Talk:Mr. Klaw "Skill" Crane Game
Contents
- 1 Likelihood of success
- 2 Usage Limit
- 3 Trophy Possibility?
- 4 Page Title
- 5 Do stats (or sobriety) affect whether you get anything?
- 6 Spade Necklace
- 7 kol holidays
- 8 probability calculations
- 9 Used after ascending?
- 10 curious
- 11 Reference
- 12 Only two attempts per rollover now?
- 13 Do holidays affect drops?
- 14 End Game
Likelihood of success
Anyone want to help figure out what the percentage is for actually getting an item out of there? I am sure someone with better chart making skills could set this up a lot nicer, until then... 3 tries yesterday, 1 item (stuffed chesire bitten) --J12601 11:28, 19 October 2006 (CDT)
- OK, here's our data collection table. When adding data, ADD your totals to the already existing totals. Be sure to update both Failures (if any) and the Total at the bottom. Please just update the table and don't leave "Here's what I got" comments - we don't need the clutter. --Gymnosophist 11:49, 19 October 2006 (CDT)
PLEASE ADD ALL THREE OF YOUR RESULTS, INCLUDING TOTAL!!!! (Data is inaccurate otherwise)
- Failures + All Items should always equal Total, else you screwed something up
- Total should always be divisible by 3, else you screwed something up
- If you can not add and/or are a moron, please do not add to this table.
- After adding your numbers in, double check to make sure the total is correct and divisible by 3
- HowTo: Find a number is divisible by 3 - add each digit of the number. If the sum is divisible by 3, the number is divisible by 3 (Example: 2427. 2+4+2+7=15. 15/3=5. Therefore 2427 is divisible by 3). -OR- , use a calculator.
Outcome | Total | Percent |
---|---|---|
Failures | 3587 | 66.50 |
club necklace | 21 | 0.40 |
diamond necklace | 19 | 0.34 |
heart necklace | 18 | 0.32 |
spade necklace | 22 | 0.40 |
pack of KWE trading card | 25 | 0.46 |
rubber WWBD? bracelet | 27 | 0.48 |
rubber WWJD? bracelet | 44 | 0.80 |
rubber WWSPD? bracelet | 28 | 0.53 |
rubber WWtNSD? bracelet | 24 | 0.44 |
stuffed angry cow | 95 | 1.75 |
stuffed astral badger | 107 | 2.00 |
stuffed baby gravy fairy | 101 | 1.92 |
stuffed Cheshire bitten | 92 | 1.68 |
stuffed cocoabo | 95 | 1.77 |
stuffed flaming gravy fairy | 104 | 1.90 |
stuffed frozen gravy fairy | 87 | 1.64 |
stuffed hand turkey | 97 | 1.83 |
stuffed MagiMechTech MicroMechaMech | 108 | 1.98 |
stuffed mind flayer | 106 | 1.94 |
stuffed scary death orb | 107 | 2.00 |
stuffed sleazy gravy fairy | 120 | 2.25 |
stuffed snowy owl | 89 | 1.69 |
stuffed spooky gravy fairy | 85 | 1.60 |
stuffed stinky gravy fairy | 83 | 1.52 |
stuffed undead elbow macaroni | 97 | 1.85 |
Total | 5385 | 100 |
Further likelihood discussion
I've been wondering about possible explanations for the drops on this thing lately, and I'm thinking that being in HC makes the drops more likely. I went through an epic dry spell during a casual run to level 30 recently, with more and more failure as I went. By the end of the run, I was going 0/3 pretty consistently (I foolishly didn't collect any data about this). But now that I'm back in HC, I've been getting much better luck, almost always getting at least one drop per day (2/3 today, and I'm going to start recording this more now). Any higher-level folks seeing the same thing? Ephesos 05:58, 4 October 2008 (UTC)
