1. #1
    former king of saturn
    Join Date
    February 2003
    Posts
    892

    Default Rare part system - WHY are there 'rare' rare parts?

    Can anyone enlighten me as to the purpose of having 'rare' rare parts? The rare pool is several hundred individual parts deep. Why is there any value in starving supply of a single item for every rare while players stock up on 10+ other parts for the rest of the rare they hope to build? They may as well just reinvent the entire rare system such that you only need a single part to drop to build your rare. It's the same damn thing in the end.

  2. #2
    Xpertz William Antrim's Avatar
    Join Date
    November 2005
    Location
    Norfolk.
    Posts
    4,112

    Default

    This topic came up a few months ago and was one of the driving forces behind the latest changes to the rare pool (the TS patch where rares now stack in one giant stack when unressed). The devs looked into the mechanics of rare drops and concluded that there is no "rarer" rare part. They all are as rare as one another apparently even though many of us have identified particularly rare parts.

    Some people gave the argument that there was a premium on certain parts to certain guns due to their viability (ATP of CS I am looking at you) and their effectiveness in pvp.

    Personally I remain undecided. I am not convinced that rare drops are totally random but I hope that the new changes will make things better for people when selling their "un-ressed" rares.
    "dulce et decorum est pro patria mori"

  3. #3
    freedom for neocron! Torg's Avatar
    Join Date
    March 2003
    Location
    wasteland
    Posts
    1,865

    Default

    Quote Originally Posted by Divide View Post
    Can anyone enlighten me....
    its a bug and will be fixed some time in the future.

  4. #4

    Default

    Quote Originally Posted by Divide View Post
    Can anyone enlighten me as to the purpose of having 'rare' rare parts? The rare pool is several hundred individual parts deep. Why is there any value in starving supply of a single item for every rare while players stock up on 10+ other parts for the rest of the rare they hope to build? They may as well just reinvent the entire rare system such that you only need a single part to drop to build your rare. It's the same damn thing in the end.
    There are no mechanisms or code or algorithms which cause the game to generate certain parts over others. Each time you obtain a rare part, you have the same chance of it being any single part in the rare pool. The resultant part is generated entirely at random so obtaining a lot of certain parts is entirely possible. Each time you obtain a rare part it has the same chance as being the same as the last one you researched, as it does being a completely different part.

    Quote Originally Posted by William Antrim View Post
    This topic came up a few months ago and was one of the driving forces behind the latest changes to the rare pool (the TS patch where rares now stack in one giant stack when unressed).
    Not quite right. Unressed parts will stack in L, T, E and MC5E stacks until researched in R#184. Not a single stack.

    Quote Originally Posted by William Antrim View Post
    The devs looked into the mechanics of rare drops and concluded that there is no "rarer" rare part. They all are as rare as one another apparently even though many of us have identified particularly rare parts.
    See above.

    Quote Originally Posted by Torg View Post
    its a bug and will be fixed some time in the future.
    Incorrect. See above.
    Trivaldi
    Neocron Support Team
    N E O C R O N - G A M E . C O M

    »I'm in a glass case of emotion!«

    DOWNLOAD NEOCRONPLAY NEOCRONFACEBOOKTWITTERIRCGET SUPPORTFORUM RULESRULES OF CONDUCT

  5. #5

    Default

    Quote Originally Posted by Trivaldi View Post
    There are no mechanisms or code or algorithms which cause the game to generate certain parts over others. Each time you obtain a rare part, you have the same chance of it being any single part in the rare pool. The resultant part is generated entirely at random so obtaining a lot of certain parts is entirely possible. Each time you obtain a rare part it has the same chance as being the same as the last one you researched, as it does being a completely different part.
    Sorry Triv - I tend to disagree with this - It cannot be random because 2 Grims spawned within 60 seconds of each other drop the same 2 tech parts. There's more to this than randomness. Not sure if its mob rank related or mob spawn related but there are definitely different rare parts within different areas of the game.

    @William - ATP CS drops off Hoverbombers. Well......mine did.

  6. #6
    freedom for neocron! Torg's Avatar
    Join Date
    March 2003
    Location
    wasteland
    Posts
    1,865

    Default

    Quote Originally Posted by Trivaldi View Post
    The resultant part is generated entirely at random so obtaining a lot of certain parts is entirely possible.
    i hear your words, Triv. while i have yet to get to believe what i hear. we have seen "random numbers" generated from spreadsheets, tho. in the past, and elsewhere, of course. depends on the algorithm, right?

