AG RNG and the math isn't adding up.

135

Comments

  • EvisceratorEviscerator Posts: 345 ★★
    Opened my second AG crystal today. Pulled my second skill AG today.
  • Maldroit2Maldroit2 Posts: 698 ★★★
    Improbable things happen. The odds of winning the lottery are very low, yet somebody wins it.
  • This RNG thread about heroes (and/or Class Awakening gems) and the similarly way-too-long RNG thread regarding why Level-Up Solo Event has not appeared after exactly 3 days off, both need to be merged (or at this point just CLOSED).
  • Speedro420Speedro420 Posts: 276 ★★
    Everyone can say what they like about statistics and such but it fairly obvious there is a correlation between what gems you need and the ones you pull. To many people have extra of the ones they don’t have champs for. Makes sense because that would entice you to spend money to get what you need.
  • So there are hundreds of thousands of players of this game, and a minority of people on the low ends (outliers) of the random bell curve post up here that there must be something wrong. Something wrong with assuming then that something must be wrong.
  • HaminHamin Posts: 2,446 ★★★★★
    Y'all do know cumulative pulls from the entire community =/= drop rates, right?

    The drop percentages are not based on all the pulls everyone has. Each pull is an isolated incident.

    The weird thing is the overwhelming anecdotal reports of people getting the same AG multiple times.

    An octillion different pulls by an octillion different people has zero effect on your pulls or the probability of you pulling multiple same class AGs.

    That's not how math works.
  • JChanceH9JChanceH9 Posts: 334 ★★
    I’ve gotten 3 cosmic, 2 tech, 1 mystic, and a skill. No mutant or science yet.
  • DiablosUltimateDiablosUltimate Posts: 1,021 ★★★
    I have been playing board games that include dice a lot and seeing someone roll same number even 10 times in row is not that uncommon, or failing to roll specific number for 100 turns, thus 4 in row is not something that is surprising me
  • DTMelodicMetalDTMelodicMetal Posts: 2,796 ★★★★★
    Thank you OP for starting this thread. Reading DrZola and DNA3000 going back and forth made my day 🍿
  • B_Dizzle_01B_Dizzle_01 Posts: 1,494 ★★★
    The first letter in rng stands for random.
  • ezgoingezgoing Posts: 264
    My take is this. While I believe in RNG, maths and trustworthy companies, there is always the case of companies not being honest despite what they might proclaim themselves to be. Or else scandals like the diesel car emission scams won’t happen.

    So while I won’t trust kabam fully (unlike some in this forum who vouch for them wholeheartedly), but neither would I say they are a scam.

    I won’t hold the mods accountable too. If there is any conspiracy, they are likely not in the loop - plausible deniability.
  • PlayerPlayer Posts: 169
    I’m fairly certain that day of the week and iso inventory contents affect the RNG.
  • GroundedWisdomGroundedWisdom Posts: 24,061 ★★★★★
    Lol. No they don't. The Crystal doesn't stop to analyze anything in your Account. It's an instantaneous process that involves you opening the Crystal, the Crystal accessing the server, and the server generating an outcome from that specific pool. It isn't affected by anything else. It doesn't have time to be.
  • PlayerPlayer Posts: 169

    Lol. No they don't. The Crystal doesn't stop to analyze anything in your Account. It's an instantaneous process that involves you opening the Crystal, the Crystal accessing the server, and the server generating an outcome from that specific pool. It isn't affected by anything else. It doesn't have time to be.

    The game is online at al times which means the server knows what we have. I routinely manipulate my iso inventory to test this and do see correlations. Perhaps is just random, but it certainly appears more frequently than randomness would suggest. Based on my inventory I can often predict the class of the next champ and am >80% right. I routinely pull tech 5* when I want to but still fail to get SL to dupe my SL.

    Try it sometimes. Stick Tier 5 ISO from every class except one and open a 4* or 5* crystal. Check to see if the class pulled matches the one class missing from the ISO inventory (I.e. the 45 ISO’s present for leveling up). Don’t include basic ISO though, it skews the results.
  • GroundedWisdomGroundedWisdom Posts: 24,061 ★★★★★
    edited April 2019
    I think you're looking for connections that aren't there. The outcome is not affected by anything you have. The server is not monitoring your Inventory. It's not affected by the day of the week, your ISO, your Roster, your Cats, anything. It's called Confirmation Bias, and it's just not true. The Crystals are not semi-intelligent entities, changing the odds of the outcome on the server based on what you have. Although, it would make for a very lovely Stephen King Novel.
  • MavRCK_MavRCK_ Posts: 302 ★★
    DNA3000 said:

    Hamin said:

    First, I'm totally aware drop percentages are independent of each other; i.e. - you have a 16.7% chance of rolling a certain class. Rolls aren't cumulative and every single time it's 1 in 6.

