Is stats pulling rigged?

  • I think that is why he wants you to list the item you are pulling. Without that named item, looking up in the database random items is effort intensive. Also, hey, you can do it yourself.

  • Please don't take the following as a request which has been the whole purpose of this topic.


    Let's say i had bad luck with stats in "Devourer Gloves of the Dawn" and want to be sure it was only bad luck, not intentionally skewed distribution. Game team can make tests and give their verdict. However Korrin wasn't enthusiastic about testing all important items and I can imagine him saying "we did enough, come back next decade after you pull 1000 PoM items and don't like something". This is not a joke btw but a reasonable assumption about "we won't redo all tests" statement.

    So I have a dilemma. Either ask to test a minor item or ignore it and ask to test something vital many players need.

    If these are the only two options then I would prefer the latter. But if Korin meant that all "vital" distributions have been tested then I can safely go back to the original issue.


    Please help me here. If testing the original EOJ item in question doesn't stop the whole thing right there then we can go there.

  • Well, I asked you multiple times for the item id* and for a definition of what should be tested.


    Unless you are cooperating with me I cannot help you.


    So please follow the steps I've described in my previous posts in this thread.


    *If you don't know how to look up an item id, please post the Romwelten url of the item.

  • For what it's worth I've done the manual round ages ago before I even knew where to look in the data.fdb for the chances - and the more I pulled the closer all possible stats on a given item came to the internal values from the game files. Ever after when I had to pull a lot of stats for whatever reason that didn't change.


    Of course, streaks do happen - that's simply to be expected. Law of large numbers comes to mind.



    And what "intentionally skewed distribution" ?

    If RW didnt want the chances as they were, they could adjust them...

    It is RWs game after all, and if they decide e.g. a Str/Patk should only have a 10% chance instead of a 25% chance then all they have to do is change the number in the freaking loottable from 25000 to 10000 and be done with it.


    Nobody except RW can read the game data anyway, as we all know .... ;)


    Since none of these chance values are actually displayed anywhere ingame and all addons and external sites use the same datamined values (IP, IP2, DungeonLoots, the romwelten Database, the English wiki maintained by Bludwyng as far as it has values instead of text descriptions ) it's all you have.


    -------

    Since it came up again despite that I had it explained back when it was fresh:


    Watermelons worked differently because they were not a direct drop but from a table - the table was already guaranteed to be used

    (so droprate% "increased" that from guaranteed to .... well, guaranteed)

    but anything a layer below is not affected anymore because it's a not a drop.


    example:

    Monster >              Watermelon1 5% | Watermelon2 10% => +400% inc. Droprate =>                Melon1 25% | Melon2 50%

    Monster > Table 100% > Watermelon1 5% | Watermelon2 10% => +400% inc. Droprate => Table 500%* >  Melon1 05% | Melon2 10%

    * (no, >100% does not mean you get more than 1 item)

    Line1: The melons are the drop.

    Line2: The table is the drop.



    [This user speaks English on a near native level.]

    [This user speaks German on a native level.]

    [This user can curse in a variety of languages.]


    In the beginning the Universe was created.

    This has made a lot of people very angry and been widely regarded as a bad move.



    Trust me, I'm an engineer.....with epic skill and epic gear

  • Devourer Gloves of the Dawn, possible Stats

    So, what you are asking is for a GM to check the probabilities by making multiple pulls to see if the table matches reality. I guess they could do that if they wanted, but why should they? Do you think RuneWaker programmed some extra actions on top of the loot table? Nothing stopping them I suppose... but why would they?
    100% [drop ID: 727264]
    25% [Ability of Dawn]
    25% [Power of Dawn]
    25% [Fire of Dawn]
    25% [Attack of Dawn]
  • According to rom-welten the item is is 236293. I'm specifically interested whether stat 514559 really drops 1/4 of the time.


    PS. I reserve the right for "i did say so" comment later.

  • I went back to the Justice vendor another day and after 11 tries I got a Protection of the Panetheon. Odds don't seem very good, but its seem about the same as it has been over previous zones before Chassizz. It's probably like a chance to get that stat from 0 to 3 times after 50 tries.

  • Brilliant. Just brilliant. All we need is id. And nothing happens when we get it.

    If there is anything that can help you establish fairness of pulls please let me know how.

  • Well, imagine if like nobody ever pulls that item except you, how can we get enough data for a statistic? :D


    I've used the Celestial Spire Token of Intelligence http://www.rom-welten.de/database/view.php?id=236148 to demonstrate the https://en.wikipedia.org/wiki/Law_of_large_numbers.

    Each of the 6 stats has the same chance.


    Expected distribution:

    Choice of the Pantheon Control of the Pantheon Pacification of the Pantheon] Promise of the Pantheon Purity of the Pantheon Keeper of the Pantheon
    16.67% 16.666% 16.666% 16.666% 16.666% 16.666%


    Distribution after checking 4 servers:


    Choice of the Pantheon Control of the Pantheon Pacification of the Pantheon] Promise of the Pantheon Purity of the Pantheon Keeper of the Pantheon
    16% 15.6% 15.1% 18% 15.6% 17%



    The more servers I check the more do the observed value converge against the expected numbers.



    grey-seal-waving-goodbye-hello.jpg?w=800


    Anyway, I've said enough and I'm not going to repeat myself all over. The stat pulling mechanism has been tested enough and bad luck is not a bug. Therefore


    > Closed <

  • Korin

    Closed the thread.