Drop rates and dupes

2

Comments

  • Matthew
    Matthew Posts: 605 Critical Contributor
    How would I go about recording my own opening on an iPhone? I'm genuinely interested in sharing my own data, though I've only got 60 KLD/AER packs, 2 Fat Packs, and 1 Big Box, so it wouldn't be as huge as some other collections.
  • THEMAGICkMAN
    THEMAGICkMAN Posts: 697 Critical Contributor
    There are several screen recorder apps. Just search the store and I'm sure you'll find one. I use Game Screen Recorder for android, dunno if its available for iOS though @Matthew
  • shteev
    shteev Posts: 2,031 Chairperson of the Boards
    Ohboy said:
    I don't know why anyone thinks it would be in their best interest to step in and clarify since it will just stir up a tinykitty storm.
    Clarify? When have D3/Hibernum ever said anything at all about drop rates and dupes that can be clarified?
  • Ohboy
    Ohboy Posts: 1,766 Chairperson of the Boards
    shteev said:
    Ohboy said:
    I don't know why anyone thinks it would be in their best interest to step in and clarify since it will just stir up a tinykitty storm.
    Clarify? When have D3/Hibernum ever said anything at all about drop rates and dupes that can be clarified?

    My phrasing might be misleading. I'm saying that there is no value in then stepping in to clarify the drop rates, because some people will just say they are lying, and stir up even more arguments. 

    For example, you yourself have already taken a similar line, assuming that they don't know that it's skewed due to faulty coding. How would they coming out to say that it's random help you? It wouldn't. 

    You need to prove that the reward drops are skewed. To that end, I'm suggesting you consolidate actual data, not opinions as part of your "mounting evidence". 
  • julianus
    julianus Posts: 188 Tile Toppler
    Ohboy said:
    People have already stated that they would choose to believe they are lying if they do officially state that it's completely random. 

    I don't know why anyone thinks it would be in their best interest to step in and clarify since it will just stir up a tinykitty storm. 

    Onus is on the accuser to supply evidence, of which none of you have supplied any(some of that mounting evidence would be helpful) . Majincob who famously states he has not converted his cards at all would be a great data point if he's willing to share his collection details. You guys can start from there. 

    While you're on the whole being scientific kick and all. 

    I don't think it's entirely accurate to say there's no evidence. I, for example, posted some data in this thread: http://forums.d3go.com/discussion/63330/more-fun-with-stats-and-duplicates#latest

    One does, admittedly, have to accept:

    1. That my description of my data is accurate
    2. That the algorithm behind the Rand() formula in Excel is a reasonable approximation of genuinely random draws (or alternately, accept the same thing for the program Volrak used for his much larger simulation run).

    And of course, nothing in that thread is concrete proof of anything. But it's strongly suggestive, and is an attempt to look at the subject in an objective and evidence-driven manner.
  • Ohboy
    Ohboy Posts: 1,766 Chairperson of the Boards
    edited April 2017
    julianus said:
    Ohboy said:
    People have already stated that they would choose to believe they are lying if they do officially state that it's completely random. 

    I don't know why anyone thinks it would be in their best interest to step in and clarify since it will just stir up a tinykitty storm. 

    Onus is on the accuser to supply evidence, of which none of you have supplied any(some of that mounting evidence would be helpful) . Majincob who famously states he has not converted his cards at all would be a great data point if he's willing to share his collection details. You guys can start from there. 

    While you're on the whole being scientific kick and all. 

    I don't think it's entirely accurate to say there's no evidence. I, for example, posted some data in this thread: http://forums.d3go.com/discussion/63330/more-fun-with-stats-and-duplicates#latest

    One does, admittedly, have to accept:

    1. That my description of my data is accurate
    2. That the algorithm behind the Rand() formula in Excel is a reasonable approximation of genuinely random draws (or alternately, accept the same thing for the program Volrak used for his much larger simulation run).

    And of course, nothing in that thread is concrete proof of anything. But it's strongly suggestive, and is an attempt to look at the subject in an objective and evidence-driven manner.


    Would you be willing to concede that there isn't nearly enough evidence yet to call it "mounting evidence"? 

    That's why I'm asking people to share their unconverted sets. Yours alone is not enough. Not by a long shot.

    Your eagerness to share your data by the way is the exception not the rule when it comes to these topics. An overwhelming amount of the time, people come in nothing but feelings and call it evidence. 
  • julianus
    julianus Posts: 188 Tile Toppler
    Ohboy said:

    Would you be willing to concede that there isn't nearly enough evidence yet to call it "mounting evidence"? 

