The myth of RNG

2

Comments

  • nigelregal
    nigelregal Posts: 184 Tile Toppler
    I have had same in LR. Pulled 10 bagman in a row
  • Daredevil217
    Daredevil217 Posts: 3,967 Chairperson of the Boards
    edited March 2018
    HoundofShadow said:
    Anyone looking forward to Bagman's PvP? Could it be this April?
    Absolutely.  I have a max-champed Baggy with a max-covered dupe ready to go.  I'm not selling the original either.  Worked too hard for that trophy.  Going to eat the 75k iso to rechamp my dupe as soon as I get a 14th cover.  If Bags to Riches rolls around it will be way more than 14 covers.
  • Kahmon
    Kahmon Posts: 625 Critical Contributor
    I've often gotten 4-5 of the exact same 1* just off seed teams, and occasionally 3-4 of a 2*, but less frequently.

    The most memorable was when it was on the 1st, 4th, 7th, and 10th fights. I do my seeds clockwise starting at the top, so each one came when I got back to the top node.
  • Alx85
    Alx85 Posts: 30 Just Dropped In
    Not sure if it counts but I got pull happy with my legnedaries last night and got 5x 5* ghost rider, 3 in a row that were green, then 2 in a row that were black. I was hoping for Jessy J.. felt very lucky, but based on my previous 5* pulls it was highly unusual, maybe a ‘happy’ glitch. Is he any good? GR was one of the ones I didn’t have so no idea..
  • Yepyep
    Yepyep Posts: 954 Critical Contributor
    GrimSkald said:
    This isn't so much a glitch as the simple fact that computers are actually pretty bad at generating numbers that are well-and-truly random.  In general, the random number a computer generates are very "streaky."  It's not bad in the long run, but it creates some really weird runs.
    Really, they can't do it at all within their own constructs. Importing observational data (lightning strikes, super-granular clock arbitrage, etc.) helps, but an architecture of binary switches doesn't lend itself to unforeseeable products. Even in a biological, seemingly non-binary machine like a human being, talking in tongues requires divine intervention.

    There is no such thing as a "random number generator" -- at best we'll get an "unexpected number generator."
  • Yepyep
    Yepyep Posts: 954 Critical Contributor
    Hadronic said:
    GrimSkald said:
    This isn't so much a glitch as the simple fact that computers are actually pretty bad at generating numbers that are well-and-truly random.  In general, the random number a computer generates are very "streaky."  It's not bad in the long run, but it creates some really weird runs.
    [1] This is just wrong, [2] computers are in fact much better at creating random numbers than humans are at perceiving them. [3] Truly random events are streaky and in 2 dimensions tend to cluster.

    http://www.dailymail.co.uk/home/moslive/article-1334712/Humans-concept-randomness-hard-understand.html

    I love your comment. And I am no mathematician but a couple of further thoughts, all of which I am making up on the fly:

    If true, [2] does not make [1] correct. Human perception as to a number's randomness (a meaningless concept, really, as a determination of randomness or correlation requires something greater than a set of one -- certainly at least four, I would think...) does not affect randomness (other than in its effect on the definition of "random" which certainly matter lol. What is "random," after all, if one has an infinite capacity to perceive relationships between numbers?). But humans are fearful of complexity and yearn for predictability and are consequently horrible arbiters of randomness...

    As to [3], "streakiness" and "clustering" are the products of (1) human perception and our powerful (and pretty magnificent) drive to organize data and find correlations amongst them and (2) limited data sets (both in quantity and in upper/lower bounds).

    So, in the case of this game we have (1) very limited data sets (2) interpreted by minds uncomfortable with complexity and thus trained to find correlations who have (3) a strong emotional impetus to confirmation bias. However good or not good the RNG in question is, we WILL find streaks and patterns in these limited data.

    But there is no such thing as a RNG being applied here. Relatively unpredictable? Yes. Random? No.
  • carrion_pigeons
    carrion_pigeons Posts: 942 Critical Contributor
    I thought it had been established at some point that the mechanism for "randomizing" covers worked more like a roll of lottery tickets than like a random number generator at all, and the reason for so many strange runs of identical covers is because of glitches that cause  the "roll" to not tick to the next "ticket".

    I swear there was a significant problem once upon a time with 5*s where people were all pulling the exact same one for like an hour? Does this not ring a bell to anyone else? 
  • kk3thess
    kk3thess Posts: 202 Tile Toppler
    Opened 20 Standard Tokens. 12 were Yellow Storm. 8 of them in a row. That isn't right, is it?
  • HoundofShadow
    HoundofShadow Posts: 8,004 Chairperson of the Boards
    Jamesh's explanation is plausible if you can understand what he is saying. Your data is residing somewhere and there has to be some kind of mechanisms to protect all those data. You have things like RAID, replication, backup etc put in place to protect your data. Somehow, there is some kind of mistiming/syncing issue that caused this kind of phenomena. It's good for the players if they are opening LT or HT. If you read some of the posts in best moments of the day thread, you see some players getting incredible covers that you wish you would have. But Bagman covers do comes in streak...I hope it's not confirmation bias.  :D
  • madoctor
    madoctor Posts: 292 Mover and Shaker
    edited March 2018
    I have probably every single one of you beat. In the LR from two weeks ago, I got 15 consecutive 2mags red covers from a single round of LR.