    I'm not talking drop percentages. I'm talking about probabilty of rolling the same number multiple times.

    As it stands, I've rolled 5* Cosmic 4 times in a row. If I'm doing my math correctly, that's a .00077% chance (1/6^4).

    This isn't adding up.

    It is unusual, but a statistical analyst would say you have to be cautious about what you're really calculating. First of all, you're calculating the odds of pulling four cosmics in a row *after* pulling four cosmics in a row. But your rolls aren't noteworthy to you because they are cosmics, but because they are all identical - in other words you'd still be here talking about it if they were all skill or all tech. The odds of pulling four identical classes in a row is not one in 6^4, but rather one in 6^3. That's one in 216 instead of one in 1296.

    One in 216 is still rare, but the second question to ask not just what are the odds of one particular person pulling four in a row, but rather what the odds of it happening at all. After all, the odds of winning the lottery are one in millions, but someone eventually wins. If *they* do the calculations after actually winning, it will always seem incredible after the fact. But in fact, the person who actually already won has a 100% chance of winning - because it already happened. Out of all the players, what are the odds of having a "winner" like yourself? If the odds are one in 1296, there should be dozens to hundreds of them. If the odds are one in 216, there ought to be hundreds or thousands of them. The difference is significant in that in a random pool of a couple dozen players, it is possible none of them are active forum participants. But in a random pool of thousands of players, the odds are at least a few are forum participants. And if any of those pull this sequence, they are much more likely to report it.

    This is actually so common of a problem in statistics it plagues even peer reviewed scientific articles, so I'm not surprised it happens to non-scientists when they try to analyze random events in games.

    Now, I've sometimes seen hints of a correlation glitch in some random generated results, but nothing that I could prove. But those were literal back to back events. Crystal openings separated by a wide amount of time can't be correlated by any bug, and if the game was deliberately attempting to generate repeats I think it would be a far more common occurrence. Most players don't report this happening, and I don't see the benefit to a game skewing the dice, but only a little bit.
    It’s for these reasons players should be upset that the number of 5* shards earned is so paltry per month (2-3) while the number of champions being added or growing is 2 per month which is huge which results in a game completely RNG-dependent (or play to win).

    Players should be even more furious Kabam took away the old feature crystal system and put in the current RNG fiesta bs one that exists today.
  • Star_Lord_Star_Lord_ Posts: 523 ★★★
    I do not hold on to awakening gems as I usually have someone to spend it on, so for me, the only AG I had in inventory was a generic. What I have noticed, and it holds true with the Intel AG, is that whatever class I have the most T4C, I would pull that AG. In this case I got Tech as I had 2 T4C tech in my stash...all other tier 4 cats were lower.

    I have found this to be the case with at least the last 4 AGs I have opened.

    just my 2 cents...

    I am curious if anyone who got repeat gems....do you have the most of that cat in inventory?
  • Ch1efsterCh1efster Posts: 463 ★★

    I do not hold on to awakening gems as I usually have someone to spend it on, so for me, the only AG I had in inventory was a generic. What I have noticed, and it holds true with the Intel AG, is that whatever class I have the most T4C, I would pull that AG. In this case I got Tech as I had 2 T4C tech in my stash...all other tier 4 cats were lower.

    I have found this to be the case with at least the last 4 AGs I have opened.

    just my 2 cents...

    I am curious if anyone who got repeat gems....do you have the most of that cat in inventory?

    I'll check this when I get and open my intel AG. I have the most Mutant t4cc right now and a mutant gem would be great to awaken my AA and not have to use my generic on him. Only other AG I have in stash now is Cosmic and debating using that before I open my intel one, but Corvus is the only cosmic unawakened I have that is good and he really doesn't need it.
  • Ch1efsterCh1efster Posts: 463 ★★
    edited April 2019
    Update: I opened my intel AG and guess what? I got the only class one I already had in inventory.....a Cosmic. :frowning:
  • Lvernon15Lvernon15 Posts: 8,350 ★★★★★
    Something is definitely wrong, my last 4 have all been mystic, my 3 before that were all mutant, and I’ve had 1 skill, no other class
  • WOKWOK Posts: 468 ★★

    i am currently re-reading the TOS, which also mentions trade secrets and consumer behavior, as well as "underlying ideas or algorithms" coming into play within the Service. isn't there room to interpret any of that to support that it is at least *possible* that it's not as simple as 1 in 6 chance for any one class or 1 in (x) chance for any one champion, where x is the total number of champions available at whichever applicable star rating in question?

