Five Star Draw Pattern

2

Comments

  • Bowgentle
    Bowgentle Posts: 7,926 Chairperson of the Boards

    From a technical point of view, I would guess that the game saves the seed of a random number generator for each person. This way the sequence of random numbers that are generated for that player are always the same.

    If the game saves only one seed, the outcome could depend on the order of tokens you open (e.g. the pre-determined numbers are, say, 123 and 456. If you open a standard token first and a legendary second, you will get whatever you get with random number 123 from the standard token and whatever you get with number 456 from a legendary token. If you open the other way round, you will get whatever you get with numer 123 from a legendary token and whatever you get with number 456 from a standard token).

    If the game saves a seed for each token type, then the order would not matter.


    It saves for each token type.
  • broll
    broll Posts: 4,732 Chairperson of the Boards
    DAZ0273 said:

    So just to get my head around this – the acquisition rate at which you are able to receive 5* characters is already set per player before making draws but the 5* cover received is still completely random?





    From my understanding what this means is that now instead of getting a dice roll every time you open a pack, the game rolls a whole bunch of dice all at once and queues them up.  I don't see a huge issue here (unless people are able to figure out what their rolls are and exploit it the opposite way).  Ultimately it's still random but the rules for how it's randomized were tweak to prevent exploits.  That sounds like the right thing to do to me.

    In theory they could also use this to better enforce the advertised odds are received (it probably won't but it could).
  • DAZ0273
    DAZ0273 Posts: 10,292 Chairperson of the Boards
    broll said:
    DAZ0273 said:

    So just to get my head around this – the acquisition rate at which you are able to receive 5* characters is already set per player before making draws but the 5* cover received is still completely random?





    From my understanding what this means is that now instead of getting a dice roll every time you open a pack, the game rolls a whole bunch of dice all at once and queues them up.  I don't see a huge issue here (unless people are able to figure out what their rolls are and exploit it the opposite way).  Ultimately it's still random but the rules for how it's randomized were tweak to prevent exploits.  That sounds like the right thing to do to me.

    In theory they could also use this to better enforce the advertised odds are received (it probably won't but it could).
    Right - so RNG is now a bit like Ozymandius from Watchmen.

    "Do it? Dan, I'm not a Republic Serial villain. Do you seriously think I'd explain my master-stroke if there remained the slightest chance of you affecting its outcome? I did it thirty-five minutes ago."
  • Straycat
    Straycat Posts: 963 Critical Contributor
    When, where, or how did this get confirmed?
  • PenniesForEveryone
    PenniesForEveryone Posts: 294 Mover and Shaker
    Straycat said:
    When, where, or how did this get confirmed?
    It was confirmed by exploiters when they were unable to change the outcome of their pulls.
  • LifeofAgony
    LifeofAgony Posts: 690 Critical Contributor
    Back around March/April of last year was when the last exploit was fixed.  They may have never spelled it out here but on other non-official medium (surprise!) there’s been confirmation of this change.
  • Jarvind
    Jarvind Posts: 1,684 Chairperson of the Boards
    edited January 2018
    jredd said:
    jredd said:
    Bowgentle said:
    It's fixed per player, but completely random between players.
    No set pattern across the whole population.
    wait a sec...

    so you're saying if i pull 50 pulls, it's already determined that i'll pull a 5* at say pull numbers 5, 15, 17, 45?

    but if you made 50 pulls at the same time you'd pull a 5* at pull 11, 17, 27, 28, 35, 41, 44 and 49?
    That's how I read it. 
    so, hypothetically speaking, if your luck was poor, you could pull the same amount as player b and always come up with fewer 5* pulls just cause?
    That's pretty much what "random" means, yeah.

    I do assume that your predetermined pulls must change when a new character is added to the pool you're pulling from. I've pulled a bunch of Nico since she was added and I definitely haven't pulled 50+ LTs, or however many pulls get determined ahead of time.
  • ThaRoadWarrior
    ThaRoadWarrior Posts: 9,456 Chairperson of the Boards
    Does it determine your character actually, or just if you're getting a 4* or 5*?
  • Calnexin
    Calnexin Posts: 1,078 Chairperson of the Boards
    Does it determine your character actually, or just if you're getting a 4* or 5*?
    It would have to be the latter.  The character pool constantly changes, so unless they've planned every character from here until the death of the game it would have to randomize the cover among the currently available pool.
  • Bowgentle
    Bowgentle Posts: 7,926 Chairperson of the Boards
    Calnexin said:
    Does it determine your character actually, or just if you're getting a 4* or 5*?
    It would have to be the latter.  The character pool constantly changes, so unless they've planned every character from here until the death of the game it would have to randomize the cover among the currently available pool.
    It determines the character in Latest.
    In classics the chars change when a new 5 enters.
    In Latest, the character leaving is replaced.

