Talk:Lost Orrian Jewelry Box/Drop rate

From Guild Wars 2 Wiki
Jump to navigationJump to search

Wrong percentages[edit]

The percentages are wrong, because you get more than one item from the box. I think it's always 1-5? unidentifiable object in there + one item (1 loadstone/ 1 obsidian/ 1 pet/ 1 potion or multiple potions of karma). --5.146.47.20 02:52, 29 January 2013 (UTC)--

I could add a parameter to turn them off if you want. Chieftain AlexUser Chieftain Alex sig.png 11:45, 29 January 2013 (UTC)
It's the chance to receive any one item from the box, doesn't matter if there is multiple items--Relyk ~ talk > 18:52, 29 January 2013 (UTC)
Still, if you're receiving 4 item a time, % results will be screwy. -Chieftain AlexUser Chieftain Alex sig.png 19:10, 29 January 2013 (UTC)
meh--Relyk ~ talk > 19:16, 29 January 2013 (UTC)
I agree, the data is skewed. Entries should be by the number boxes, not by the total number of items received. i.e. Unidentified object should have 100%+. Data does not tell us chances of obtaining an item from a box, but rather it tells us how much of an item we will have according the total number of items we received. Jemi > 23:47, 29 January 2013 (UTC)
Heh, someone reverted my code addition to the footer, perhaps I could instead add a hover note to explain why the percentages may not reflect the real chances. Chieftain AlexUser Chieftain Alex sig.png 01:54, 30 January 2013 (UTC)
How about this? --> [?] <--- (hover to see the message. I'd stick this beside the 100%) Chieftain AlexUser Chieftain Alex sig.png 02:06, 30 January 2013 (UTC)
We need a field for number opened if it's needed, a generic count parameter in SDRL would work. Although you could derive the count from the numbers. Then people can do stuff with it even if our poor abused template can't.--Relyk ~ talk > 04:58, 31 January 2013 (UTC)
I wasn't sure exactly how you wanted the percentages to be handled, but User:Chieftain Alex/sandbox. -Chieftain AlexUser Chieftain Alex sig.png 12:45, 1 February 2013 (UTC)
One solution that would take care of both the original intention (+1 for each type for each opened box), and actual number of items received in the stack is adding a separate coloumn (or line) under each item. The downside is of course the doubling of entry fields and the already skewed values. My suggestion would most likely require that all current data is voided in order to ensure more accurate values, which would very likely not be received well by those who have spent lots of time entering their data... LarsV 00:46, 29 April 2013 (UTC)
The best solution is to stick everything in a bloody spreadsheet that was designed to perform data aggregation like this, not in some wildly complex wiki templates using simple functions that were never intended for this sort of thing. I've been saying this all along, but no one seems to agree with me. —Dr Ishmael User Dr ishmael Diablo the chicken.png 02:27, 29 April 2013 (UTC)
BUT WE NEED DROP TABLES ON THE WIKI ISHMAEL--Relyk ~ talk > 02:36, 29 April 2013 (UTC)
I've nothing against simple drop tables where we just track one vector of variables. But trying to track both the type and the quantity of items dropped on the wiki is insane. —Dr Ishmael User Dr ishmael Diablo the chicken.png 02:42, 29 April 2013 (UTC)
Like u said, percentages are wildly wrong. Why don't u just add the numbers manually? I can't, don't understand enough about this wiki-pages. I calculated percentages per box manually for 16316 boxes listed on this page (I dropped the ones that did list their things according to description and didn't list multiple items per box, I also excluded the boxes that didn't list karma items or unidentified objects but only for the calculation of the things that they did not mention). Here the numbers per box (and thus the only thing people will be interested in). More than 100% means per average box there are more than 1 that will drop of that item type.
Item | Number of Drops | Percentage chance of drop per box
Drop of Liquid Karma | 18653 | 126.71%
Taste of Liquid Karma | 4254 | 28.90%
Vial of Liquid Karma | 1651 | 11.22%
Swig of Liquid Karma | 1844 | 12.53%
Obsidian Shard | 2464 | 15.10%
Charged Lodestone | 54 | 0.33%
Corrupted Lodestone | 59 | 0.36%
Crystal Lodestone | 74 | 0.45%
Destroyer Lodestone | 53 | 0.32%
Glacial Lodestone | 70 | 0.43%
Molten Lodestone | 74 | 0.45%
Onyx Lodestone | 70 | 0.43%
Pile of Putrid Essence | 44 | 0.27%
Unidentifiable Object | 50954 | 314.74%
Charged Potion | 14 | 0.09%
Cold Potion | 9 | 0.06%
Hard Potion | 4 | 0.02%
Warm Potion | 10 | 0.06%
Mini Risen Priest | 171 | 1.05% Rink (talk) 23:46, 9 May 2013 (UTC)

