CS Rarity List [WIP] - status update on page 75

[ThunderCedar]

may I ask if you guys are still looking for someone with statistic knowledge? ^ - ^
I'm not doing a stat major or anything but 3 of the years I did in highschool had stats and probability as main subjects. So anything from frequency to percentage and etc, I can do ;u;

[Solloby]

Yes it would really help confirm if our calculations are correct once they are released, but also if you can help us work out if there's a valid way to compare pet rarities across litters.

I've gone through the July data and compiled the summary into a separate spreadsheet. I hope to make a new tab for each month (July / Aug / Sept). Here's what I have so far; is this useful?
http://sta.sh/0h20rk8m231

[Law]

So much data I'm going cross-eyed @.@ but also SPARKLY EYES *O*

Did you grab December 24, 2011?

I totally forgot OTL

Have the data now though! (keeping all the links together so they are easy to locate) :D

Spreadsheet links (with sampled ID range)
2008: 810000 to 820000
2009: 7270000 to 7310000
2010: 25432641 to 25631198
2011: 54673270 to 54831930
2011 (12-24): 55494774 to 55648442


Some notes:
- Looks like there was a slightly higher chance of adopting URs during 12-24 (but there were so many outcomes that this might not be entirely accurate)
- During both rereleases, no cake dogs were adopted. Only the palomino horse was released. I'm pretty sure I've seen cake dog rereleases (not necessarily from 2011), so this is a bit strange

[EmberzTale]

Is any help currently needed for counting data ranges?

[Solloby]

The pet counts are finished, we are just looking at the data now.
I'm not sure what to make of the rerelease data tbh.
The July data is posted on the previous page.

[EmberzTale]

All the pet counts are finished? As in we're not doing any more?

[Law]

All the pet counts are finished? As in we're not doing any more?

Well, we may need to count more in the future, but for now they are finished.

~~

I'm not entirely sure how to take rerelease into account either... for one it's hard to tell if 10-20k is a big enough sample size for the number of possible outcomes. :P

If the sample is accurate, then would the analysis look like this?
For the 2011-12-18 rerelease, I took 28,621 samples out of a range of 158,660 IDs. Lets compare the sunjewel to the new rereleases (UR cucumber and UR lion).

Sunjewel: 11/28621 = ~61/158660
UR Cucumber: 46/28621 = ~255/158660
UR Lion: 44/28621 = ~244/158660

There were ~120 sunjewels released in July 2008, but we also have to include the sunjewels that were rereleased during 2008, 2009, and 2010:
2008: 76
2009 (~40000 total): 32/11818 = ~108/40000
2010 (198558 total): 15/20785 = ~143/198558

Does that mean there are 508 sunjewels total on the site as of 2011-12-18? And we haven't taken off-site adopts or inactive users into account. That could even mean there are only 100-200 sunjewels available (or on the flip side, 400-500).

Overall, it's pretty hard to conclude how rare URs are relative to the sunjewel.

[Solloby]

I'm pretty sure UR pets are rereleased in different amounts on different years as well - look at UR Cat going VR one year randomly. Perhaps we should ignore the rerelease data and just look at the original release data for July/Aug/Sept?

[nickjr]

Thinking quickly, I'm not sure what to make of the data either OTL Will try to think more when I have more time; classes are gonna start in a few days and I haven't even gotten back to campus yet lol

Edit: actually I'm probably gonna be MIA or super sparse for at least a week...

[SabrinaB]

I think (juat my theory) that during the rerelease 0,02% is sunjewel, 0,005% tiger dog etc.
This %is based on the amount of members, so every year more pets will be rereleased since there are new members every year (and some leave, but lets not make it too complicated).
However, the % staya the same, so they dint mess up the rarity / value.

Sidenote: i picked some random numbers as example, these should not be taken seriously in any way.