    So a Gambit will change to a Ghost Rider, while a Thor and an Archangel will stay the same.
  • ThaRoadWarrior
    ThaRoadWarrior Posts: 9,456 Chairperson of the Boards
    And it does the same for 4*s? 
  • Bowgentle
    Bowgentle Posts: 7,926 Chairperson of the Boards
    I have no idea, I don't care enough about the 4s I draw to keep track.
  • jredd
    jredd Posts: 1,387 Chairperson of the Boards
    Jarvind said:
    jredd said:
    jredd said:
    Bowgentle said:
    It's fixed per player, but completely random between players.
    No set pattern across the whole population.
    wait a sec...

    so you're saying if i pull 50 pulls, it's already determined that i'll pull a 5* at say pull numbers 5, 15, 17, 45?

    but if you made 50 pulls at the same time you'd pull a 5* at pull 11, 17, 27, 28, 35, 41, 44 and 49?
    That's how I read it. 
    so, hypothetically speaking, if your luck was poor, you could pull the same amount as player b and always come up with fewer 5* pulls just cause?
    That's pretty much what "random" means, yeah.

    I do assume that your predetermined pulls must change when a new character is added to the pool you're pulling from. I've pulled a bunch of Nico since she was added and I definitely haven't pulled 50+ LTs, or however many pulls get determined ahead of time.
    what we are talking about here is that pulls, and specifically 5* are no longer random though.

    i feel like a big can of worms has just been opened here. it'd be nice to actually have some official confirmation this is happening and how it's actually working.
  • Hadronic
    Hadronic Posts: 338 Mover and Shaker
    If I was gonna code this up, it would go a little something like this.

    old method: cover was rolled when token was opened.
    rerolling changes outcome.

    new method: cover is rolled when you acquire the token.
    rerolling does not change outcome.

    I know what you ask next. If they are rolled on acquisition how do i get new characters out of my hoard of old tokens? That is easy. You don't store the result of the role as green 5* thor, you store the result of the roll as a number, say .078456.

    You convert this number to a cover using a loot table.

    When the token updates, you change the loot table itself, not the rolled number.
    so .078456 corresponds to 5* thor in todays latest, but .078456 corresponds to 5* archangel when the tokens update.
  • bluewolf
    bluewolf Posts: 5,826 Chairperson of the Boards
    Just wanted to point out that the above is completely valid for tokens, but a large chunk of 4 & 5 pulls are using CP (all Classic pulls, for example) which are earned and spent at will between two (or more) stores. You can spend 20 in Classic, 25 in Latest, then 25 in the GR store tomorrow if you choose.  This point based draw system does something to eliminate rolling again for different results as well.  So a pre-determined table for each player makes the most sense.
  • Pr0spect0r
    Pr0spect0r Posts: 291 Mover and Shaker
    Xax said:
    Is this documented data provided by the devs somewhere (and if so, where) or information that's been data mined? I've never read anything about this on this forum. I'd love to read the details from the source if someone could point me in the right direction.
    there wont be, i believe they have a policy regarding data mined info. such as there have been warnings in the character discussion thread about recent character releases that may have been leaked so they were warning certain posters about info they wrote being possibly too specific.
  • Bowgentle
    Bowgentle Posts: 7,926 Chairperson of the Boards
    This information does not come from datamining.
  • Pants1000
    Pants1000 Posts: 484 Mover and Shaker
    This is just a guess, but it could be pre-rolled a few at a time in advance, like PVP matchups.

    As long as it’s random (or pseudo-random), it doesn’t really matter whether it’s rolled real-time or in the past.  People buy scratchcards all the time, and they don’t complain that it’s not fair because it was determined months ago if that card would win.
  • Pr0spect0r
    Pr0spect0r Posts: 291 Mover and Shaker
    True but then they develop unhealthy addictions chasing that win. 
  • Starfury
    Starfury Posts: 719 Critical Contributor
    Does it determine your character actually, or just if you're getting a 4* or 5*?


    What most likely happens is that the games stores a random decimal number between 0 and 1.

    Each token type has a list that "translates" a number into an outcome.

    So for an LT it might mean:

    0.85 - 1.0 = 5* character (further split into 0.85-0.9, 0.9-0.95 and 0.95-1.0 for the three respective characters)

    0.425 - 0.85 = vintage 4* (split into ~50 ranges for each of the vintage 4*)

    0.0 - 0.425 = latest 4* (split into 12 equal ranges for each of the latest 12)

    (probably the ranges are further split in three to determine colors as well)


    This would mean that when the contents of a token change, your predetermined pulls might point at a new character. (since the 2nd color of the 4th character in LT might now be ME Hulk instead of Nightcrawler)