230k Karma spent, and the final result is....[edit]

already used 50 boxes and the the things i got was 1 glacial lodestone, and 4 obsidian shards( which i could get for 9k karma), and 169 uni objects(wtf?), from that boxes got items thats returns me almost 70k karma, that was a loss of 150k karma. So far this item seems to be a karma hole nothing more, or something wrong with drop rate, or this is one more unusefull item in the game so we can spend karma without any profit .... kisses <3Eternity.png (Max / talk) 16:52, 29 January 2013 (UTC)

Anet introduces "lucky" items that have buggerall chance of a good item all the time. Coffer of Whispers -Chieftain AlexUser Chieftain Alex sig.png 19:19, 29 January 2013 (UTC)

72 boxes down...[edit]

I have no clue how to add to the work already done, so I will post my results here and hope that someone can add it, I'm sorry they aren't fully documented, I didnt realize there was drop research until after I had sold my unid items stored stuff etc etc. What I DO know is that I opened 72 of them and recieve 1 charged loadstone 1 crystal lodestone 1 onyx lodestone and 3, (yes 3) mini balth pets. I was shocked to see such a low % listed here when I got so many out of only 72 boxes. GL all!

Thats good news, but unfortunately without the other tiny bits of data we'll be unable to add your findings. Still, it may give hope! -Chieftain AlexUser Chieftain Alex sig.png 00:09, 30 January 2013 (UTC)

Inaccurate results due to estimations[edit]

The entry added by Mejais Souleater is clearly an estimation, meaning the data was likely constructed after the fact. This also means the data is useless. Should this entry be removed, or is an estimation "good enough" for the purposes of this page? --71.94.244.71 07:10, 2 February 2013 (UTC)

Pile of Putrid Essence[edit]

Should be included on the table, I got 5 out of the boxes I opened. -Cursed Angel 熱 18:52, 18 February 2013 (UTC)

Looks like an oversight on my part, apologies. I've now added the option. -Chieftain AlexUser Chieftain Alex sig.png 20:59, 18 February 2013 (UTC)

Early analysis[edit]

Unidentified objects
  • 61%
Karma items
  • Drop - 25%
  • Taste - 6%
  • Vial - 2%
  • Swig - 2.5%
Materials
  • Obsidian shard - 3%
  • Each type of lodestone - 0.1%
Tonics
  • Each type of tonic - 0.01-0.02%
Miniatures
  • Risen - 0.2%

evidently since you receive multiple unid objects at once + people have recorded them as <multiple> instead of +1, the data is skewed towards the karma + unids.. relative percentages of the remaining stuff might be ok. -Chieftain AlexUser Chieftain Alex sig.png 13:58, 20 February 2013 (UTC)

+ after applying some very rough estimations on how the multiple items have skewed the data:
  • 25% junk
  • 10% obsidian shard
  • 1% mini risen
  • 0.25% each type of lodestone
  • 0.05% each type of tonic
  • 33% Drop of karma, 17% taste of karma, 8% vial of karma, 4% swig of karma
-Chieftain AlexUser Chieftain Alex sig.png 14:30, 20 February 2013 (UTC)

Always drops 1 set of karma pots and 1 set of non-karma things[edit]

People have mentioned that the percentages are wrong because some things are dropped in piles, but they're also wrong because they conflate the karma potions with the other items. Every time you open a box, you get one set of karma potions, and one set of other items. They need to be recorded (or at least, percentaged) separately. 24.91.0.228 00:49, 14 March 2013 (UTC)

That would be very useful to note. -Chieftain AlexUser Chieftain Alex sig.png 01:08, 14 March 2013 (UTC)

Some additional drop data[edit]

I bought and opened some boxes today, and took the liberty of recording the drops and doing some basic stat work. The karma gained was with guild buff, banner buff, and BLC karma buff active, to make it a bit more worth dumping all of my karma into this. :P

The reason I didn't just add this to the table is that the percentages are skewed because items can be dropped in varying multiples, and I've a personal standard of not contributing data to inaccurate statistics. -Adilor 05:57, 21 April 2013 (UTC)