I think that certain items will only drop for certain classes/signs. When I ascended for the first time, I started seeing three or four new drops that I hadn't gotten in several months of previous Mr. Klaw-ing. Unfortunately, I didn't have the presence of mind to keep track of what was dropped. --IAmATurnip 16:29, 6 December 2007 (CST)
3/3, 0,3 next day, Dec 13 2006 - sepharoth
It seems to almost always work 1/3 the time. I have had a 2/3 once. Has anyone ever struck out 0/3? or 3/3'd? --JRSiebz (☎|§|‡) 22:01, 19 October 2006 (CDT)
- Shadowhntr7 seems to have garnered an oh-for-six (really, two 0/3s). --Gymnosophist 22:12, 19 October 2006 (CDT)
- I just got a 3/3 on my multi (4/6 for the day). 2 owls and a macaroni. Huzzah. --J12601 23:13, 19 October 2006 (CDT)
- i've failed 9 times in a row :/--Bigbenwd 08:41, 20 October 2006 (CDT)
- 0/3 thrice, 1/3 and 3/3 once. --Hurax 09:23, 20 October 2006 (CDT)
- it appears to me as if the bracelets and the necklaces are more rare than the plushies.--Bigbenwd 16:16, 20 October 2006 (CDT)
- I got a 0/3, a 1/3 (stuffed baby gravy fairy) and a 2/3 (Plush Angry Cow and Club Necklace.) --Finners101 07:03, 22 October 2006 (CDT)
- Does having a certain familiar with you increase your chances of getting their stuffed counterpart? --Flare 19:43, 22 October 2006 (CDT)
- I'm on my first run, and I'm quite litteraly 0/24 --KAmeRe 18:01, 6 December 2006 (CST)
- I lucked out and got 3/3 my first day (two spade necklaces and a stuffed flaming gravy fairy), but the next day I went 0/3. I've gone 1/3 everyday since then. --Fionafairhame 04:37, 14 March 2007 (CDT)
- I quite frequently get 0/3. I have horrible luck - you can always count on me to get the worst results! --KiraCatgirl 12:23, 7 September 2007 (CDT)
Is it possible that you always fail when attempting in hardcore/ronin?--Starspiritgeno 05:01, 23 October 2006 (CDT)
- No. My multi is in ronin, and just went 2/3. --J12601 06:14, 23 October 2006 (CDT)
- And my main (and namesake) is in Hardcore, and just got 1/3 - stuffed baby gravy fairy, yayy!--Bork 09:01, 23 October 2006 (CDT)
- I've had a couple of Hardcore 2/3. Finners101 04:15, 24 October 2006 (CDT)
- I got a 3/3 in Hardcore.--Geek22 17:05, 24 October 2006 (CDT)
- I got 2/6 (1/3 on both of my guys). No hardcore or anything, since everyone else is posting that too. --Ηúηţêř ťĥẽ Ŕéļáţīv€ļ¥ ĜΘΘÐ (Ţâĺķ | Ćņťβ§) 20:20, 24 October 2006 (CDT)
- So far in softcore, I'm 10 for 24. 1 Cheshire Bitten, 2 astral badgers, 2 MagiMechTech MicroMechaMech, 1 scary death orb, 2 sleazy gravy fairy, 1 stuffed snowy owl, and 1 stuffed stinky gravy fairy. Azp2k32 21:23, 24 October 2006 (PDT)
- I got a 1/3 tonight in Hardcore. Spade Necklace added to my collection. Only need the Club Necklace now.--RicFule 23:56, 24 October 2006 (CDT)
- 0/9 :( Discworldian 16:42, 28 October 2006 (CDT)
- 0/12 Stupid RNG. --Gnomesquid 11:28, 7 November 2006 (CST)
- I'm a level 8 first run Disco Bandit. As my stats have increased I've noticed general increases in pickups from Klaw, but I still sometimes get 0/3 or 1/3... No noticeable increase in the quality of items when I get less, despite the increasing stats.
