User talk:Starwed

From TheKolWiki
Jump to: navigation, search

Recipe

A test.

Uncle-30px.gif 2 googly eyes 4 stuffings 2 felts
Equals.gif Rag doll

Your infobox template

I really like what you have there with your infobox, however I suggest you make it float so that it does not affect the centering of the rest of the page. (I am not sure if what I just said makes sense, but you get the idea :))--SomeStranger (t|c) 22:39, 12 October 2006 (CDT)

It is floating, but still affects the centering.  :( I know that there are CSS styles which would work here, but I think it's hard to get them consistant across browsers. --Starwed 01:31, 14 October 2006 (CDT)
  • Now that we're moving towards adopting your infobox idea, I've set up copies of your monster and location templates as stand-alone templates. The new templates are now cat agorized as test templates, been tweaked to put in some standard template "stuff", and have been renamed in accordance to our standard template naming conventions. Would you mind revising your example to use them while also tweaking Discussion/Infoboxes so that links use these new templates? Thanks. --Gymnosophist 03:38, 16 October 2006 (CDT)
    • Done and done. But anyone should feel free to modify the mockup stuff I've done w/o asking. (Especially because I'm likely to become extremely busy with HW in the next couple of days.)--Starwed 08:03, 16 October 2006 (CDT)
  • kudos on the template-wrangling, but would it be possible for you to mark your edits as minor? you've pretty much filled the recent changes page. --Evilkolbot 06:37, 17 October 2006 (CDT)
  • Is it jumping the gun to be pre-making all the location infoboxes? We're still discussing how infoboxes will be implemented. One possible change in particular, the use of /data pages for locations, might make obsolete much of your advance work. --Gymnosophist 16:14, 10 November 2006 (CST)
    • Nah, it takes about 30 seconds per box; I did most of that while waiting on something earlier today. And the problem is, there hadn't been any discussion on the format for about two weeks. I said a few days ago that if no one made any comments to the contrary, I'd simply start putting them in pages, and this is basically prep work to do that over the weekend. Now, I personally think that there hasn't been enough discussion on the matter, but this seemed the easiest way to spur that discussion. ^_^ Additionally, using data pages would simply mean altering the name of the template; the actual format of the parameter assignment would be completely unchanged. --Starwed 16:28, 10 November 2006 (CST)
    • Yeah, getting movement on discussion can sometimes be like pulling teeth! I'd been a bit slow myself recently on getting the bit between my teeth on infoboxes - my biorhythms had been at a low ebb. :) FYI - infoboxes will happen - in an offline discussion we (Wiki staff) decided to try to get them implemented by early December (once again, many thanks to you for your idea and your efforts in developing it). Speaking of discussion, what are your thoughts on the various open infobox issues? --Gymnosophist 21:50, 10 November 2006 (CST)
    • The problem with stuff like the item data pages, is that I don't know how they're planning to be used well enough to comment. --Starwed 14:30, 12 November 2006 (CST)


{{{item1}}}HOLLA* Bones

Stone spheres

Now that you've taken all the "when used" information off all the stone sphere pages and put in that link to an empty page, will you be putting anything in that page any time soon? Or would you mind if I started filling in the basics? -- Old Ned 21:21, 4 July 2007 (CDT)

  • It's working now, thanks! -- Old Ned 22:04, 4 July 2007 (CDT)

The Wossname

You're right; the fortune cookie idea is a valid one. However, the way it was phrased, it didn't make sense; you're not making the adventure semi-rare, you're waiting for the adventure to be semi-rare before you use it. As such, I had to remove it from the table, but thank you for the clarification. --Unnatural20 00:33, 16 July 2007 (CDT)

Effects from Monsters

I think that the "Potential effect: ______" might work better, because what you put is usually on the actually on the monster's page. However, if you look at DoD, all of the messages are in the same form. That is why I'm reverting it to my version. --CG1:t,c,e 17:48, 21 July 2007 (CDT)

Wad/nugget/powder source mockup

List of useful things to pulverize for elemental powder/nugget/wad in HC.

