The New VocaRan, 28 Days Later

ポテチの手 by Mizushiro Kazuya

It’s been 4 weeks since the Weekly Vocaloid Ranking entered its 2010 Season and implemented new changes, namely the MyList Modifer.  As you can see, results were mixed.  Four out of five like that there was change, but half didn’t like this specific change.

Our Vocaloidism number monkeys have now had time to figure out what’s what and to see how Modifier B has shaped the VocaRan landscape.  Prepare to be dazzled with math!

In order to properly analyze the B-Mod, we first have to determine why it’s there in the first place.  My personal hypothesis was that it was to emphasize new songs and to push the Melts and the RomiShins of the world out of the spotlight.  It makes sense in theory; viral videos (which is what these are) truly do spread in a rampant fashion before reaching a certain threshold after which it dies off.

Part of LOL's nicochart with translated column headers

I’ll use “LOL” as an example, since that was the song I used on my paper on viral videos :3  The ratio of MyLists-to-views are immediately quite high, 1 in 4 for the first day, but then drop over the course of the following days and weeks.  This is true for most any other song, and the B-Mod amplifies that.

So does the new Modifier do what it’s supposed to do?  Short answer, yes.

Since the B-Mod, the number of songs older than 2 weeks have generally decreased while the amount of new songs have increased.  An interesting note is that the songs 1- and 2-weeks old stay about the same.

This is all well and good, but it’s not without its side effects.  For example, take a look at overall scoring.

The above is a look at how many songs scored a certain amount of points.  The number of 30k+ and 50k+ scores have noticeably decreased.  In the last 10 weeks of 2009, there was an average of 41 songs of at least 30k and 17 of at least 50k.  Those numbers have dropped to 30 and 14 since.  This is partially attributed to the above notion that the B-mod rewards the newest of new songs, but it can also be chalked up to what is most likely the B-Mod’s biggest criticism.

Because the B-Modifier encourages high MyList scores without repercussion (the 40 pt cap isn’t much of a penalty), MyList spamming is a real possibility.  I’m sure you remember Meltdown, right?  Well, we may have a repeat performance, as indicated by taking the average B Modifier from the Top 30 songs.  I can only hope that the increase in B-Mod average coincides with the number of new songs in the Top 30 (the dip in last week’s average sure seems to imply that).

So in conclusion, the new MyList modifier does what it set out to do, but it’s by no means perfect.  By fixing one issue, it opens itself up to another, but for now it doesn’t appear to be that bad.  Let’s hope it doesn’t get worse.

About: Joe Mello


4 thoughts on “The New VocaRan, 28 Days Later”

  1. *sigh* Because of the flaw mentioned, the new VocaRan system will probably have to be revised again. At least it's handling the other problems better than it used to, I like it when the rankings get good new songs, and having more new songs while decreasing the amount of older songs increases the chance of a new song popping up that I like. The side effects…eh, I can't really understand this technical jargon, but I know that the system is not without its flaws.

      1. Or maybe I just fail at understanding things, take your pick. I don't have a very good track record of understanding things.

  2. While MyList spamming may be potential flaw in the system, I wouldn't say that there is any sign of it so far. Statistically, there is no significant evidence to say that there is any kind of increasing trend in the average B-modifier among the top 30 songs. From what I have seen so far, the trends have been more or less stable. New songs will tend to have lower view counts than the long-runners, but will have large B-modifiers to compensate for that fact. If a song gets to the point where it enters long-running status, it will tend to have a large number of views (20k+) and lower B-modifiers. The fact that the average point score since the ranking system change has decreased is probably based almost entirely on the fact that the long-runners, including those that are in the 31-50 ranks, are losing so much from the modifier moved from 20x to 10x or even 5x, a difference that amounts to 5-10k or larger, depending on the song.

    All in all, I don't think there's anything to worry about at the moment. The system works at getting more new songs up to the top 30 ranking. It may not be perfect, but that would be a rather difficult thing to set up, no?

Leave a Reply

Your email address will not be published. Required fields are marked *