- I have gotten 0/3 multiple times, and mostly I get 0/3 or just 1/3 (NOTE: I've only added my numbers to the table starting today, but I found out about the Mr. Klaw at least a week ago). The first time I used it, I got 3/3, however. Granted, I've not done much adventuring, but I am a lvl 12 Accordion Thief (I've beaten the NS but I didn't break the prism yet). If the "stats relate to n/3" is a true statement, with higher stats yielding n/3 approaching 1, I either have rotten luck, or else the true relationship is between n/3 and Δstats, not the actual value of the stats themselves.... --Rhiannon, posting under shylarah 08:36, 12 June 2007 (CDT)
Currently the score is 518/732=.70765. Note that: 1/Sqrt(2)=0.7071. That number appeals to me more than 70%, but there's probably not enough data to get it within even 1% yet. (Although I can't be bothered to actually run the statistics. ^_^) --Starwed 07:49, 25 October 2006 (CDT)
- Well, now the success is 944 / 2880 = 0.328. I think 1 / 3 chance of success is a good guess. Vovan 16:31, 26 December 2006 (CST)
Could it have anything to do with when during the day you use the machine, or how much has been taken from the machine already for the day, possibly compared with the number of people with access to the machine? I just went 0/9 on three characters, and I'm sure that most people have already had their chance to use the machine before me.--Lostcalpolydude 06:21, 27 October 2006 (CDT)
- I highly doubt this. Personally I have gone 3/3 later in the day, or 0/3 right after rollover. I am currently testing daily on four accounts, and the time I am able to get on there varies from day to day, but my success seems to hold steady at about 1/3 between them all. Given the ability to ascend and use the game again during the same day, I have also gotten various amounts of things later in the day than I did the first time, and vice versa. --J12601 20:18, 28 October 2006 (CDT)
I hit the jack pot!! I got a pack of KWE trading cards followed by nothing followed by another pack of cards!! What can I say. The RNG must just like me!! --Tornan Steel 07:27, 4 November 2006 (CST)
- 4/39 in HC. perhaps it's less likely to get things in HC?--Kiltan 22:46, 7 November 2006 (CST)
Not true I'm in hardcore and my success rate is around 2/3 each time btw when I got one bracelet (WWJD?) I got the rest conscecutively-anyone else got that? --Syandoggy 12:32, 8 November 2006 (CST) --
I am trying to equip item drop increasers when playing Mr. Klaw, not sure if it is helping any though. Seem to get at least one prize per day since I started making sure to equip them. Might be something to think about for those who have any type of item increasing skills. --Fettucini carabinieri 22:59, 17 November 2006 (CST) Today went 0 for 0 with the same item drop increasers equipped, so maybe it doesn't help much, will keep trying. --Fettucini carabinieri 22:36, 19 November 2006 (CST)
Heh. I got 3/3 with a WWSPD, a cocoabo, and a spooky gravy fairy.--Disgaea Dude 8:54, 23 November 2006 (CST)
Has anybody noticed anything related to the moon phases? I can't say for sure, but it seems like I have gotten less when Grimace was full and gotten more items when Grimace was new.--Fettucini carabinieri 23:35, 10 December 2006 (CST)
Is it just me/my math/my calculator, or is that total wrong?--RPGMarker35 16:17, 24 February 2007 (CST)
Woo! a heart necklace! --Deranged Cat 11:21, 9 March 2007 (CST)
Is it possible that the number of days you use it affects your chance of success on later days? It seems like when I try using it day after day after day I only get 1/3 at best. But on this run I waited till I was out of Ronin (took me 9 days on oxy), and on my first try, I got 3/3. Also, the first time in an ascension or when creating a new multi usually gives me the best results... on my latest multi, the first try gave me a bracelet, a necklace, and a stuffed thingee. Coincidence? Luck? I really want to get more pack of trading card... :| --Jameslovelock 11:57, 13 March 2007 (CDT)
Wow.... 2 accounts, and for 5 days I got 3/3 everyday of the same item (rubber WWJD? bracelet) has anyone ever gotten this before? 20 of a single item? ever?
