Difference between revisions of "Talk:Adventurer Theories"

From BattleMaster Wiki
Jump to navigation Jump to search
(→‎Gathering: Speculation)
m (→‎Gathering: (signing))
Line 96: Line 96:
 
The "research" on gathering is very simplified. Among other things, there's no fixed 25% chance. The actual chances of finding stuff and what you find depend on various factors, none of them the amount of hours you invest. But region type, for example, is a factor. There's a reason for it having different "you find nothing" texts. It '''is''' a difference if you find nothing or are disturbed by militia, for example, and the chances for each are different in different region types. --[[User:Tom|Tom]] 22:41, 16 January 2007 (CET)
 
The "research" on gathering is very simplified. Among other things, there's no fixed 25% chance. The actual chances of finding stuff and what you find depend on various factors, none of them the amount of hours you invest. But region type, for example, is a factor. There's a reason for it having different "you find nothing" texts. It '''is''' a difference if you find nothing or are disturbed by militia, for example, and the chances for each are different in different region types. --[[User:Tom|Tom]] 22:41, 16 January 2007 (CET)
  
* My personal bet, based on how I'd design it, is that there's a flat 50% chance of finding nothing for an hour ("You find nothing.")  There's also a chance to be disturbed, based on troop amounts and peasant population, possibly modified by adventuring skill (unknown how to test this; we can't get adventuring skill estimates).  Then, there's a random roll based on hidden region statistics (how fertile it is for items) and adventuring skill as to whether or not your search finds anything ("You find nothing ''of value''" versus "You find a foo!").
+
* My personal bet, based on how I'd design it, is that there's a flat 50% chance of finding nothing for an hour ("You find nothing.")  There's also a chance to be disturbed, based on troop amounts and peasant population, possibly modified by adventuring skill (unknown how to test this; we can't get adventuring skill estimates).  Then, there's a random roll based on hidden region statistics (how fertile it is for items) and adventuring skill as to whether or not your search finds anything ("You find nothing ''of value''" versus "You find a foo!"). [[User:Majromax|Majromax]] 07:30, 26 January 2007 (CET)

Revision as of 07:30, 26 January 2007

Thoughts on layout, readability, ease of use or just general comments go here... Any observations should be made on the page itself. --The1exile 23:23, 21 December 2006 (CET)

I'll fill it up once I start resting again... Damn champion seriously wounded me... and I wanted a unique :( -Chénier 23:29, 21 December 2006 (CET)

Actually, I haven't insisted that it's non-linear. I did, however, wonder why you think it is. Not many things in BM are linear... --Tom 00:08, 22 December 2006 (CET)

Hmm.. This reminds me of my university time. True, you can approximate an unknown function, and often within a limited range the easiest approximation is a linear function. Very interesting. I'm surprised nobody has applied higher math yet, there should be enough geeks in the game. --Tom 15:27, 26 December 2006 (CET)
Yeah, there is one here. -- Gsklee 16:08, 26 December 2006 (CET)
I hadn't crossed out the idea, but liniar functions with a minute random factor seems to fit nicely. Also, for most games, server speed is an issue, I guessed that expodential functions weren't likely. --Ambeco 01:56, 27 December 2006 (CET)
CPU power is one of the least problems I have. I/O and memory are the main issues of the database. Calculations for the PHP code are small compared to that. --Tom 18:46, 27 December 2006 (CET)

Math

Seems like to be exponential functions; correlation coefficients are high when the data are approximated this way. Also fit into my experience with the game.

On the Floor

f(x) = 2.8837e0.1607x

Primitive Room

f(x) = 6.1564e0.1366x

Common Room

f(x) = 4.8468e0.2208x

-- Gsklee 16:54, 26 December 2006 (CET)

You guys are impressive. Very impressive. I love this game. :) --Calvin November t/c on 04:09, 27 December 2006 (CET)
If those are the correct formulas, what are the restrictions on the value of e (unless you mean Euler's (big chance of spelling mistake)), we could, and should, simply make the calculations to see the maximum and minimum possibility for every value of x and we'd be able to fill out the chart accuratly. -Chénier 21:13, 31 December 2006 (CET)
Euler's value doesn't work out. I even wonder how he got to those strange numbers, it doesn't seem logical to me that Tom just filled in random numbers, not when you try to imitate a real sleeping process. Foreign Curs 10:15, 1 January 2007 (CET)
I'm thinking e is not Euler's, but a random number between x and y, seing as how sleeping the same ammount of time the same way have removed different ammounts of fatigue. But I'm not sure, I didn't come up with them... -Chénier 18:51, 1 January 2007 (CET)
pretty sure e means "Natural Log"... --Grancourt Family 14:02, 06 January 2007 (CET)