All

Spooky Powder

Item Conditional Odds Result
Spooky
3ball.gif
3-ball
1/3 Scpowder.gif
spooky powder
Clownpants.gif
bloody clown pants
1/3 X powder
Buckler1.gif
clownskin buckler
Hatpants.gif 1/3 X powder
Harness.gif
clownskin harness
Hatpants.gif 1/3 X powder

Boss bat vs. guard stats

normal

gbgbggbggbgbgbggbbgbbgbbbbbbbgbbggbggbbbbbbgbbbg

Odor extractor

ggggggbggggggggbgbggggggggggggbgggggggg

Data pages

Yes, adding by hand because I'm an idiot :P

Hp, Attack, Defense, and Initiative come from MonsterStats. HP, Attack and Defense are from Kittiwake's data. Initiative may not be reporting to base correctly, some of it is undefined, etc.

Items come from MonsterStats from Yiab, mostly, except the new ones (like the kitchen drops).

Images are from the wiki; location's being taken from the compendium we have up.

Name is from the wiki. I tend to take the page name except when it's wrong, like the frat boys (and the unadded ninja snowmen).

Meat is...tricky. I'm not taking any solid value, normally taking the minimum/maximum values I can find. MonsterStats using Ragnok's data disagrees with the wiki's monster pages, which also tend to disagree with the combat template in the locations they're in. Also, MonsterStats sometimes just doesn't show meat even when it is dropped, so that doesn't help - also, some of the wiki combat pages don't show a meat drop when there is one, such as black adder. We should probably focus on a single site for this, or just retake all values ourselves, because they don't look right as is.

...of course, now that I actually check, it appears the wiki has mass parseable information pages, all linked relevant for monsters. Missing the new NS13 ones, but most of it's there I think.

Ryo Sangnoir 06:14, 23 August 2007 (CDT)

An offer

Would you like the privilege of wiki admin-hood? I'm looking to expand the team slightly, and you have been helpful and worked well with existing admins since you've been here. Let me know one way or the other. --jin 12:52, 13 October 2007 (CDT)

TPTBW

Would you mind moving tiny plastic the Big Wisniewski... Jick changed the name yet again, and the Wiki complains when I try to fixy. --Bagatelle 14:34, 2 January 2008 (CST)

Drop Locations

Sorry: I didn't see the discussion on it... Maybe a template would help... In any case ask Quietust to see if he can use QuietBot to help... --CG1:t,c,e 14:52, 4 February 2008 (CST)

Talk:Player Hit Messages#New Ones

  • obviously, by "most exist", i meant out of nine proposed, four exist, which is stretching the meaning of most. shouldn't fill in explanatory text before you've finished. oh well. --Evilkolbot 05:13, 10 February 2008 (CST)

Statistics Question

With regard to your statistics page, I've been wondering why the typical method of weighting when combining results uses 1/StdErr2 for weights. It seems to me more natural to use sample size (when that number is available, as in item drop data gathering). I bring this up because I implemented the statistics you outline for my data about Scary pirate (cursed) and the weighted means are coming up differently.

Weighting results by sample size alone gives me a drop rate of 0.0019072183
Weighting results by 1/StdErr2 gives me a drop rate of 0.0016551708
and a 95% error margin of 0.0002978118, barely including the former in the 95% certainty interval for the latter.

