Talk:Search Odds

From The Urban Dead Wiki
Revision as of 01:28, 13 March 2008 by Airborne88 (talk | contribs) (→‎Searching on streets?)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to navigationJump to search

Archives

Thoughts on New Search Rates

Okay, just because it's fun, Let's use what we know now, to see if we can predict what our results will be with the two new search rates.

We're fairly certain that the base rate in most locations is 1/5 success, and we have evidence that suggests that the Bargain hunting rate is 1/3. If this is the case, we're looking at frational rates, not decimal rates, and likely any increase or decrease will probably change fractions, rather than a standard subtract or add.

So, my current expectations for the Search rates in all conditions, from best to worst:

Condition Rate
Mall, Bargain Hunting, Lights On 1/2
Mall, Bargain Hunting, Lights Off 1/3
Mall, Bargain Hunting, Lights On, Ransack 1/3
Mall, Bargain Hunting, Lights Off, Ransack 1/4
Other Location, Lights On 1/4
Other Location, Lights On, Ransack 1/5
Other Location, Lights Off 1/5
Other Location, Lights Off, Ransack 1/6

Naturally, a lot of further testing is going to be required. -- Odd Starter talkModW! 06:51, 30 April 2006 (BST)

  • I don't think it has been mentioned anywhere here, but changes were made to NecroTech buildings at the same time the syringe rules changed. I gathered some data using DHPD officers, but the files are at home. In any case, it should be noted that prior to April 29th books were eliminated from Necrotech buildings and syringe find rates averaged roughly 11% (Of total searches). I think a complete new survey of NT buildings needs to be done.--Bill Turner 14:46, 30 April 2006 (BST)
  • A system based on changing the divisor is inherently unfair to the zombies (e.g. 1/5 to 1/6 is -3.3%, 1/5 to 1/4 is +5%) so we should consider denominator changes too. I seem to recall someone suggesting the non-bargain search rate might actually be 2/9 (making bargain hunting an exact +50% as well). If there's enough present data to pass judgement on that hypothesis, it might be a pointer to what we should expect. - Dashiva 03:15, 1 May 2006 (BST)
  • Yep, I think it is more likely that the rates will be changing from 2/9 to 3/9 to 4/9. Also, 50% seems like a scary-high find rate for Caiger Mall... --Tycho44 W! 03:25, 1 May 2006 (BST)

Template Change for New Search Rates

On the search pages the current instructions for recording if the building is ransacked or lit is to add a letter (or letters) next to your number of searches. Instead of doing this, why not change the ‘running generator’ column to ‘condition’? Also, I suggest starting a new data table on each search page for post update data. (I’ve changed NT buildings and Hospitals to the above so far.)--Gamer.g33k 17:27, 1 May 2006 (BST)

That was the other possibility, but I figured that all up it was probably just quicker and simpler to add it to the no of searches field. With that said, if you want to go through all the pages and set up the new column, feel free :) -- Odd Starter talkModW! 04:41, 3 May 2006 (BST)
Since there are no objections I'm getting started. Gamer.g33k 02:15, 4 May 2006 (BST)
Okay, I'm done! Over the next few day I'll check to see if I left anything out or messed anything up (right now I'm about to fall asleep in my chair). Also, I'll try to search in each of the diferent buildings over the next few weeks so I can replace the "Your name here!" placeholders I put in (I'm currently in a mall, so I'll start there). Gamer.g33k 03:07, 4 May 2006 (BST)
Update: I'm slowly making progess, but I need someone without bargin hunting to go through to mall. I'll need to spend the next few days emptying my inventory so I can search some more. Gamer.g33k 16:24, 8 May 2006 (BST)
Question/suggestion? On the Search data pages, the codes are L => Lights on, R => Ransacked. However, I wonder if there shouldn't be an explicit "Lights off, not ransacked" status. I think this would clarify, "did someone just forget to fill in the status? Or was it neither lights on, nor ransacked. E.g., you can simply add "N" as a code for "neither" or "No lights, Not ransacked". ZaqWer 16:53, 13 May 2006 (BST)
Well, the idea was that each code specified an active condition, and these could be combined, so the absence of both conditions would equate to "regular conditions". I mean, we can add a new code, but it's not require, IMHO. -- Odd Starter talkModW! 10:20, 14 May 2006 (BST)
I like the suggestion as it should help with the data interpretation later on. I've changed 'building conditions' part of the template (adding "n" for nothing and "lr" for a ransacked building with lights on {I just found out this was even possible}). Gamer.g33k 15:27, 17 May 2006 (BST)
Good grief, I feel like such a n00b. Yesterday when I searched in a hospital, I thought the “building is in disarray�? message meant that it had been ransacked. Looking at the wiki today I saw that for a hospital, it will actually say ransacked. So I guess we still don’t know if a “LR�? building condition is possible. Should I leave the LR or should I yank it? Gamer.g33k 12:21, 19 May 2006 (BST)
There is every reason to believe a place can be lit, yet ransacked. Starting up a generator is an independent action from clearing ransackedness (word)? That's what Odd Starter meant by "these [two conditions] could be combined." SearchDerelict 21:56, 19 May 2006 (BST)