Your statistics are inaccurate for the same reason, you might want to reconsider your personal standards.--Relyk ~ talk > 15:58, 21 April 2013 (UTC)
Perhaps Relyk meant: "thanks for sharing this huge amount of data, even though this wiki doesn't have a way to include it in the drop rate tables."
@Relyk: care to explain why you think Adilor's stats are inaccurate?
I see two different and incompatible goals. The wiki only cares about *if* something drops, which is useful in predicting how many boxes you need to open before you see a particular drop. Adilor and most of the posts on Reddit or Guru care about *how many* of each item drops, which is useful in predicting karma -> gold conversion rates and other types of economic efficiency. Both methods have value and probably there's only room for one method on the wiki. 75.37.19.169 16:13, 21 April 2013 (UTC)
He didn't track how many times he received items that dropped multiple times, including Unidentified Object and Karma Drop, which is why he's getting over 100% The fact the item itself varies in the amount receives doesn't impact the statistics, only the average number you will receive. If you did track it correctly, you could find that average (Total/Number of times received).--Relyk ~ talk > 16:19, 21 April 2013 (UTC)
He's tracking total amounts, so the % will necessarily add up to more than 100%. The fact that the item varies in the amount received *does* impact the stats for computing expected *value* -- the 100% total system only covers the expected result of yes|no. The fact that the OP is posting totals > 100% has nothing to do with accuracy. Similarly, it was unfair for the OP to call the wiki's data "inaccurate" since it's collecting a different type of data.75.37.19.169 18:06, 21 April 2013 (UTC)
Truthfully, it wasn't unfair of the OP to call it inaccurate - it has been highlighted several times that different users are blatantly logging different things. Its hard to compare the results of Adilor's data and the wiki page since evidently Adilor's data has counted "receive 5 unids at once" as 5 units and not 1 (wiki should record 1 unit here). I tried to "fix" the data here. (also we shouldn't have put the karma items in the same table as the other stuff :/ ) -Chieftain AlexUser Chieftain Alex sig.png 18:30, 21 April 2013 (UTC)
The easy solution is we can have people list the number of trials as a parameter because the totals get skewed they way they are calculated (Guess what, it's calculated automatically!). If someone wants to design a template that does that and deal with the how terrible a wiki is for statistics and spreadsheets, they can do that on their own time. You could then ignore the statistics and manually figure out the total number of chests opened from the drop table then add that as a parameter, assuming people followed the rules. If you have a problem with inaccurate statistics (generated automatically), the calculate them yourself or propose a better way for the wiki to deal with it, the data is all there.
tl;dr just don't post information the wiki can't use and then bitch about what's wrong with the wiki, we can't help you there.--Relyk ~ talk > 19:13, 21 April 2013 (UTC)

Collapse User Entries[edit]

The user entries section is getting quite large. Does anyone object to collapsing it with a hide/show button, leaving it hidden by default? --Moto Saxon 03:28, 29 April 2013 (UTC)

If we do have it collapsible, it would be expanded by default. 95% of tables aren't this large.--Relyk ~ talk > 03:53, 29 April 2013 (UTC)
But it's only going to be getting larger. IIRC, gw1 wiki had this type of data on a separate page with only the percentages on the actually wiki page. I haven't been around much, but have we discussed doing that with all of the 'drop rate' data pages? Would that be an acceptable solution? --Moto Saxon 15:49, 29 April 2013 (UTC)
pft. Disregard my silly comments. For some reason I thought this was on the lost orrian box page and neglected to see this is a separate drop rate page! Not sure what I was thinking. --Moto Saxon 16:17, 29 April 2013 (UTC)

1400 boxes results, 30 MAY 2013[edit]

I didn't record stats on the karma vials, only the non-consumables.
4441 Unidentifiable Objects
216 Obsidian Shards
14 Risen Priest of Balthazars
12 Piles of Putrid Essence
8 Molten Lodestones
7 Onyx Lodestones
7 Crystal Lodestones
6 Charged Lodestones
5 Glacial Lodestones
3 Corrupted Lodestones
2 Destroyer Lodestones
2 Charged Potions
1 Hard Potion
1 Warm Potion
1 Cold Potion

Post cleanup[edit]

Ok ok I gave in and finally put user input totals in. I've taken the liberty of removing incomplete data (e.g. data without any karma items), which means that:

  • Average of 3 unid per box
  • 0.01 minis per box - about 1% - this ties in with some individual data provided above
  • On average, you'll receive 2335 karma points per Box - i.e. you should get 50% of the 4550 spent karma back.

-Chieftain AlexUser Chieftain Alex sig.png 15:45, 25 November 2013 (UTC)

as much as I love this set of drop rates, it takes 25sec to load after saving an edit. it might be worth combining all of the existing rows in terms of wiki-editing friendliness. any objections? -Chieftain AlexUser Chieftain Alex sig.png 23:58, 6 April 2014 (UTC)
I think we could archive it and start over, on the basis that the drop rates are likely inaccurate due to some people recording their total amount of items and some people recording the number of incidents as they were instructed. And in the new version, combine all lodestones into a single entry since they are all equally likely. Psycho Robot (talk) 00:02, 7 April 2014 (UTC)
Sure the way that it's used is different to most of the other drop rate tables, but when setting it up to be manual, basically every user had written their manual totals by their signature already. I don't think that after 25k entries that its statistically possible for users who have mistyped 1 or 2 totals to have ruined the data. As to combining lodestones, they'd average out at 98 if we combined all of the "... lodestones" with the "pile of putrid essence" - would you be including the putrid essence? seems much lower on first glance. -Chieftain AlexUser Chieftain Alex sig.png 07:51, 7 April 2014 (UTC)
I'm not talking about averaging them, I'm talking about recording the total amount of lodestones dropped, regardless of type. So instead of having "charged lodestone: 88; corrupted lodestone: 91; etc. etc., we'd have just a single entry - lodestone: 724. This also records the probability to receive a lodestone at 2% which is a much more useful number to users. In addition to making the data easier to digest it reduces the amount of fields required by half. And no the essence is obviously not the same drop rate so it'd still be different. Even if it were the same, I'd still separate it because it's not immediately obvious that its a "lodestone". Psycho Robot (talk) 18:25, 7 April 2014 (UTC)