Japan's Weekly Blu-ray & CD Rankings for April 29 - May 05

RyanSaotome said:
I honestly never would have predicted Attack on Titan to sell so well. I was thinking like 5k tops before this season... it really doesn't seem like the kinda show otaku would rally around. I was wrong though...

RyanSaotome said:
How has the reception been for Gargantia Ep 6 in Japan?

RyanSaotome said:
Symbv, how has the reception been for Gargantia Ep 6 in Japan? I'm seeing the same "Its terrible because its all fanservice, dropped" stuff that that I saw for Ep. 5 on MAL and other places, but I wonder if Japan also universally likes this one.

Ejc said:
Stalker with a tabbed design now? =O I guess it looks cleaner now.
What does everyone else think?

symbv said:
It was not there when I used it less than one hour ago... Anyway, it does not affect my use so much, although somehow I slightly prefer the old way as everything is in one page, particularly when I can see anime of two seasons together, but the impact is really very small. I guess it is done to keep one page shorter?

EDIT: Just find that even inside each title's page, the data are now separated into 4 tabs. Hmm.... I definitely prefer seeing everything in one page there, but I guess I can get used to it quite soon..

symbv said:
One more thing I don't like for the new layout: I usually check the final sales prediction of each work by clicking into the titles one by one in the anime list of spring season. Now when I go back, the page is the current ranking by default, so I need to click the "Spring Anime" tab to get to the anime list of spring season. And that sucks.

Ejc said:
Do you use Mozilla Firefox, Google Chrome or a browser with tabs? You could just open everything one by one and switch between them so you don't need to keep going back to the "Spring Anime" tab.

symbv said:
I use Firefox but my Firefox already has >50 tabs open for various things including a bunch of Japanese anime forum, news and blog sites, so I am hesitant to open everything in new tabs.

jmal said:
I usually look at these for the graphs, not the wildly unreliable (at this point) final estimates.

jmal said:
When I say it's unreliable I mean it's not reliable (beyond extremely broad ballparking) at being the thing it's named: 最終予測累積ポイント. At least until you're within ~2 weeks of release. That's why I prefer to look at the graphs, and want them to be on the first page.

jmal said:
I don't have an issue with it being used for showing whether a title has gone up or down in the past week (though it can't tell us why).

jmal said:
Personally I would prefer to look at a title's median rank over the past week but that takes more time/work to unravel since you need to calculate it from the hourly ranking data.

symbv said:
Anyway, Ejc, I think you may have missed one GirlPan DVD in the daily DVD ranking ;-)

Ejc said:
Huh really? I swear I copied today's ranks correctly, lol.

symbv said:
This is because it is a DVD that is very special ;-)

Check the #3 of the DVD daily list http://www.oricon.co.jp/rank/dg/d/2013-05-12/ also and this video



A really really special case I should say, particularly this is not really a GirlPan DVD but also so GirlPan (if you know what I mean).

I am sure if it is not for GirlPan it won't even reach top 20 or even 50.

jmal said:
Oh, stalker just updated and it shows how much the final prediction can change at a moment's notice...

Shingeki v1 BD is now predicted at 98516pt. A 20k jump in an instant, because the latest day got a 2400 point boost for averaging #1, and the small 346 point jump between May 6 and may 7 rolled off and the 2400 point jump from May 13 to May 14 rolled on.

Now the average gained per day is 1,335.14, and the predicted additional points are 84,114 and the final prediction is 98,516.

It's just a number you need to be realllllllly careful with. Not saying you aren't, I know you put caveats in your comments, but it's not just obsessed sales watchers like us who look at these numbers and make assumptions based on them.

Man I really just wish Stalker had a more sophisticated final prediction algorithm.

jmal said:
And while it is most visible at the top tiers, honestly the only thing more casual sales watchers tend to care about are 1. big sellers and 2. big shifts, the two things Stalker's estimations overreact to most.

jmal said:
I'm certainly not telling you how or what to post, as aggregating any degree of this stuff is a lot of work as I know all too well. But if you can find a way to easily provide more statistical context than just the final prediction it'd be helpful to a lot of people. That's all.

jmal said:

Not sure but I always assumed it was based on the average rank through the previous day not just midnight rank. If it's midnight only that's kinda problematic.

jmal said:
What we intend and what people get out of something are often different. Not saying you're responsible for people who don't want to read all the caveats and minutiae, but if we can make things easier by providing more or better metrics, it's worth at least considering.

jmal said:
Not all of Stalker, no, but the final predictions this far out, for titles that sell enough for the majority of people to care, yes.

jmal said:
What it comes down to for me is, if I want to tell people how a title has performed just over the past week, a measurement that only covers a week is better information than one that projects the week for months into the future even if both technically can be used in the same way.