Searching on streets?

Street.JPG As seen in the picture, it is actually possbile to search for things on streets, but possiblly at a really low rate. Personally I think this will be a rather interesting thing to research on.

There is a page for Search_Odds/Outside already. Your shotgun shell will be only the second on record found outside (on this wiki). If you have interest in doing more 'on the streets' research, I'm sure everyone here would be interested in seeing it. Gamer.g33k 19:11, 4 May 2006 (BST)
Sorry, did not realise it is already covered. Will try my best to do something about it. --Changchad 19:48, 4 May 2006 (BST)
I have found a shotgun shell & a pistol clip outside. Did it when I first started playing and had no idea what I was doing.. So, no screen shot.--Airborne88Zzz1.JPGT|ZC|MI|E! 01:28, 13 March 2008 (UTC)

Museums

I was in a museum that housed a collection of European art (according to the building description) and all I was able to find were European paintings, torn paintings, and burned paintings. Do different museums offer different items? This is worth researching.--Koppie 17:45, 21 March 2007 (UTC)

You can fnd some information on this here and here, though there were no research of search odds yet... --Duke Garland 20:36, 21 April 2007 (BST)

Greasemonkey Script

I haven't played Urban Dead in.... a year? Two? But I did join a game called Cities a while back, and recently I wrote a data-collecting Greasemonkey script for it. I was active on the Search Odds pages for a while, and I greatly appreciated the treatises on statistics. Therefore, I present a zipped copy of the greasemonkey script, the PHP data-receiver script, a (slightly old) copy of the database, and the PHP webpages used to display the data. The results can be seen over here, which is a PHP webpage generated on the fly from the database.

The script collects data from 17 possible sources, so for a basic search odds script, it can be pared down massively. Appropriately enough, the "Searching" section is probably the most useful, though the "Mining" section is also pretty close to what you're looking for.

File layout:

The greasemonkey script is the file citiesbigbrother.user.js.
The PHP data-receiver script is dbaseInserter.php.
The database (for reference only, really) is potatoengineer_db 3.29.07.sql.
version_check.php is the file that the greasemonkey script checks once per hour, to see if a new version of the script has come out. (Very important for preventing old, buggy versions from fouling the database!)
The rest of the .php files are the various webpages to display data.

This setup requires somebody with a webspace that does PHP and has a MySQL database. I use 110mb.com, because it's free and I'm cheap, but a paid site would have better uptime and ping times, and there are other free sites out there, too. The other minor detail is that this script was written for Cities, not Urban Dead, so somebody is going to have to change it. Hopefully the comments I've scattered through it are enough. PotatoEngineer 12:24, 9 April 2007 (BST)

Ruined buildings worse then ransacked?