Grethko 11:57, 27 March 2007 (CDT)
- 2 accounts, and for both accounts 5 days of 3 (out of 3) successful pulls, all of which were the same....Where's the "20 of a single item" coming from? Maybe it's a typo, but I'd think that 2x5x3=30. Unless perhaps it was 2/3 of the WWJD? bracelet and 1/3 of something else? That would make for a total of 30 successful pulls, but only 20 of the given item... --shylarah 10:01, 18 June 2007 (CDT)
Okay, someone's screwed the table up bad. The corrected total number that Call me ishmail edited to fix adds up just fine. It's just not divisible by three. Suggestions as to what to do about it? --Jensachan 15:53, 11 April 2007 (CDT)
I believe it might not just be one factor but a combination. I got 3/3 with a diamond necklace, a WWtNSD bracelet and a pack of KWE trading card right before rollover. I am also in hardcore. It was also a Moxie day.
Jayvee-10:42 12 May 2007 (EST)
Usage Limit
Either the limit of three is for the entire clan, or there is protection built in that prevents newly-joined clan members (even white-listed people) from using the machine until after a rollover. The same applies to the snack machine. --sl1me 04:58, 18 October 2006 (CDT)
- ...but if you ascend, you can use it again ;-) --JRSiebz (☎|§|‡) 05:02, 18 October 2006 (CDT)
- I just checked, it's the latter (there's protection). Pretty cool. :) --sl1me 05:06, 18 October 2006 (CDT)
- The limit of three attempts seems personal, not for the whole clan. As of now, our clan has pulled five Klaw items between three people. --Dawnshadow 15:42, 18 October 2006 (CDT)
It's a person thing, each player can use it 3 times.jcowley(#1319374)
Trophy Possibility?
Anyone consider that the plushies from these machines result in a trophy?
- It might include the stuffed items from the gift shop as well. --Thoby123 10:54, 20 October 2006 (CDT)
- There doesn't appear to be. I have all the plushies (including the gift shop ones, the stuffed doppleshifter, the stuffed shoulder parrot, and the stuffed spooky mushroom). Then again, I still don't have all the other things. I only have two of the bracelets (Sneaky Pete and Jarlsberg), and 3 of the necklaces (I don't have the Heart). Natenife 14:30, 17 December 2006 (CST)
- The machine does give KWE Trading cards, which are very difficult to collect. Perhaps the cards may be worth a trophy. --The Pretender 7:11, 13 January 2007
- Maybe, though I kind of doubt it, since you can just buy the cards in the mall. Surely someone's tried purchasing the full collection by now? Natenife 04:47, 8 February 2007 (CST)
- according to jickenwings (and the inestimable epeterso2) 46 people have all the cards, no trophy. i know of at least one who has all the mr klaw items, and still no trophy. --Evilkolbot 06:10, 8 February 2007 (CST)
- Maybe, though I kind of doubt it, since you can just buy the cards in the mall. Surely someone's tried purchasing the full collection by now? Natenife 04:47, 8 February 2007 (CST)
Page Title
Why was this page renamed to include "A" at the beginning, which the image clearly does not include? --Quietust 11:51, 18 October 2006 (CDT)
Do stats (or sobriety) affect whether you get anything?
I have two characters in a clan, my main and my alt. My main was at the end of the first day of a new ascension (level 5 Disco Bandit, still in ronin) and drunk, my alt (level 11 Turtle Tamer) was completely sober. My alt pulled two items, but my main failed all three times. Is the Klaw -really- a game of skill, or does the RNG just hate me? --Dawnshadow 15:40, 18 October 2006 (CDT)
- I got on my multi, which has the same level, and same drunkeness. My main pulled two items, while my multi pulled none. RNG just hates us. --Mareusio 17:31, 18 October 2006 (CDT)
- I've observed something, although most likely it is a quirk. But my rate seems to improve with my moxie; I put it up to 193 today and got three out of three. It seems to be a trend, but I wonder if that is coincidental. I don't think sobriety has anything to do with it, as I was falling-down drunk at the time. Anyone else? --Berith 09:47, 8 January 2007 (CST)
- I just tried the Mr. Klaw with 1076 buffed moxie (828 unbuffed) and 14 drunk as a level 29 AT, and went 0/3. -ThaSupafly 20:20, 28 September 2008 (UTC)
I have almost 300 moxie but still usually only get one or two.