    You read it right, 15...all red 2mags... consecutively...in one LR round.

    EDIT - I have gotten very few 2* covers in LR since then though. Maybe 4-5 in total.
  • Crowl
    Crowl Posts: 1,580 Chairperson of the Boards
    GrimSkald said:
    This isn't so much a glitch as the simple fact that computers are actually pretty bad at generating numbers that are well-and-truly random.  In general, the random number a computer generates are very "streaky."  It's not bad in the long run, but it creates some really weird runs.
    Actually, a truly random sequence should also include streaks even if such a thing doesn't feel random to us because random does not imply evenly spaced, so while a random number generator that eliminated these streaks would seem more random to us, it would actually be doing a worse job.
  • Hadronic
    Hadronic Posts: 338 Mover and Shaker
    Yepyep said:
    GrimSkald said:
    This isn't so much a glitch as the simple fact that computers are actually pretty bad at generating numbers that are well-and-truly random.  In general, the random number a computer generates are very "streaky."  It's not bad in the long run, but it creates some really weird runs.
    Really, they can't do it at all within their own constructs. Importing observational data (lightning strikes, super-granular clock arbitrage, etc.) helps, but an architecture of binary switches doesn't lend itself to unforeseeable products. Even in a biological, seemingly non-binary machine like a human being, talking in tongues requires divine intervention.

    There is no such thing as a "random number generator" -- at best we'll get an "unexpected number generator."
    You just defined what is commonly known as a Pseudo Random Number Generator (PRNG).
    For the record True Random Number Generators (TRNG) exist and are commonly used in cryptography.

    https://en.wikipedia.org/wiki/Hardware_random_number_generator
  • veny
    veny Posts: 834 Critical Contributor
    RNG? Where? https://i.imgur.com/fl4i4FQ.jpg

  • Momo988
    Momo988 Posts: 15 Just Dropped In
    You can also use lava lamps to generate random numbers. 

    https://www.popularmechanics.com/technology/security/news/a28921/lava-lamp-security-cloudflare/
  • Yepyep
    Yepyep Posts: 954 Critical Contributor
    Momo988 said:
    You can also use lava lamps to generate random numbers. 

    https://www.popularmechanics.com/technology/security/news/a28921/lava-lamp-security-cloudflare/
    This is just so very fantastic... Thanks for the link!
  • jamesh
    jamesh Posts: 1,600 Chairperson of the Boards
    Yepyep said:
    Momo988 said:
    You can also use lava lamps to generate random numbers. 

    https://www.popularmechanics.com/technology/security/news/a28921/lava-lamp-security-cloudflare/
    This is just so very fantastic... Thanks for the link!
    It's basically a recreation of something Silicon Graphics did back in the 90s (Lavarand).

    Note that the problem with most hardware random number generators is that there is a limit to how fast they can generate entropy.  If you are taking pictures of a lava lamp once a minute, it might be difficult/impossible to predict what one image will look like based on the last image.  If you take pictures 60 times a second, it will be easier to predict the next image: you're not suddenly producing 3600 times more random data.

    So most of these systems generally use the hardware RNG to periodically reseed a PRNG.  That's how SGI's Lavarand worked, and presumably how Cloudflare's one does too.  The PRNG can produce random looking data at a faster rate, and if there are any bugs in the PRNG that make the output predictable, they will only work as far as the next reseeding.
  • The rockett
    The rockett Posts: 2,016 Chairperson of the Boards
    A few months back in LRs, I only used the top node and got 12 Hawkeye Purple straight.  Weirdest thing I ever saw.  
  • DAZ0273
    DAZ0273 Posts: 10,286 Chairperson of the Boards
    I pulled a blue Prof X from an heroic, got a bonus hero America Chavez green. Opened an LT straight afterwards, exact same result. I had to stop there as now Prof needs champing...
  • Bcorm
    Bcorm Posts: 36 Just Dropped In
    The main issue is true RNG's like the one Cloudlare uses are extremely intricate. Most RNG's just start with a few seed numbers run through an algorithm to get the result. The second problem is human nature. People instinctively have a need to find order in everything, whether it exists or not, therefore they kind of invent patterns where there aren't any. True chaos or randomness is something our subconscious isn't hardwired to comprehend.
     I used to work at a casino and the roulette wheel had a LED display showing the last 20 spins. The regular guests constantly said they knew the pattern of the spins even though there really isn't one. They wanted to see one so they did. Roulette is a game invented by casinos with the worst odds of any game they have. Each spin is truly random, any pattern or cluster of similar results is literally a freak occurance. The odds of ten of the same number in a row are the same as ten different numbers. 
  • Michaelcles
    Michaelcles Posts: 100 Tile Toppler
    RNG mechanic in games is terrible.  If you do something you should get the reward.

    Would you want to be paid a random amount each week at your job?

    if this game was based on a weaker property than Marvel, it would have been extinct years ago.

Categories