I haven't gotten enough data points to say yet, but does anyone suspect that Ruined builds have even worse search rates then Ransacked ones? --Morgan Blair 00:48, 18 August 2007 (BST)

CMIIW, but what can be below zero? --~~~~T''' 09:42, 18 August 2007 (BST)
IIRC Ransack search odds aren't quite zero percent, they're just close enough that nobody bothers to search. --  T   10:31, 18 August 2007 (BST)
I've found things in ruined buildings, but the rate is so low that I think it's worse then the ransacked rates. By the way, did Odd Starter's estimates about base search rates prove to be true? 1 in 5 for no condition, 1 in 4 for lit, and 1 in 6 for ransacked was what he predicted. --Morgan Blair 22:26, 20 August 2007 (BST)
Found a first-aid kit and two newspapers in a ruined hospital today, within about ten searches. I don't think the rates are any less than for ransacked buildings. Besides, if you think the search rates system works according to 1/9, 2/9, 3/9, up to 4/9 for bargain hunting a lit mall, which makes some sense, then you can't get any less than 1/9 anyway. --MeatHead 21:07, 25 August 2007 (BST)

I doubt this question will be ever answered... --~~~~ [talk] 10:19, 2 January 2008 (UTC)

Prototype of a redesign for the Search Data table

Hey, some—most, probably—may not think it is needed, but what with the recent addition to both conditions as well as searchable items, I thought I'd try my hand at a redesign, so that the data tables would more clearly outline what types of data have been collected (search conditions), and subsequently make then a tad more usable, in that it would be easier to visually estimate search rates for items under a specific condition, something that is difficult under the current design. So, if you care, meander on over to Search_Odds/Fort_Data/Storehouse_Data, where I've put up a prototype (the page didn't exist until yesterday, and no data has been collected for this location, so it shouldn't ruffle any feathers). While it may look a little over-complex, it's really just as simple as the old table. But if you have any constructive suggestions, please make them on the talk page over there. Oh, I suppose you could make non-constructive criticisms as well... --Morgan Blair 22:39, 20 August 2007 (BST)

See my comments in there. Many table for one need is bad. The common system is much more comfortable --~~~~T''' 09:03, 22 August 2007 (BST)

Toolboxes

Will anyone be volunteering to ascertain the search odds for toolboxes? We have very little search information on toolboxes right now. Forlorad 16:02, 31 August 2007 (BST)

tables are ready, waiting for data income... --~~~~T''' 16:27, 31 August 2007 (BST)
Bleh, there's probably less motivation for people to collect search data for toolboxes because they're like flak jackets or radios - one is enough. Forlorad 01:17, 5 September 2007 (BST)

I've resurrected SearchDerelict2 who only does these kinds of jobs. So, I'll start accumulating some data on toolboxes in Auto Repair and Factory. (Was trying to accumulate data on Ransacked, but it was too hard to stay alive.) --SearchDerelict 17:36, 28 October 2007 (UTC)

Does it help knowing that im currently trying to find a tool box in a unpowered ruined building and have found nothing after well over 100 searches?--Carnexhat 20:24, 26 February 2008 (UTC)
No, unless you provide a type of building you are searching on and more info about how you search, other alts possible influence, etc. --~~~~ [talk] 00:18, 29 February 2008 (UTC)

Total Renewing

If you watch the page, you might've seen what i was doing with it recently. Well, data tables are as much uptodate as possible with current (lack) of data. I'll also enhance analysis section, but so far i've done renewing. --~~~~ [talk] 15:35, 25 September 2007 (BST)

Mansions are searchable

proof --~~~~ [talk] 21:41, 5 October 2007 (BST)

Does Fog Affect Searching?

Appears to me that fog (introduced for Halloween) severely limits searching? Either that or I had an unusually unlucky day. Inside a *lit* Auto Repair 35 searches yeilded one newspaper and nothing else. I've decided not to add this data to the data tables as it may skew the results. --SearchDerelict 23:50, 31 October 2007 (UTC)

I find faks in unlit hospital with pretty normal rate right now - 8 times nothing and then 2 in 8 searches. Anyway, if there are any concerns - it's better not to add any data until fog lifts --~~~~ [talk] 07:08, 1 November 2007 (UTC)

New Search Text on Hospitals?