  7. #7
    Dream Cast Cursed Shadow's Avatar
    Join Date
    November 2005
    Location
    London, United Kingdom
    Posts
    223

    Default

    Quote Originally Posted by Coffeecup View Post
    Sorry Triv - I tend to disagree with this - It cannot be random because 2 Grims spawned within 60 seconds of each other drop the same 2 tech parts. There's more to this than randomness. Not sure if its mob rank related or mob spawn related but there are definitely different rare parts within different areas of the game.

    @William - ATP CS drops off Hoverbombers. Well......mine did.
    With regard to the part stacking you mention Coffee Cup.

    There IS a mechanism involved, or so I have found. If you kill two monsters of the same type, that spawned from the same location, within a few seconds of one another then their loot tables will be the exact same.
    I have tested this and replicated this over and over.

    Trivaldi, I'd like to believe that there is a complete equality between parts.
    Given the data Brammers provided we can interpret two things in this case:

    People are building certain weapons immediately and therefore they are not kept within the rares database.
    OR
    There is a general trend here.
    See spreadsheet

    https://docs.google.com/spreadsheet/...1E&usp=sharing

    There are two tabs, the first shows the rarity of individual parts.
    The second tab shows you the likelihood from researching any tech part, whether you will get a certain weapon. (I have gone to the extent of working this out for each L, T and E-based rare)*

    *All data is from Techhaven Blue Pages, and is for statistical analysis of the current market, and therefore may not be a true reflection of tech part rarity, but more so interprets current demand.
    Last edited by Cursed Shadow; 16-09-13 at 15:14.

  8. #8
    We are the CityMercs
    Join Date
    September 2002
    Location
    J01, Bunker
    Posts
    371

    Default

    I was waiting for that post CS.
    Of course the rares data is skewed, but only in the relations between Items whose market has crashed (Freeze Stuff, melee, most pistols) to "useful" items like Implants, HC Weapons etc.
    If an Item is built, then the numbers go down evenly over the individual techparts.
    That wouldn´t explain the difference between those individual parts then.

    I´m no statistician, so i don´t know wether the numbers check out. almost 300k rareparts and there are still individual parts that are low, while others seem to be high.

    If we see a a trend clearly reversing in one year, i think we know that it was only noise and there just weren´t enough parts in the pool to give out even numbers.

  9. #9
    Registered User Agent L's Avatar
    Join Date
    September 2002
    Location
    Plaza 1 Medicare
    Posts
    917

    Default

    I see 2 possible explanations:
    1) Parts are dropped really random. Its human mind's fundamental incapability of understanding randomness that plays tricks on us. Same shit as sacrificing animals for different weather and so on.

    2) The drop formula is equal and devs did everything right, but they've used crooked RNG without even knowing it. Or someone did it intentionally to save CPU cycles.
    I've even seen Middle-square method used in a game o_O! This one has bias so huge that 50% chance equaled to percepted "pretty much never" and 51% turned out as "almost always" : ) This way all the formulas seem right in code, but since they're fed with biased input, the outcome is biased as well.

    AAAAnd even if algorithm is good, some regular events like daily server restart and reseeding can still crook it.

    Quote Originally Posted by Trivaldi View Post
    There are no mechanisms or code or algorithms which cause the game to generate certain parts over others.
    There are no *intentional* mechanisms or code : )
    Unless you meant that someone ran a huge test sample and got pretty flat result.


    *disclaimer: by "RNG" I meant "pseudo random number generator".
    Last edited by Agent L; 16-09-13 at 15:21. Reason: disclaimer

  10. #10
    Dream Cast Cursed Shadow's Avatar
    Join Date
    November 2005
    Location
    London, United Kingdom
    Posts
    223

    Default

    Yeah, as I said, the purpose of my spreadsheet is to identify current demand within the market.
    Not to try to prove some kind of rarity.

    I cannot personally believe that there is a completely random ruling to tech parts.
    Firstly, there is an inherent rarity applied. L parts are generally rarer than E or T parts from monsters but there are fewer items that require L parts.
    T parts, from recollection, make up over 60% of all available weapon/item builds.
    So as a result, this skews the fairness already.

    My personal opinion is that there should be no L, T or E distinction, that would guarantee much more fairness in items. There should simply be a distinction between an Unresearched Tech Part and an Unresearched MC5 Part.