It seems to me that the net result of using n/p(1-p) as a weight instead of n alone is to give more weight to values closer to 0 (or 1). Maybe you can help clear up this stuff for me - I've never been very good at understanding statistics. Thank you for your time.--Yiab 13:31, 3 April 2008 (CDT)

I'm going to take a longer look at this, but one point is that, in finding the (weighted) average for the base rate, the %item drop affects the standard error of each sample. That's what's really important for finding the weighted average, otherwise you could just combine them directly.
The p(1-p) term is supposed to take into account the average variation you'd see, based on the "real" rate. For consistency, that "real rate" should be the same in each and every sample, meaning that if the samples differed by only N, it would reproduce what you'd get by just adding them together.
Here, it's not the base rate we're observing, but a rate augmented by a multiplier m (caused by +item drop): p → p' = p×m. We want to use this rate p' when calculating the error of each sample. Luckily, when p is really damn small, we can readily approximate p(1-p) by just p, since the p2 term is 3 orders of magnitude less than p!
So we now know that the correct estimate of the standard error, and weighting factor for each sample will be
σ = 1/m × √( m×p/N)
W = 1/σ2 = N×m/p
Since p is in common to all weighting factors, we can instead use a weighting factor of:
W = N×m
I think this is the consistent way to define the weighting factor here. But statistics is damn tricky, and I'm not actually trained in that area, so I'll have to think/research a bit more about it. (I feel like there's probably a more formal method to find the most likely base rate here.)
Final note: With this method I calculate a rate of 0.25%, +/- .03%. But I get the same result when simply weighting by N (since most of the item boosts are about the same, this makes sense.) This disagrees with the number you quote for weighting by N: I could be misunderstanding how your data is presented, or one of us could be just miscalculating.  :)

--Starwed 15:11, 3 April 2008 (CDT)

Well, I have about twice as much data now than I did when I last updated the table on the wiki. I'd be happy to type up a pdf description of the spreadsheet I'm using for this (or just send over the spreadsheet) if you'd like. I appreciate all the help you've been so far.--Yiab 14:53, 4 April 2008 (CDT)

Garbage and blank sections

What's going on here? Somebody's putting junk in your talk page... --Wrldwzrd89 23:18, 28 November 2010 (UTC)

advmin/max

Hi Starwed, the vast majority of the Food pages has the form adv: x-y and only a few have the form advmin: x advmax: y As far as I know the reported average is computed (assuming a symmetrical distribution) from these two numbers and can not be explicitly specified. So the information in both forms is identical and computing the average of two integers lies within everybody's possibilities.

Therefore it seemed to make sense to use one form for everything and eliminating the rarely used form seems to be the easiest solution.

You have not yet convinced me that I am doing something wrong. Eliza Dane

  • It's not wrong per se. Just less useful. advmin/advmax gives the wiki reader more information, without requiring them to go out of their way, and lets the template do stuff that it can't do with an input of "20-26". The latter would numerically be evaluated to -6, for example! It's also part of a larger project in mind to increase the power of templates and data pages, and separating the min/max makes it a lot easier for the wiki to auto-detect when, say, an item's in-game quality does not match up with the adventure gain data on the wiki. And many other things in many other places. Such as vastly improving and simplifying Best Foods pages. --Flargen 09:41, 11 December 2010 (UTC)

Template:Combat

Hey, I noticed that you were working on User:Starwed/Combat, and just thought you should know that I've been doing quite a bit of work on Template:Test/Combat today, which includes "tag" support.

--Eleron 09:05, 11 December 2010 (UTC)

I was mostly trying to figure out how to get the indentation working correctly (see a bit more about that at Discussion.) Saw you working on the combat template -- glad you're abstracting all the Data stuff into a separate template, that's much cleaner than the complete mess that the current Combat template is. --Starwed 09:32, 11 December 2010 (UTC)
Flargen helped me add indentation to the CSS and pushed the new combat template, so now everything should hopefully be working :) --Eleron 10:58, 11 December 2010 (UTC)

Template:GroupMonster

Thanks for putting that remark on my talk page about the beatles being a group of 4. I fixed the monster page. --Wrldwzrd89 21:35, 28 June 2011 (CEST)

Conditional Encounters

Could you start a page about encounter conditionals and how they work? Merging/revamping Superlikely, Floors and Ceilings, including e.g. rampaging adding machine and just adding how things work and removing/improving the text. --Eleron 20:37, 22 January 2012 (CET)