View source for Talk:Statistics Notes

:*Yiab, it would be very helpful if you would include the raw data somewhere within the wiki, preferably within talk pages. If not, the information you gather must be dated so that when things change, we know whether given data is still valid. --[[User:Snickles|Snickles]] 06:58, 13 January 2006 (Central Standard Time)
::*Considering the raw data is currently over 70000 entries in my database (and I'm less than a third of the way through the project), I think that posting the raw data here would be too much space for the benefit. I do agree with posting the dates, though, so I'll put that information on this page, if nobody minds.--[[User:Yiab|Yiab]] 03:23, 21 January 2006 (Central Standard Time)
:::* A summary of the data would be nice to have (See [[Talk:Knott Yeti]]), but a record of the dates is certainly the most important item to record. With that said, I'd be interested in getting a copy of your data, if it is not a problem. ;) --[[User:Snickles|Snickles]] 14:02, 21 January 2006 (Central Standard Time)

While this is useful information, the level of precision is much too high. For example, the yeti fur droprate at the peak is listed as 30.8%. I haven't studied statistics for about 7 years, so I don't know the formulas to calculate the accuracy of data, but to be certain of that level of accuracy would probably require tens of thousands of adventures. The numbers for the drop rates of yeti and penguin fur are suspiciously close to 30% and 10% respectively... which I suspect is closer to the actual drop rates than 30.8% and 9.4%. I don't want to sound like an annoying maths teacher, but in times like this, less can be more. Again, the stats are really interesting and useful, but I think it would probably be better if it was rounded to at least the nearest 5% (at least in cases as clear as the peak.)--[[User:Blibbler|Blibbler]] 02:18, 22 April 2006 (CDT)
*While I agree with your reasoning in principle, it's something that I considered when this project was in its infancy and decided against. My reasoning for not rounding to the nearest 5% was that yes, I am aware that most items have actual drop rates that are multiples of 5%, but I would be wary of listing an item whose drop rate is 1/9 as 10%. Unfortunately, rounding to the nearest fraction is meaningless and rounding to the nearest "smallish" fraction is extremely imprecisely defined for my taste. Also yes, I am aware that the accuracy on the statistics is not within 0.1%, but I would guess that it is typically within 1.5% of actual value. Really, all this is why I jumped on the statistics tables when they came up and started posting my actual drop numbers - anybody who just wants the chances can see the page, rounding off in their head however they want, anybody who wants to deal with the nitty-gritty can get the summary before calculation in the discussion page. Of course, if you can suggest a reasonable rounding method which preserves such fractions as 1/9 but still maintains the benefits of rounding, I'd be happy to begin implementing it.--[[User:Yiab|Yiab]] 06:08, 22 April 2006 (CDT)

Note: I recalculated the drops rates of all double-drop monsters in the kingdom to reflect the possibility that the double-dropped item might have different drop rates for the two drops. After sleep I will do the triple-drop items similarly. I chose to do this now because I was bored at 2am while listening to an old Jick show on which it is mentioned that every [[The Spectral Pickle Factory|pickle monster]] drops the same item twice, with one dropping at half the other's rate, so I actually got around to making an Excel spreadsheet to calculate drop rates with both single- and double-drop numbers for me. And I'm ranting due to lack of sleep.--[[User:Yiab|Yiab]] 03:47, 8 October 2006 (CDT)

*Update: After some formula twisting I've decided to just forget about this method of finding different drop rates on non-8-bit triple-drop monsters and I'll get to another method once I'm done finding the drop rates in the haunted mansion.--[[User:Yiab|Yiab]] 13:24, 9 October 2006 (CDT)

== Icy Peak ==

For a while I've been curious about the drops at the icy peak.

The icy peak lists as 30% penguin skin and 30% yeti fur.  With drop increasers, I seem to max out at 33% penguin and 66% yeti.  This makes me think there is more than one yeti.  Say, one with 90% penguin and 45% yeti, one with 45% yeti, and one with no penguin or yeti.  I haven't really come up with a good way to verify this.

--[[User:Crowther|Crowther]] 10:02, 21 January 2006 (Central Standard Time)

* The drop rate of penguin skins has changed. I haven't kept many statistics, but it seems to be around 10% now, rather than 30%. Your data would indicate a 15% drop rate, which is certainly possible. Either way, the values need to be changed. --[[User:Snickles|Snickles]] 14:02, 21 January 2006 (Central Standard Time)

:* My data indicates approximately a 30% yeti fur drop rate and 22.5% drop rate for penguin skin. Of course, that may be mixed between pre-change and post-change, so I should really whipe my icy peak records and check again.--[[User:Yiab|Yiab]] 16:06, 21 January 2006 (Central Standard Time)

::* The change was first noticed between the 29th and 31st of December (to the best of my knowledge), so your data likely fell before the change or in between. --[[User:Snickles|Snickles]] 17:43, 21 January 2006 (Central Standard Time)

== Meat Drops ==
It is worth noting that, counter to my expectations, meat drops appear to follow a Gaussian distribution rather than a constant distribution.--[[User:Yiab|Yiab]] 21:49, 19 May 2006 (CDT)

== Statistics distributed throughout the wiki ==

I'm not sure if this is the right page for this but I can't find anywhere else with this info.  What I would love to see is the percentage rates of encounters and item drops distributed among all the Item and Location pages on the wiki.  Such as, on a Location page, list the percentage rate of each encounter listed on that page.  I know it gets skewed a little depending on super-likelies etc. but these could surely be accounted for, such as giving a percentage range, or ignoring special encounters when calculating normal encounters.  On each individual encounter page, repeat the same percentage chances.  And the most useful suggestion would be to include percentages on every Item page.  In the section on the Item pages that list all the locations and monsters/encounters that drop the items, multiply the percentage chance of that monster/encounter appearing with the percentage drop of that item.  There must be a really easy way to do this from the Parseable Area and Item statistics pages.--[[User:Urutsini|Urutsini]] 03:13, 23 May 2009 (UTC)

== Distribution of meat drops ==

After talking to RoyalTonberry, it appears that meat drops follow a triangular distribution rather than a normal. See here: [http://ben.bloomroad.com/kol/smash/black_knight_meat.html]. As such, this page has been modified. --[[User:Irregular|Irregular]] 07:21, 1 March 2010 (UTC)
*Yes, they do.  [http://alliancefromhell.com/viewtopic.php?p=25799#p25799 Eleron] had established that some time ago.  It's used in {{tl|meat}}, as well, to generate the proper meat distribution range (and standard deviation, up to me making some silly typo).  I'm sure there's a KoL forums post about the formula, as well, but I don't have the link handy at the moment.  --[[User:Flargen|Flargen]] 08:29, 1 March 2010 (UTC)