Spade Necklace
- If you notice, each necklace relates to the 'type' of player (club adds power, diamond is for collecting, heart is for friends... Why do I have a feeling that we're going to have to find out how to get the spade item? --Barstool 18:31, 18 October 2006 (CDT)
- Seems likely. The necklaces are likely a reference to the well-referenced article by Robert A. Bartle: HEARTS, CLUBS, DIAMONDS, SPADES: PLAYERS WHO SUIT MUDS. From the article: "...achievers are Diamonds (they're always seeking treasure); explorers are Spades (they dig around for information); socialisers are Hearts (they empathise with other players); killers are Clubs (they hit people with them)". This could be added as a reference to the necklace pages. --xcorvis 22:07, 22 October 2006 (CDT)
- Oh duh, it's already referenced. Kind of. --xcorvis 16:17, 23 October 2006 (CDT)
- Nope. I got a spade necklace without anything special, except maybe great RNG...I got both a spade necklace and a trading card in one round. --Fryguy9 09:40, 24 October 2006 (CDT)
- Could Necklaces have to do with how one plays the game? My Heart Necklaces drops are far above average. I (used to) spend a lot of time in /games until it became a /beggers. And I have always been very active in sending kmails to my clan. Maybe players who speed ascend (like Bashy) get Club Necklaces to drop more often. It also could be porportional to ascention. And the "increased odds" could reset after getting a Necklace. --typelogin 15:15 24 February 2007 (CST)
- I don't think so - I just got a club neclace today and I haven't ascended yet. Perhaps it is slightly altered by your class, or whether you can cast buffs on other players? --Froggybuster
- Seems likely. The necklaces are likely a reference to the well-referenced article by Robert A. Bartle: HEARTS, CLUBS, DIAMONDS, SPADES: PLAYERS WHO SUIT MUDS. From the article: "...achievers are Diamonds (they're always seeking treasure); explorers are Spades (they dig around for information); socialisers are Hearts (they empathise with other players); killers are Clubs (they hit people with them)". This could be added as a reference to the necklace pages. --xcorvis 22:07, 22 October 2006 (CDT)
- Or Maybe It's Just RNG Luck.
It is human nature to see intention in clustering, they are probably coincedence that you get so many--El taco 00:12, 19 September 2007 (CDT)
kol holidays
Has anyone gotten any special prizes from the crane game for St. Sneaky Pete's Day? Maybe while in St. Sneaky Pete's Day Stupor?
- There aren't any. It is sort of funny that I got a WWSPD? bracelet today, but that has nothing to do with the day. Oh, and sign your posts like this: --Ηúηţêř ťĥẽ Ŕéļáţīv€ļ¥ ĜΘΘÐ (Ţâĺķ | Ćņťβ§) 18:03, 25 October 2006 (CDT)
probability calculations
so I calculated a few probabilities off of the chart we have so far and I get the probability of getting
- Nothing = 66.3% ± .8%
- Any stuffed item = 29.3% ± .8%
- Any rare item = 4.3% ± .4% Though this has made a significant jump in the last 1000 data points. I think people may only be adding data if they get a rare (or at least biased in that direction).
individual item frequency varied a lot, but averaged 1.8% ± .2% for a specific animal, and .5% ± .1% for the rares.
--Univector llama 4 May 2007
Used after ascending?
In the first section it states "It may be used three times per day and like other once-a-days, it may be used again after ascending."