    That's why I'm asking people to share their unconverted sets. Yours alone is not enough. Not by a long shot.

    Your eagerness to share your data by the way is the exception not the rule when it comes to these topics. An overwhelming amount of the time, people come in nothing but feelings and call it evidence. 
    Absolutely. It's a small and isolated sample and not at all comprehensive - but it's something.

    I don't consider myself part of any group pushing an interpretation, for the record. I did this mostly because I like dabbling in stats and was curious what that particular analysis might suggest.

    My guess is that the draws are weighted, but I'd also assume that a priori simply on a basis of business rationale. If the company has never said otherwise (I don't know whether they have, I'm just assuming from other statements I've seen from long-term players), then it makes sense. Lots of computer-based games of chance have weighted odds - electronic slot machines, for example. I don't think anyone plays those thinking that that they have an even chance at every outcome.

  • Ohboy
    Ohboy Posts: 1,766 Chairperson of the Boards
    edited April 2017
    Slot machines are an interesting example. 

    The casino declares that x% of people wins $A, y% of people wins $B for example. If we draw a connection between that and mtgpq, I would say it's more a correlation between rarities. The casino doesn't target people within the x group and weight against them further, not just because it's against the law, but because there's no incentive to do so. All they have to do is tweak A to get the same result they want. 

    Same thing in mtgpq. All they have to do is reduce drop rates to get the results they want. Or change the rares they want to make "more rare" into mythics. There's no need to do all this fancy shady stuff people are accusing them of doing.

    This accusation is ludicrous because it 1) was started by someone who met someone else online who told them they saw it in code on client side(and then conveniently stopped playing when questioned further), and 2) makes zero sense for the devs to do because it requires extra work and subterfuge when the existing infrastructure already allows them to tweak the drop rates the way they want. 

    [Moderator edit: removed potentially controversial comparison. -Dayv]
  • shteev
    shteev Posts: 2,031 Chairperson of the Boards
    Unfortunately, as you can see from my avatar, I have been 'jailed' for posting off topic, and am therefore now vastly restricted in the number of times I may post during a week. As such, I will be unable to take further part in this discussion until I have thrown a double.
  • AngelForge
    AngelForge Posts: 325 Mover and Shaker
    I read in an science article that, I think it was in the second half of last year, they made a step forward to "true" RNG for computers. But that was in a fancy lab.

    What I want to say is, there is no real RNG.
    And if that RNG they use is someone related to a number that is fix and related to you, like the UID and/or Device ID, then  you will have repeating pattern.

    Or am I wrong there?
  • Ohboy
    Ohboy Posts: 1,766 Chairperson of the Boards
    AngelForge said:
    I read in an science article that, I think it was in the second half of last year, they made a step forward to "true" RNG for computers. But that was in a fancy lab.

    What I want to say is, there is no real RNG.
    And if that RNG they use is someone related to a number that is fix and related to you, like the UID and/or Device ID, then  you will have repeating pattern.

    Or am I wrong there?

    You're not wrong. Pseudo random is what we call it, and people like to pounce on the pseudo part to mean it cannot generate random numbers.  The simple rand() function people use for generating random numbers works as such:

    There is a pre-generated list of numbers which is very long and "random". Now, if we only want to generate one list of random numbers, then just reading off from the top of the list would be fine as long as you weren't concerned about security (this list is available to the public, and people can predict what your next "random number" will be if they know you're just reading from the top of the list.)

    What most people will thus do is to start reading the numbers at a random part of the list. This is called the seed. A common way to do it is to convert the current date+time into a number and use it to point to the xth number in the list to start reading. Note that this has the same problem outlined above... Anyone who knows when you're doing this can reverse engineer and predict what comes next. This is in fact the way one of the biggest online poker cheating was done. But for generating a list of random numbers for personal use, it's perfectly fine. 

    So the random here is random, but predictable if we know the seed. That makes it not really random, because by definition you cannot predict a random number. However, for what we want to do here(generate a random card), it's perfectly fine because we don't care if someone knows what card is going to be generated(OK we care, but it's really not a big deal unless someone creates a hacked apk to abuse it). That's why it's pseudo random. 

    Shteev's hypothesis is that they are reusing the seed. You can see the problem if that's true. When you put in the same seed, you always start reading the list from the same place and hence always get the same number. Mistake happens more often than you think, but it's usually in year 1 computing when people first start writing programs that use random numbers. It is a simple mistake, and I don't think I've ever heard of anyone who does it twice. Because you only need to seed once and seeding again and again is actually extra work, it's not hard to never repeat that mistake. 

