Talk:Lunar Fortune/Drop rate

The previous years shouldn't be combined. The drop rate for the rat was stated to be higher than that of the pig. [1] 66.190.15.232 20:11, 5 January 2009 (UTC)

While I don't think there is a need to combine them, the info linked doesn't say what you think it does: the extra pigs came from other sources. (district giveouts and so) Backsword 20:21, 5 January 2009 (UTC)

Ah, I didn't know that. We could split them in two rows if you want to. It's more for display and temporary quick reference. Sorry about moving around your comment - i was confused and didn't see this was the Talk page XD

@Backsword: Gaile said "drop rate" though.. Do you mean that the percentages were the same but that there were more methods of getting one? -- Karasu (talk) 20:23, 5 January 2009 (UTC)

I'll go through the statistics later and split them into two entries. poke | talk 20:26, 5 January 2009 (UTC)

Unreported Aura Counts

Sorry. I tried to keep track of each of the three blessing effects as I was opening all my fortunes, but I lost count early on and didn't keep screenshots of the full chat log. I'm sure of the other totals because my inventory was empty before opening all my fortunes and I took before and after screenshots. LicensedLuny 11:35, 31 January 2009 (UTC)

Unfortunately doing so would break the Statistics. Thus I didnt include them. --SilentStorm 20:44, 31 January 2009 (UTC)

Other Statistics for the Droprate

These are very useful, but what if what month the character was made in affects the drop rate for the item? Perhaps we should include that when someone enters data, such as another box which if filled out yes/no if the character you opened the package with is of that month? Moo

Old data

Well, I found that I have all my opened lunar fortunes drop written on the paper, what about adding archival data :P? The other problem is that I'm always too lazy to calculate it all up and then write it down here :P.

Adding this years totals

I didn't see any drop rates for the Tiger data, except for the claims made on the talk page. So, I have moved the Ox totals to the archive and have started a new table for the Tiger data. I noticed a lot of miscalculations in the archival totals, but did not change them. However, I did correct them when totaling up the previous years totals for the main page. I also did not include any totals that were incomplete. If someone wishes to edit the pages, to make them look better or something, be my guest. Hopefully everyone is ok with me doing this. FloppyJoe 00:02, 14 February 2010 (UTC)

Rabbit data problems - plan to move them to separate table

I propose to create a second reporting table for this year and move incomplete data sets (and text-only reports on talk) to that table. (Unless, of course, there are objections.)

A number of people have added rows to the 2011 table, but left off how often they received a bonus effect. This throws things off when determining percentages. Similarly, a number of people have posted to talk indicating how many fortunes they opened and how many bunnies they found (e.g. 0, 2, etc), without indicating what else they found.

Accordingly, I think we should create a new table just for bunny sightings and move the incomplete rows to that table and add the bunny-sighting data. Then, for that table, calculate just the likelihood of finding a rabbit drop. We can be reasonably sure that the bunny-only table is going to have more selection bias, but there's no particular reason to throw out the data entirely.

Since Stumme has confirmed a 0.45-0.65% drop rate, we can use both tables (full data sets and bunny-only) to get a sense of reporting bias...and potentially use that in the future to help make predictions for this and other items.  — Tennessee Ernie Ford (TEF) 16:40, 6 February 2011 (UTC)

The people who didn't count their blessings seem to have put in the total lunar tokens though. So you can still pull statistics for the bunnies and firework items since you know the total number of lunar tokens of those people. That way you can make a general prediction of the separate items and the general chance to have a blessing.--Fogeltje 18:33, 6 February 2011 (UTC)

I think it might be worth the trouble for bunnies, but I don't think it's worth the effort for everything else (data cleansing is tedious, time-consuming, and error-fraught; it's also controversial). We probably have enough reliable data for non-mini items and blessings; we are always short reliable data for mini drop rates.  — Tennessee Ernie Ford (TEF) 18:55, 6 February 2011 (UTC)

Of course it is questionable how reliable this is at all considering we only have a very few samples. Not to mention my bunny drops seem to be somewhat of an anomaly in the data. With a lot more data that wouldn't be so bad (or perhaps turn out not to be so high after all).--Fogeltje 19:37, 6 February 2011 (UTC)

yeah - I only know my fortunes and bunny data - 824 fortunes burned, 1 bunny dropped (and bunny dropped around 743 as I was clicking through the end of my third stack). In hindsight I should have bought the bunny by selling lunar tokens, as my luck stinks. --Falseprophet 17:58, 8 February 2011 (UTC)

Incomplete Statistics

Should the 5 rows with a lot of tickets, but no entries for the Lucky Aura, Lunar Blessing & Spiritual Posession by removed from the table & statistics caluclation? I'm asking, since their inclusion will clearly bias the total calculations... I'm discounting the row for 6 tickets, so few tickets seems plausible to have no occurence of the effects --BramStoker (talk, contribs) 13:11, 7 February 2011 (UTC)

See above. Unless anyone objects, I plan to create two tables: one that has full data on each row, one that has partial data (and will include the bunny sightings data from talk:Lunar Fortune. However, I'm waiting until next week for three reasons:

1. It's easier for contributors to have only one table to choose from.
2. Previous year information gives us a good idea of the drop rates.
3. People will probably still be opening fortunes for another week or so; there's no particular rush.

I think we should keep the incomplete rows because it might give us some insight into reporting bias.  — Tennessee Ernie Ford (TEF) 17:10, 7 February 2011 (UTC)