Didn't I say these are approximated results? If you want a universal function it could be represented in the following way:
η = ⌊αeβx + r
With x being the hours spent, α and β are constants that vary in each resting method, r a randomize factor, η the fatigue restored, and ⌊⌋ for a floor function to obtain an integer result. -- Gsklee 21:49, 1 January 2007 (CET)

Ha! Now you lost me! That's alright, I'll let you handle those things... -Chénier 21:58, 1 January 2007 (CET)
  • These obviously can't be the real formulas, since no sane programmer would come up with them. However, exp(.2208) == 1.247 ~= 1.25, so I think we're on the right track. My guess for the general formula is int(A + C*k^t + rand), where A, C, and K are rest-dependant factors and rand is a random perturbation. The randomness could also be multiplicative in the C. How does the correlation work out with that functional form? Majromax 07:03, 10 January 2007 (CET)
  • Two examples: 10 hours sleep in common room reduced fatigue by 32, 14 hours sleep in common room reduced fatigue by 42.

why functions ?

Why do you all assume it's a simple function at all? There are only 16 possible values, it could be a table. Or it could be a more complex or multiple overlapping functions - I quite like those. There are several places in the game where the function is y = sqrt(x)+x/10 or something like that. There are also various places where it's like y = min(a(x), b(x)).

Then again, I might just be having fun with you and it actually is 2*x + rand(1,3)... :-)

--Tom 16:08, 2 January 2007 (CET)

One day a physicist was bragging to his friend about the Almightiness of The Physics, claiming it being The Answer to Life, the Universe, and Everything... The friend snorted, asking the physicist to describe the motion of a chicken.
"Of course!" The physicist replied in excitement. "Assuming the chicken is a perfect spherical rigid body..."


You see, that's the spirit! ;) -- Gsklee 20:45, 2 January 2007 (CET)

Actually, why even bother? I mean we've got half of the table already written down, what use is it to find the function behind it? If we've got the numbers in front of us, I won't calculate the solution to that function to check again... Neither will I try to memorize all those diffrent functions for each type of sleeping. This is useless work actually, as we can just check the table for the answer in two seconds :D Foreign Curs 22:31, 9 January 2007 (CET)

Investigation Texts

Whoever thought it was a bright idea to combine undead and monster investigation texts - it wasn't. Do you have any proof whatsoever that the same texts are used for both? There is exactly one investigation text that is identical for both, and a few who are similar. So whoever did it, you acted based on either baseless assumptions or one data point each, or on not looking close enough ("you hear many rumours" and "there are many rumours" is not the same text...).

And you others: You accepted it without questioning.

It's always interesting to see how things work out, and how people go about finding hidden information.

--Tom 16:02, 2 January 2007 (CET)

Oops, that was me. At that point we had one text for monsters and two for undead, and the one monster text was one of the two undead texts (a 100% success rate!). We've got to make assumptions at some point. For example, do we have *proof* that an infiltrator assaulting a soldier has the same chance of success as an infiltrator assaulting a cavalier, all other things being equal? We don't, and the chances might not be equal, but we've got nothing that tells us they aren't and it makes sense for them to be the same. The texts all seemed like they could be used for either type, and you investigate monsters and undead at the same time with the same link, so it seemed reasonable to assume that the same texts were used for both. Clearly that was wrong though. And looking in the history, the two I thought were the same had a slight difference: "There are some rumors *of* monsters" vs. "There are some rumours *about* undead." That would seem to indicate that different code handles each. I guess we've got to be more careful. The fact that the monster and undead code are not the same might mean that there are some more important differences between them too. Any ideas? -- Mcglynn 18:56, 2 January 2007 (CET)

Sage Items

I highly doubt the usefulness of listing the sage items. These are unique items - don't you think the offers are likewise unique? Once the sage moves on, it's gone. --Tom 13:26, 6 January 2007 (CET)