    Tldr : It's not completely random because output can be predicted if seed is known, but for just generating (a small-ish set) of numbers, it's OK to assume it's random.
  • shteev
    shteev Posts: 2,031 Chairperson of the Boards
    Ohboy said:
    Shteev's hypothesis is that they are reusing the seed.
    If that were indeed my hypothesis, then why are you encouraging people to post their collections without any time collection data?
  • Volrak
    Volrak Posts: 732 Critical Contributor
    shteev said:
    If that were indeed my hypothesis, then why are you encouraging people to post their collections without any time collection data?
    Maybe because large-scale collection data, with or without timestamps, is really really useful.

    Speaking for myself, I'm currently far more interested in making a determination between (A) "uniform random" and (B) "not uniform random" than I am between (C) "pseudo-random with time-based bias" and (D) "random, or non-random in some other way".

    Although whether C is indeed your hypothesis remains somewhat murky, I feel obliged to further state that even if we could somehow establish what your hypothesis was, my own personal priorities as stated above would remain unchanged.

  • Ohboy
    Ohboy Posts: 1,766 Chairperson of the Boards
    shteev said:
    Ohboy said:
    Shteev's hypothesis is that they are reusing the seed.
    If that were indeed my hypothesis, then why are you encouraging people to post their collections without any time collection data?


    Because they are unlikely to have them? And I'm pretty sure even you are unlikely to have yours. If consolidated data shows there's a certain skew, then we can narrow it down moving forward. Why force answers we know aren't there yet? 
  • Ohboy
    Ohboy Posts: 1,766 Chairperson of the Boards
    Just so there can be no doubt (shteev wants to be scientific!) , this is Shteev's hypothesis
    Shteev said:

    But, actually, I don't really suspect that the drops are calculated specifically to annoy us. What I think is more likely is that Hibernum keep re-seeding the RNG with the system clock in an ill fated attempt to make it 'more random'. I worked at a place where development did that once. You would not BELIEVE how difficult it was to persuade them that their random number generator was demonstrably not random.


  • shteev
    shteev Posts: 2,031 Chairperson of the Boards
    Ohboy said:
    Just so there can be no doubt (shteev wants to be scientific!) , this is Shteev's hypothesis
    Shteev said:

    But, actually, I don't really suspect that the drops are calculated specifically to annoy us. What I think is more likely is that Hibernum keep re-seeding the RNG with the system clock in an ill fated attempt to make it 'more random'. I worked at a place where development did that once. You would not BELIEVE how difficult it was to persuade them that their random number generator was demonstrably not random.
    Well, that's certainly something I posted in another thread.

    As a result of that thread, I have received some very interesting messages from a variety of sources, and so that's no longer my only working hypothesis.

    Please, continue to do science! Science is good. Information is good! Sharing information is excellent! My own ability to share on this particular forum has sadly been hamstringed as a result of posting the following, highly damaging joke:
    Shteev said:
    "If you get 5,000 likes, you can buy a LOL pack, which will give you a 3% chance of opening a 25 LOL badge."
    but I shall continue to be an active member of the community and do my science elsewhere!


    Also: a bit rude, speaking for me, I thought!
  • Ohboy
    Ohboy Posts: 1,766 Chairperson of the Boards
    Thought you were muted and did you a solid bro.
  • shteev
    shteev Posts: 2,031 Chairperson of the Boards
    Ohboy said:
    Thought you were muted and did you a solid bro.
    The more time goes on, the more I think that the dev's suspision that you and me are the same person is probably correct...
  • Gunmix25
    Gunmix25 Posts: 1,442 Chairperson of the Boards
    shteev said:
    I'll just pin this here.



    Any time you want to chip in, @Brigby , you go right ahead.
    And the second shooter behind the grassy knoll was..... 

    Conspiracy theories are abound with each and every game that has random packs. WotC MTG was accused of fixing packs for years since it's inception.  Pokemon was the same. Mage knight was huge in the early millennium and their figure packs for Rares and dupes thereof caused an out cry on the forums.  If I have this right, What he is essentially asking is that he wants 100% of a useful pack rather than one with any dupes. It doesn't happen and whether or not that the game is electronically based,  to even ask that they manipulate the packs "prior" (despite the insinuation that they are now to provide dupes) is opening a can of worms.  Especially on the power  level.... imagine top players with tinykitty near everything in their collection topping off with literal guaranteed masterpieces because they can afford to do so. My two cents on the matter

Categories