    i say that because I know i have personally seen in my alliance many times the same champion being pulled multiple times in a brief span. or my aforementioned theory about the algorithms attempting to help you save up any one resource. once i finally had a couple t1 alphas saved up I started getting t1alpha fragments much more often than i had previously in my 4 hour free crystals.

    based on my personal stats, i seem more likely to receive units from an arena crystal if i have just spent some units, more likely to receive 10,000 gold from an arena crystal if i have just ranked someone up, and more likely to duplicate a 3star champ from a PHC if i am very close to enough shards for a 4star hero.

    i also know a bit about game theory, decent amount about carl jung's coincidence theory, and am well familiar with cognitive biases particularly Confirmation bias; but knowing all that I still am left wondering about the rng. i'm not angry or holding a pitchfork, i know enough about life to manage my expectations to be realistic, i just enjoy thinking/talking about this kind of stuff. any time i see the phrase "trade secrets" or a company mention "consumer behavior" my brain gets to thinking, thats all.

    I've spoken to people from several alliances over the years and they have spoken with others and we've all come to notice and agree that the improbable "anectodal evidence" seems to definitely be real
  • MMCskippyMMCskippy Posts: 330 ★★
    I'll throw this into this wonderful punch bowl of a thread:

    I play Marvel Puzzle Quest and the only similarities are that it's Marvel based and there is RNG.

    The way the developers over there seem to run the RNG is that champs you're going to need in upcoming content seem to be loaded into the pool at a higher rate than other champs (2* and 3*).

    For example, each Event has a required champ at each level (2*, 3*, 4* and 5*). They don't publish a list of what 2* champs will be required each event, but you can tell because all of your 2* pulls seem to hit around the same couple of champs and this pool of champs changes throughout the year. I can go for a while without seeing a specific 2* or 3* champ and then... there's a whole bunch around and, what do you know, they're essential characters in the next event.

    This seems to happen to a certain extent with 3* champs as well.

    4* and 5* champs seem to resemble true RNG for who you pull, but that's only because the sample size is small.

    I guess what I'm trying to say is that there SEEMS to be code that exists that SEEMS to be able to weight RNG toward certain outcomes. I only say "seems" because only the creator really knows.
  • LegendsendLegendsend Posts: 92
    I would have to agree that it seems like you tend to get the ones that you already have... I'm not sure why exactly... but I'm sitting on a couple science AGs that who knows when I will get to use... At the same time I literally have 0 duped mystic champions...0... my first 5 star out of everyone was magik and still no mystic is duped... Mind you, I am deep enough in the game to be cavalier... I just chalk it up to bad luck

  • henrieshenries Posts: 55
    Have rolled the last 2 times on 5* skill ag
  • DNA3000DNA3000 Posts: 10,701 Guardian
    WOK said:

    I've spoken to people from several alliances over the years and they have spoken with others and we've all come to notice and agree that the improbable "anectodal evidence" seems to definitely be real

    Question. Scientific studies have been performed which demonstrate conclusively that people cannot properly distinguish random events from correlated ones. They overwhelmingly see non-randomness in random events, and predict randomness when there is overwhelmingly high amounts of non-random but well distributed events. In other words, they see any clumpiness as evidence of non-randomness, and "nice" well distributed results as being random when randomness cannot produce such non-clumpy or non-bursty results.

    The question is: do you believe these studies prove all of your anecdotes to be suspect? In other words, do you believe all the players you talk to because they all say the same thing, or do you believe all the cognitive scientists in the world, because they all say the same thing?
  • MavRCK_MavRCK_ Posts: 302 ★★
    edited April 2019
    Another thread filled with comments and arguments filled by the usual suspects... so reminds me of Macbeth and its famous soliloquy.

    Do you ‘trust’ Kabam’s claims? Yes or No.

    If yes, end of argument.

    If not, ask for evidence that supports the claim. If no evidence can be found to support the claims, where do you stand as a result?
  • Dtl7714Dtl7714 Posts: 425 ★★★
    edited April 2019
    Given the way most people pull the one they dont need which is already in there items it would stand to reason that there is an algorithm in the game code that increases the odds of pulling one that is already in your items. By making you pull one you eont need it increases odds of people spending money to get more
Sign In or Register to comment.