Shouldn't this actually be "after rollover"? You do not need to ascend to use the machine again. Or am I missing something? mugsiam 18:00, 26 November 2006 (CST)
- No, it means that you can use it, ascend, and use it again. --Ricket 18:05, 26 November 2006 (CST)
- in other words, you can use it 3 times per rolover, but if you use it three times, then acend you can use it an additonal 3 times b4 rolover--CSDragon 20:53, 11 April 2007 (CDT)
curious
My multi was able to equip a stuffed cocoabo. I have Knob Goblin Tongs equipped as well, but whenever I changed weapons I couldn't equip one of my off hand items... anyone? --Dcfiregirl 11:13, 1 December 2006 (CST)
- Were any of the other weapons one-handed? --Ion42 12:58, 1 December 2006 (CST)
Reference
Considering the mechanical/robotic claw, is there a chance that the name might also be a reference to "Klaw" - Inspector Gadget's nemesis? --Grimdel 11:15, 25 January 2007 (PST)
- It's possible, I suppose, but where I come from these games are "klaw" games, with the "k" instead of a "c". The only exception I can think of is the popular kiddie-version, the Candy Crane ("play 'til you win", it says, and if you get nothing you do get to try again -- but a single tiny-size Tootsie roll counts as a "win"). And that...doesn't have "claw" (or "klaw") in the name....
I thought that "Mr. Klaw" was "klaw game" + "Mr." (like Mr. Store, Mr. Accessory, and so on, with the cascade of references applicable to something referencing said "Mr. (Stuff)"). --shylarah 11:12, 11 June 2007 (CDT)
Only two attempts per rollover now?
Over the last couple of days I've only been able to use the Mr. Klaw twice per day before it broke down. Anyone else notice this? --Eitheladar 06:37, 16 February 2007 (CST)
- I don't think this is the case. Are you perhaps using Firefox with a Greasemonkey script which is automatically refreshing the game three times in one perhaps? Or some other script? Or perhaps forgetting that you already clicked (unlikely)? Or something else? Or perhaps it really is variable? Well, all I know is today I got three attempts for sure. --TheDotGamer 08:08, 16 February 2007 (CST)
Probably some script's fault indeed. Sorry 'bout that. --Eitheladar 07:04, 17 February 2007 (CST)
Do holidays affect drops?
It seems as if holidays might affect the drops. On El Dia De los Muertos Borrachos I got 0/3, the same with Generic Summer Holiday. On today, St. Sneaky Pete's Day, I got 1/3 - a rubber WWSPD? bracelet. Coincidence? -Fadookie 04:08, 17 March 2008 (CDT)
End Game
Can we please end this page? With several thousand samples, the odds have been determined. You have a 1/3 chance of getting an item. Items are divided into two catagories: rare and common. Rare items are the necklaces, bracelets and KWE cards. Common items are everything else. Common items are each 4 times as common as rare items. -QuantumNightmare 18:32, 14 November 2008 (UTC)
- I think the last statement is the only one there technically isn't enough data to support. Kind of hard to pin down something which appears to have a rate of about 1/80 amongst the successes. We need a much larger sample size for that; an order of magnitude larger. Unless there's a way to use the correlations to cut down on the error margins. In any case, as far as not being insane about things goes, then yes, I'd say this is the conclusion we should draw. --Flargen 22:52, 14 November 2008 (UTC)
- There's plenty data. It's clear that the summons are divided into two catagories, with every rare summon having the same odds and every common summon having the same odds. Once that conclusion is reached, we can find the odds of each easily. Out of 1625 successes recorded here, 205/1625 are rare summons. Note that there are 9 rare summons and 16 common summons, therefore:
- Average rare rate is (205/9)/1625 = 1.40%
- Average common rate is (1625/16)/1625 = 5.46%
- Comparing the two we find that the common rate is 3.89 times the rare rate. Thus the fact that commons are 4 times as likely as rares is supported by 1625 samples, which is more than enough.