At the very least, though, won't they tell us the kind of items we should keep around if we want to have a chance at making a unique item? You probably won't need a bedroll, it looks like, but be sure to hold on to those rare woodland flowers and monster parts, and probably in general the rarer "common" items. -- Mcglynn 15:42, 6 January 2007 (CET)
Tom, dunno if this is usual, but I saw the same sage in Batesaor no less than 3 times. --The1exile 16:10, 6 January 2007 (CET)
I also have seen a sage many times in Cutnipaniel(?), and twice they asked me for items, but i'm in Prison atm so can't check. --Grancourt Family 17:09, 6 January 2007 (CET)
I've now encountered a sage who asked for the same recipe two times in Gemke. -- Mcglynn 01:03, 8 January 2007 (CET)
Idem here, and that was after a two week interval, so I'm pretty sure he stays for a very long time.
I have removed the item recipes from the main page. One, it's nonsense to list them there because they do and will change. Two, there are hundreds of them that can and will happen over time and in various regions. So even if it could make sense to list them despite their changing, it's just not feasable. --Tom 15:19, 12 January 2007 (CET)

Adventurers and Infiltrators

Recently, an Adventurer in Watto was attacked by what I am assuming is an infiltrator. This strikes me as odd. Adventurers would be a fertile training ground for Infiltrators because they have no guards, making them easy to attack, and also cannot do Scouting or anything else to find a hidden Infiltrator. It sounds like no-risk training for infils. Am I way off the mark here? --Habap 16:29, 12 January 2007 (CET)

  • No, you are right, infiltrators can attack adventurers at will. It's the risk of being a commoner, we have no real security of men who swore alleigance to us, we are always in danger. I think though that adventurers, just like 'travelling nobles' are harder to track down. --Foreign Curs 20:28, 12 January 2007 (CET)

Gathering

(redirected texts from the article)

After trying out different methods of gathering i have found out that usign less hours results in a bigger find. For example, 4 x 2hours of searching will get you more items than 1 x 8 hours.

Ive tried this several times and it seems to work. This is with the standard boring option (cant think of the name) option. Has anyone else found this out? ScottSabin 18:22, 14 January 2007 (CET)

How much evidence is this based on? Gathering results seem to be calculated individually for each hour you spend, so why should it matter how much you spend total (unless maybe fatigue accumulates and prevents you from concentrating on the later hours)? -- Mcglynn 13:19, 15 January 2007 (CET)
Look, it has something to do with 'probability'. Each time you gather there is a chance of finding something. Now, probability doesn't have a memory. You'll find 1 item each four hours, at 25% rate of chance for each hour. At 3 hours you'll not have that 100% chance of finding something, because 3 hours x 25% is only 75%. As probability has no memory, it won't add 75% + 75% the next time you gather. SO each time you'll only have 75% chances at finding something. Resulting in multiple events when you find nothing (there is 25% you'll find nothing remember). The average is 20%. That's a loss. I can give you the full details about 5 hours too, and 6 and 7 and ... Foreign Curs 16:33, 16 January 2007 (CET)

Based on feedback Tom left on the page, I have removed Foreign Curs' gathering theories from the page and turned it into a more descriptive and basic guidelines approach. Basically, he said that the theory was simply wrong. (Quote: The above speaks with so much authority and yet, after just looking into the code to make sure, it isn't so. ) --Indirik 16:09, 17 January 2007 (CET)

There's still a lot of... well, guesswork in that section, and much of it is wrong. Maybe the difference between "I _know_" and "I _think_" should be made more obvious? I don't mind guesswork on the Wiki, I just prefer it clearly marked. --Tom 19:57, 17 January 2007 (CET)

Gathering

The "research" on gathering is very simplified. Among other things, there's no fixed 25% chance. The actual chances of finding stuff and what you find depend on various factors, none of them the amount of hours you invest. But region type, for example, is a factor. There's a reason for it having different "you find nothing" texts. It is a difference if you find nothing or are disturbed by militia, for example, and the chances for each are different in different region types. --Tom 22:41, 16 January 2007 (CET)

  • My personal bet, based on how I'd design it, is that there's a flat 50% chance of finding nothing for an hour ("You find nothing.") There's also a chance to be disturbed, based on troop amounts and peasant population, possibly modified by adventuring skill (unknown how to test this; we can't get adventuring skill estimates). Then, there's a random roll based on hidden region statistics (how fertile it is for items) and adventuring skill as to whether or not your search finds anything ("You find nothing of value" versus "You find a foo!"). Majromax 07:30, 26 January 2007 (CET)