this community etc. I don’t disagree with your reasoning that anime bloggers watch less anime because they do other things, like eat food, go to work, or blog or communicate with other fans. But this varies greatly from person to person and from “community” or sub-communities to other groups of people. Uttil you tell us what the average is, how can we take that reasoning seriously? It’s like in baseball, you gotta have a WAR or some such before you can say blogging time displaces anime watching time.

    Now where does the assumption that anime blogging types watch more than the community? I can’t speak for Kaikyaku but I think anyone watching more than 3-4 shows a season probably beats the average individual in the community.

  5. @Omo I don’t think it has anything to do with “food making” or “post writing”.

    I generally think more anibloggers “try” many shows but how many they end up watching is pretty few. There are not 75,000 anime bloggers and the top 100 songs (constituting 80 different anime) each had an average number of viewers just slightly over that. I don’t think all those views are coming from anime bloggers. It is my experience that casuals will stick with shows that are immediately discarded in the blogger community. If I assume they also watch the good shows then the picture starts looking in favor of the “community”. My case would be say Blood-C which despite being heavily panned has 23,000 completes but that is part of an established franchise so more people may have stuck around anyway. So how bout Ro-Kyu-Bu! Original Anime about loli’s playing B-ball: 11,824 completes. Kurogane no Fucking Linebarrels has 10,260 completes. I don’t know many bloggers who own up to having completed it. Hell even fucking Miracle Train has 3661 completes.

    I wouldn’t discount that anime bloggers are still watching these and just not saying anything but even if they were, there aren’t 3000 bloggers that I know of. But then do you branch out the blogging community also to their viewers and commenters? Would even that be enough to match some of the numbers set by the more watched series?

  6. @omo I thought that one was a little odd too.

    @kadian1365 Are you saying you… cheated? And on something as deadly serious as the animusic tourney!

    @cyth & omo & taka Perhaps I should have been more clear, but by aniblogging community I meant the broader spectrum of those that write, read and discuss anime through blogs and other social media (since that’s how the tournament was promoted, rather than though something like a magazine). It is certainly far from a well-defined group, but I thought it was necessary to consider where the nominations came from. My guess is that they generally came from people who watch fansubs, keep up with current/newer series and have a good general understanding of the breadth of the anime industry, as opposed to other people who might only watch dubbed anime on TV or might only have seen the most popular shows, for example. The fact that the tops songs don’t directly correspond to the most watched series suggests that the nominators had seen more than just the most famous series. I was trying to highlight some of the pros and cons of having the noms come from this group of people and illustrate that the results reflect the anime watching preferences of the community that generated them.

  7. Recency bias shows up a lot in talk about the results, but I’m actually surprised how little of that bias there is. The top 100 song age average is 7 years? Compare that to the average “blog generation” – 2 years. Or character-based tournaments like ISML, where 2006 is not the average, but the current cut-off point for the oldies division instead.

    I think that were the tourney not blogger-centric, we’d get stronger bias for series up to 2-3 years old. A lot of people get into anime for a short time only, after all.

  8. 2011 was a great year, and so are Hacking to the Gate, Magia and even Secret Base. The latter and Toki wo Kizamu Uta have also a strong emotional power for those who were receptive to Anohana and Clannad.

    A Cruel Angel’s Thesis and Tanks are legendary.

    “Kimi no Shiranai Monogatari” is a very good piece of Pop/Rock, soft and more musically melodious than brainlessly catchy, with a super nice touch of piano harmony, I totally understand the appeal.

    “My Soul, Your Beats” is another Jun Maeda / Lia entrance, I think they deserve it, even if I find Kanon’s “Last Regrets” and “Kaze no Tadori Tsuku Basho” much more meaningful, innovative, or well-constructed, and of course Air’s Tori no uta and Aozora superior. For me, their best songs by far.

  9. […] My Thesis on one being the loneliest number. A rundown of top-slot songs that only got one vote. Behind the Nihon Review on recency bias reflected in the nominations […]

  10. @Taka

    I watched both of those shows, just as a point of reference. I am also not recording my views on MAL.

    I think as far as viewing anime, bloggers are probably as representative as any other group of people in terms of what they have watched, if not a little more diverse than the average single community.

    Recency bias is pretty strong I think. ISML is a good comparison as I think it’s out of similar reasons why they created a separate category for old characters–people are biased with ‘width’ but the classics tend to stand out (either because they pass the ultimate test (of time) or because of the impact of casuals).

  11. […] in music (:)), so at least my potential nominations might have made things interesting, although an analysis of the results of the nominations point out that such a decision might have been moot. This is too […]

  12. […] results. But taking into account the tournament as a whole so far, along with the few analyses and musings offered by some bloggers on their own sites, I hardly saw any mention of lyrics besides the […]

Leave a Reply

Gravatar enabled.