Searched 25 times in a hospital today. 2 times found a FAK, 1 time found a newspaper. 2 times got new flavor text: "Searching the wards you gather together enough supplies for a first aid kit." I figured this was just new flavor for "You find a first aid kit." However, at the end of my turn I noted I only had 2 FAK, not 4 as expected. Not sure what to make of this. Will try again. --SearchDerelict 18:14, 10 November 2007 (UTC)

sigh... News#14th_August_2007: Some minor flavor was added to messages, in particular for a successful FAK search --~~~~ [talk] 18:32, 10 November 2007 (UTC)
Well, today, I only had the new flavor text and confirmed each one associated with finding a FAK. I "Found a first aid kit" twice and "Found enough supplies to put one together" twice but only confirmed my inventory as +2 FAK at the end. --SearchDerelict 23:26, 11 November 2007 (UTC)
Hm..? old text "found a first aid kit" shouldn't appear any more, unless it's a "and discarded it as useless"... Bug? --~~~~ [talk] 23:30, 11 November 2007 (UTC)

Effects of light in NT

Maybe it's just lack of data and limitation of having to take in account only recen "state N" data, but it seems from current numbers that powering up NT increases search odds for DNA Exs and GPSs, but search rate for syringes remains the same... That's quite weird --~~~~ [talk] 17:48, 11 November 2007 (UTC)

New page: Park

Let's see how often dices roll out trees --~~~~ [talk] 21:51, 12 December 2007 (UTC)

That is the weirdest sentence I've seen all day. :D --Toejam 22:23, 12 December 2007 (UTC)
You didn't know that all the randomity in UD is in real life Kevan throwing dices? :P --~~~~ [talk] 07:39, 13 December 2007 (UTC)

Malls searchs nerfed?

Lately, since about 3-4 days ago searchs in Malls became worse than they used to be. I have a suspicion that BH was nerfed a bit. --~~~~ [talk] 22:06, 22 January 2008 (UTC)

toolbox searches quietly nerfed?

don't seem to find them as much as before these days... maybe just no luck, though --~~~~ [talk] 21:17, 28 January 2008 (UTC)

RESPONSES REQUESTED: Search rate variability? (Feb 2008)

I've done some looking and can't find this referenced anywhere, but I'm concerned that there may be more depth to the search algorithm than we know. I just created an alt to hang out an listen in on all ten 25.9X channels hoping to pick up surprising military broadcasts, yada yada. I've spent my first 100AP in one of the PDs searching to get radios, and I've noticed something very strange: Almost *everything* I've found are shotguns and shotgun shells. Like, two radios, four shotguns, and 40 shotgun shells. This is way outside of the search probabilities we'd expect to see, both in success rate and in types...is it possible that different locations have different probabilities more specific than just their building type? Has anyone else seen anything comparable to this? --Hungrygrok 00:18, 16 February 2008 (UTC)

Generally it's the smallest measurable variable on the tool used to get the data. And also keep in mind that the RNG is well noted for its tendency towards streaks. As for specific locations having higher rates, no, nothing like that has ever been checked as far as I can see but, I really doubt it's the case.--Karekmaps?! 03:03, 16 February 2008 (UTC)
Karek, can you expound on your first sentence for me? --Hungrygrok 02:20, 20 February 2008 (UTC)
It's quite impossible to check, we have barely enough data to make some assumptions on search rates of PDs based on all searches scrobbled, and you want to have that much from concrete separate PDs... Also, there's no reason for Kevan to program different values for different PDs. By the way, that isn't way outside. --~~~~ [talk] 12:41, 16 February 2008 (UTC)
Am I misreading our expected rates? I was seeing a 6% likelihood of shotgun shells per search, which makes a 50% find rate over 100 iterations more than an expected anomaly. --Hungrygrok 02:20, 20 February 2008 (UTC)