- Remember kids, it's not just the size of the boat, it's how you use it.-QuantumNightmare 00:09, 15 November 2008 (UTC)
- What I meant is that there's still a lot of error margin on the projected 1.4% chance of any particular rare. To get 95% confidence for an anticipated average of .014 with an error of margin of .001 takes over 53000 trials. Right now we've got a 95% error margin of .0033. Which is a pretty substantial portion of .014. So is .001, actually, but it's at least at about the scale I personally stop giving a damn. It's needing the 53000 trials to cut down the error margin by a third is what makes me feel like we should just call it like you say. As I was trying to say, it's the correlations between the probabilities of the items that I don't know how to take proper account of. An "average" probability for the category doesn't establish that that is, in fact, the probability for each item. I think it works out like you say in the end, but I never cared enough about statistics and probability to know the formalisms for that case. I might be able to just infer them, but I continue to not care enough. --Flargen 00:30, 15 November 2008 (UTC)
- I suggest that key rate is just the proportion of rares, which is 205/1625 = 12.61%, or about one-eighth. That would give a 1/72 chance for each rare, and 7/128 for each common, so each common is 3 15/16 times as frequent as each rare. (The 95% confidence margin of error, based on 1625 trials and an anticipated average of .125 is about .0161.) But again, I agree that trying to nail down the exact figure is far too much effort for too little gain. —Yendor 05:18, 15 November 2008 (UTC)
- There is an error of margin in the success rate, as well, (of about .013 at 95%; which is pretty decent for a 1/3 estimated probability) so that's not the true error of margin. That's just the conditional probability and error of margin amongst the successes. The true error of margin is larger because the conditioning probability is also uncertain. And, again, simply averaging "all rares" in no way establishes that all rares occur with probability of that average. You're just assuming that whoever coded the probabilities went for a lot symmetries, and there's plenty of evidence both for and against this (so it's a questionable assumption at best). Unless you happen to have a radio transcript or something where they say that the probabilities are perfectly correlated. Or you can explain to me how you use strongly correlated outcomes to reduce the number of trials needed for a certain level of precision, and then do so. Not like I care all that much beyond the academics and technicalities. I wouldn't mind QN's numbers being stuck in the actual wiki page at all. --Flargen 05:36, 15 November 2008 (UTC)
- If the probability of each item in a catagory is exactly equal, we can look at the catagories in general. While it would be incredibly hard to prove a 1.4% drop, it's much easier to prove a 12.6% rare drop rate and we get the exact same result from the two. If we run with the 95% confidence interval on the 12.6% rare drop rate, we find that commons are 3.38 to 4.56 times more common than rares. It's much more likely that 4 is the correct value, instead of either 3.5 or 4.5 which are near the edges of our confidence interval. Meanwhile if we choose 4 times as being the most likely result, we receive an expected 12.3% rare drop rate (which compares well to the 12.6% drop rate we encountered in 1625 samples). We can very reasonably declare that this is the correct result.
- While this is only true if all items in each catagory have equal odds, this fact can easily be seen from the data. This isn't an assumption: the data very accurately matches two constant values, one for each category. While it is possible that this is incorrect, there isn't anything in the data to assume otherwise.-QuantumNightmare 09:37, 15 November 2008 (UTC)
- There is an error of margin in the success rate, as well, (of about .013 at 95%; which is pretty decent for a 1/3 estimated probability) so that's not the true error of margin. That's just the conditional probability and error of margin amongst the successes. The true error of margin is larger because the conditioning probability is also uncertain. And, again, simply averaging "all rares" in no way establishes that all rares occur with probability of that average. You're just assuming that whoever coded the probabilities went for a lot symmetries, and there's plenty of evidence both for and against this (so it's a questionable assumption at best). Unless you happen to have a radio transcript or something where they say that the probabilities are perfectly correlated. Or you can explain to me how you use strongly correlated outcomes to reduce the number of trials needed for a certain level of precision, and then do so. Not like I care all that much beyond the academics and technicalities. I wouldn't mind QN's numbers being stuck in the actual wiki page at all. --Flargen 05:36, 15 November 2008 (UTC)
- Remember kids, it's not just the size of the boat, it's how you use it.-QuantumNightmare 00:09, 15 November 2008 (UTC)
- Almost a year (and 175 results) later, I think we can call this one done. I tallied up 1800 item drops, 227 rares (12.611%) and 1573 commons (87.388%). Knowing Jick's love of numbers that make sense, I'd say 12.5% and 87.5% would be a safe estimate. Factoring in the 2/3 chance of failure, this would give an overall 29.16% chance of a common and 4.16% chance of a rare. I'll put this on the main article some time later today. Thoughts? Comments? --TechSmurf 05:59, 29 October 2009 (UTC)