jmal said:
It's very common for first timers to misunderstand the number. Whether they think it's a solid prediction, or misunderstand how it's calculated. But realizing it will change is one thing, realizing just how much it can change is another. The latter is where a lot of people get (quite understandably) confused.

jmal said:
Uh, that's because it came up as a side comment in a post about Stalker's new tabbed UI, not as part of an essay about algorithms.

jmal said:
I think you're overreacting to criticism of Stalker, and not for the first time.

jmal said:
Don't take it personally - and maybe you aren't but that's how it comes across There are certainly unfair criticisms I've seen made of Stalker, but in my opinion complaints about the final prediction number are for the most part definitely justified.

jmal said:
Far from being a "clear and concise summary", it's too simple without reams of context you just have to cross your fingers and hope people actually read, making it very very tricky to use.

jmal said:
symbv, I am rolling my eyes really hard right now. Assuming my statement was intended to personally disparage you rather than it being an opinion I've held about an overly simplistic statistic on a website since before I'd even met you is the definition of taking it way too personally. You are not Stalker.

jmal said:
That aside, now that the argument is being turned into "If you dare to criticize this single specific data point you might as well throw all sales discussion out the window", we've clearly exhausted whatever constructive debate there was to have.

jmal said:
I'm deliberately ignoring your metaphor because debating metaphors always becomes about the metaphor itself rather than the original issue and the disagreement isn't complicated enough to need abstracting out to start with.

jmal said:
Of course it all requires explanation, there's no way you can seriously think I have ever said or felt otherwise. But the final estimate requires the most because it assumes the most and is most susceptible to fluctuations and statistical overreactions.

jmal said:
So no, I don't like it, and think Stalker should 1. improve it, 2. offer alternatives calculated different ways, or 3. not calculate it more than a month out. That said I still tried to take the middle ground and just recommend some more numerical context but apparently that makes me biased against all of Stalker, okay then.

jmal said:
if you're going to give prominence to the final estimate, other people should be able to say - without you assuming they're attacking you personally - that they think there's serious problems with it and that it's not the best way (easiest, but not best) to represent what you're intending trying to show.

jmal said:
No, symbv. Life is not a constant state of total war where compromise is tantamount to capitulating to an enemy. I have no intention of engaging in some intellectual running firefight with you every week on this, nor should you interpret it as such if I voice concerns about the statistic in the future. I wish you wouldn't interpret everything as being antagonistic all the time, especially from people like me who generally enjoy discussing things with you and honestly aren't out to get you just because there's a disagreement over something stupid like this.

scytheavatar said:
Am I right to say that things are looking extremely grim for Railgun2 and Oreimo2? Both are series you'll expect to hit 20k or at least 15K with ease, and the numbers suggest that the otakus are getting sick and tired of the saturation of light novel adaptions.

symbv said:
On the other hand, the real action for Railgun is only just beginning. And the guess is that the Stalker estimate is likely to be an underestimation, so a 15k sales is entirely possible. And if the coming episodes indeed prove very popular, its sales forecast will also benefit. Its OP CD has taken No.1 in Oricon Singles CD Weekly Sales Ranking with >26k sales, just another indication of the popularity of the series.

jmal said:

Estimates would very very very rough at this point. An estimate for, say, Shingeki could easily end up off by 20,000 units either way. Sequels can be guesstimated based on previous seasons, at best.

But to completely pull some estimates out of my ass, not even looking at Stalker's estimates:

20k+
Shingeki no Kyojin

15k-20k
To Aru Kagaku no Railgun S
Uta no Prince-sama Maji Love 2000%
Ore no Imouto ga Konna ni Kawaii Wake ga nai.

10k-15k
Suisei no Gargantia

5k-10k
Kakumeiki Valvrave
Hataraku Maou-sama!

0-5k
Everything else

Guess for averages, not just Vol. 1.

There are some other potential 5k sellers like Date A Live or HenNeko or Nyaruko W or Azazel-san Z.

Railgun and OreImo I have a feeling will drop below 20k. At least OreImo anyway. UtaPri should clear 20k easily for v1 with the event ticket but drop lower for 2-5 like season one.

Shingeki could be 15k, could be 50k, fuck if anyone knows at this point.

I'm probably not giving Gargantia enough credit but it's really hard to know how much of its rank comes from the act that its MSRP is higher (which results in higher ranking). Hell it could be 20k for all we know. Only 3 boxes and one format, so it shouldn't see a lot of release attrition.

There is lots and lots of time for this to change completely. Therefore my guesses are meaningless!

(And in my heart, Yuyushiki will always be 1,000,000+.)

jasque said:
How would Henneko do before the korean stuff?

jmal said:
Henneko shows that bigotry is powerless when compared to Kantoku bedsheets.

jmal said:
Henneko shows that bigotry is powerless when compared to Kantoku bedsheets.

jmal said:
Henneko shows that bigotry is powerless when compared to Kantoku bedsheets.