    Overall, I was more interested in drilling down into the data, this really doesn't affect me too much.

  11. #11
    Tech Haven Network Brammers's Avatar
    Join Date
    February 2003
    Location
    UK
    Posts
    3,812

    Default

    Quote Originally Posted by Coffeecup View Post
    Sorry Triv - I tend to disagree with this - It cannot be random because 2 Grims spawned within 60 seconds of each other drop the same 2 tech parts. There's more to this than randomness. Not sure if its mob rank related or mob spawn related but there are definitely different rare parts within different areas of the game.
    Sorry you are not fully understanding the word "Random". A random number generator can generate a random number, and it can also generate the same number twice in a row. (Unless it's coded not to generate the same number in succession) Considering there is 376 parts in the normal pool, and 42 parts in the MC5 pool there is a good chance the random number generator will pick the same number between 1 and 376 twice.

    I'll let the stats people quote the numbers, but I though the definition of "random" needs to be clear up.

  12. #12
    Registered User Agent L's Avatar
    Join Date
    September 2002
    Location
    Plaza 1 Medicare
    Posts
    917

    Default

    Quote Originally Posted by Brammers View Post
    Sorry you are not fully understanding the word "Random". A random number generator can generate a random number, and it can also generate the same number twice in a row. (Unless it's coded not to generate the same number in succession) Considering there is 376 parts in the normal pool, and 42 parts in the MC5 pool there is a good chance the random number generator will pick the same number between 1 and 376 twice.

    I'll let the stats people quote the numbers, but I though the definition of "random" needs to be clear up.
    Yep, having drop of 2 parts identical as previous drop should have exactly same chance of happening as any other pair of parts.

    There is this thing Blizzard used, pseudo random distribution. This thing is to make random look more fair to humans. Which is done by making it LESS random : )

  13. #13
    I am BlackMaze SilentEye's Avatar
    Join Date
    July 2003
    Location
    NeoCast HQ
    Posts
    1,732

    Default

    I have just used Random.org to generate 100 times a random number between the numbers 1 and 376.

    Below are the top 3 numbers that came back.

    306, 3 times
    332, 3 times
    244, 2 times

    This doesn't mean that part 306 and part 332 are more frequent

    I wasn't convinced so I did it again but asked it to generate a random number 10.000 times.

    Part 5 returned 43 times while the other numbers were all between 30 and 40 times. Still part 5 is just random
    Last edited by SilentEye; 16-09-13 at 16:11.

  14. #14
    Registered User Agent L's Avatar
    Join Date
    September 2002
    Location
    Plaza 1 Medicare
    Posts
    917

    Default

    Quote Originally Posted by Cursed Shadow View Post
    My personal opinion is that there should be no L, T or E distinction, that would guarantee much more fairness in items. There should simply be a distinction between an Unresearched Tech Part and an Unresearched MC5 Part.
    I agree with you. So far, what a part is was predetermined during drop - so the L-T-E distinction did nothing but vary the research difficulty. I think this variable difficulty feature is worth less than having the parts in one stack.

    Please note if the L-T-E distinction was a RESULT of predetermining the part during drop, there is no unfairness involved.

  15. #15
    Registered User Agent L's Avatar
    Join Date
    September 2002
    Location
    Plaza 1 Medicare
    Posts
    917

    Default

    Quote Originally Posted by SilentEye View Post
    I have just used Random.org to generate 100 times a random number between the numbers 1 and 376.

    Below are the top 3 numbers that came back.

    306, 3 times
    332, 3 times
    244, 2 times

    This doesn't mean that part 306 and part 332 are more frequent

    I wasn't convinced so I did it again but asked it to generate a random number 10.000 times.

    Part 5 returned 43 times while the other numbers were all between 30 and 40 times. Still part 5 is just random
    I though of a sample of 100k at least, but your result are still quite flat with 40:30 ratio. Looking on first few Blue Pages reveals abundance of CS-comp to CS-ATP *consistent* across almost all accounts. OFC, ratios of 30:1 (summed) seems scary, we must keep in mind that many, many CSes have been built in the past. The real drop ratio is probably closer to 130:101 (no idea here how many CSes are in fact built).
    However, 130:101 still means 30% bias, which is unacceptable for any RNG.

Page 1 of 7 12345 ... LastLast

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •