PWI's randomizers

24

Comments

  • Airyll - Dreamweaver
    Airyll - Dreamweaver Posts: 2,882 Arc User
    edited April 2010
    Davitiel - Heavens Tear wrote: »
    True, but does it really work? I've never tried to prove it correct, but I've always thought it was just a random rumor. b:surrender

    The Duke's little tips on where to refine have been proven to be exactly what you called them: rumours. There was a refining-based thread which delved into the different ideas behind refinements and three or four people went out their way to test Duke's little bits of information and found no change in the refining capabilities; the chances were neither better nor worse.
    [SIGPIC][/SIGPIC]
  • Magicgabe - Lost City
    Magicgabe - Lost City Posts: 225 Arc User
    edited April 2010
    fyi I did a search on randomizers. Couldn't come up with something useful. From what I've seen, luck has little to do with crowdedness. It seems to follow an alternating pattern based off time.
    and bro...fyi this isn't a story.

    All people are idiodic but some are more idiotic than others.
  • Michael_Dark - Lost City
    Michael_Dark - Lost City Posts: 9,091 Arc User
    edited April 2010
    Alasen - Heavens Tear wrote: »
    Edit: Ty grow up, i was trying to state something i read elsewhere and michael jumps in right off the bat insulting me.

    Please point out ANYWHERE I've been either as insulting, offensive or childish as you.
    Alasen - Heavens Tear wrote: »
    Go crawl back under the rock u crawled out from under you 3-toed grub eating piece of sh it.

    And you're telling someone else to grow up?

    lol
    I post in forums. This one and others. That's why I post.
  • Alasen - Heavens Tear
    Alasen - Heavens Tear Posts: 1,874 Arc User
    edited April 2010
    Michael_dark - Lost City wrote: »
    Someone said something somewhere like if that something happens that it won't work because something someone said that sometimes something doesn't work...

    Could you be more vague and unintelligible?

    How bout that michael? that was your first post to me. not my fault you're a total f'ing moron that can't read.

    EDIT: although it's so far gone u prolly wont' see this Michael, i said TY as in the GUY THAT POSTED "TY" a page back, not you. Once again you fail at reading.
  • Devoted - Lost City
    Devoted - Lost City Posts: 3,634 Arc User
    edited April 2010
    @OP,

    I can pretty much guarantee you that perfect world did not create their own randomization function. Why waste resources creating their own when almost every programming languages' standard library will come with their own randomization function? I'm sure pwi is coded in c++ so they most likely use rand(). This will provide much more entropy than a simple sin wave, not to mention a sin wave is faaaaaaaaar from random. Granted, no randomization function to date can provide a truly random number, so consider the number pseudo-random.

    Fyi, most randomization functions are more complex than you think. They use large algorithms that take variables from multiple sources that offer a decent amount of entropy (ie. based on your computers real time clock, the movements of your mouse, etc..)
  • Thelas_Carr - Sanctuary
    Thelas_Carr - Sanctuary Posts: 448 Arc User
    edited April 2010
    Devoted - Lost City wrote: »
    Fyi, most randomization functions are more complex than you think. They use large algorithms that take variables from multiple sources that offer a decent amount of entropy (ie. based on your computers real time clock, the movements of your mouse, etc..)
    Sadly "rand()" is not one of those that usese multiple sources offering a decent amount of entropy. It's not confirmed that PWI uses "rand()" but it seems pretty likely; as you said, why write your own or adopt a "known good one," when the language has one built-in?

    There's also the question about the often-observed "cluster" of success values... in particular of getting a string of critical hits all in sequence or quickly picking up drops and the first 4-6 all go to the same person in a random distribution squad. It's as if some bias in the pseudom-random sequence is being experienced. Or perhaps there is, as another person suggested, a global cached value for "the next" random number that only gets updated once a second or once every 3 seconds or on some heart-beat interval.

    Generating a pseudo-random number usually involves a multiply, an add, and a modulo function and isnt' nearly as processor intensive as calculating a sine or a square root, but if it's called a huge number of times per second, then it makes sense to look at ways to optimize the function... either try a different algortithm that has fewer computations (reduce the time to execute, which may give rise to periods of predictable bias if the algorithm is less "random") or cache the data and only update it on a fixed global heartbeat (reduce the total number of calls, which may result in "clusters" of failures or successes when a number is required and you draw the same stale value until the time-mark to draw a new number).

    All speculation, but it's an interesting discussion.
  • Elenacostel - Heavens Tear
    Elenacostel - Heavens Tear Posts: 2,822 Arc User
    edited April 2010
    Devoted - Lost City wrote: »
    Fyi, most randomization functions are more complex than you think. They use large algorithms that take variables from multiple sources that offer a decent amount of entropy (ie. based on your computers real time clock, the movements of your mouse, etc..)

    I don't think rand() does; it just takes a single seed.
  • Daose - Harshlands
    Daose - Harshlands Posts: 69 Arc User
    edited April 2010
    Sorry for the spam, but this
    Go crawl back under the rock u crawled out from under you 3-toed grub eating piece of sh it.





    was hilarious.



    BTW, the first post was a wall of text. Couldn't finish it, my eyes hurt.
  • Alasen - Heavens Tear
    Alasen - Heavens Tear Posts: 1,874 Arc User
    edited April 2010
    lol ty :D..... i tooka page outta my faction leader's insult book when he rips idiots apart and CONSTANTLY wrong chats it in the faction lol...... lol ****'s hilarious sometimes when he does that.
  • Michael_Dark - Lost City
    Michael_Dark - Lost City Posts: 9,091 Arc User
    edited April 2010
    Alasen - Heavens Tear wrote: »
    How bout that michael? that was your first post to me. not my fault you're a total f'ing moron that can't read.

    Sorry, that doesn't even come close to being as childish and insulting as my new sig.

    I've also never used foul language. I don't need to call you names. Yes, I'm also better and more intelligent than you as well, seeing how you're completely oblivious to the truth, I felt I needed to point that out.

    b:chuckle
    I post in forums. This one and others. That's why I post.
  • Alasen - Heavens Tear
    Alasen - Heavens Tear Posts: 1,874 Arc User
    edited April 2010
    LOL whatever man... keep making yourself look like you think you're superior to everyone when yet in fact you're making yourself look like a total ****. you're better and more intelligent then i am? LOL wow.... go ask your mommy for permission to go play outside you snot-nosed little brat. I think u need some sunshine.
  • Devoted - Lost City
    Devoted - Lost City Posts: 3,634 Arc User
    edited April 2010
    Elenacostel - Heavens Tear wrote: »
    I don't think rand() does; it just takes a single seed.

    I never said it did, that's why I had a separate paragraph. The second was just to point out to the op that randomization isn't as simple as a sin wave. Rand() is just the real time clock afaik. b:surrender
  • Untamedf - Lost City
    Untamedf - Lost City Posts: 304 Arc User
    edited April 2010
    lol this flame war well never end b:surrender


    EDITFlameWar.jpg
    [SIGPIC][/SIGPIC] hmm hi?
  • Michael_Dark - Lost City
    Michael_Dark - Lost City Posts: 9,091 Arc User
    edited April 2010
    Alasen - Heavens Tear wrote: »
    LOL whatever man... keep making yourself look like you think you're superior to everyone when yet in fact you're making yourself look like a total ****. you're better and more intelligent then i am? LOL wow.... go ask your mommy for permission to go play outside you snot-nosed little brat. I think u need some sunshine.

    Usually I have no issues sinking down to someone else's level, you've made me realize that even I have a limit that I won't go past. lmao

    Keep raging. lol
    I post in forums. This one and others. That's why I post.
  • Elenacostel - Heavens Tear
    Elenacostel - Heavens Tear Posts: 2,822 Arc User
    edited April 2010
    Thelas_Carr - Sanctuary wrote: »
    There's also the question about the often-observed "cluster" of success values... in particular of getting a string of critical hits all in sequence or quickly picking up drops and the first 4-6 all go to the same person in a random distribution squad. It's as if some bias in the pseudom-random sequence is being experienced. Or perhaps there is, as another person suggested, a global cached value for "the next" random number that only gets updated once a second or once every 3 seconds or on some heart-beat interval.

    It is quite possible that the developers fell victim to a programming fallacy: reseeding the generator.

    The built-in C/C++ random number generator should only be seeded once for a program execution, and never again. However, many programmers erroneously believe that reseeding the function increases its "randomness". In actuality, doing so reduces the "randomness".
    Devoted - Lost City wrote: »
    I never said it did, that's why I had a separate paragraph. The second was just to point out to the op that randomization isn't as simple as a sin wave. Rand() is just the real time clock afaik. b:surrender

    Ah, my apologies, Lady Devoted.
  • Alasen - Heavens Tear
    Alasen - Heavens Tear Posts: 1,874 Arc User
    edited April 2010
    If you tell that hamster in your head to stop moving for 2 secs and think.. you'll realize you started this for no reason at all. how bout u just shut up and move on little boy
  • Devoted - Lost City
    Devoted - Lost City Posts: 3,634 Arc User
    edited April 2010
    Elenacostel - Heavens Tear wrote: »
    It is quite possible that the developers fell victim to a programming fallacy: reseeding the generator.

    The built-in C/C++ random number generator should only be seeded once for a program execution, and never again. However, many programmers erroneously believe that reseeding the function increases its "randomness". In actuality, doing so reduces the "randomness".

    I wouldn't put it past them. I'm pretty sure the devs that currently work on the code don't care much for optimization of previous code. Really as long as it's playable it's profitable, I have a strange feeling making the game run smoothly is as important as fixing jungle ruins. The game isn't very graphically intensive compared to games of it's time yet it requires a beefy processor to be able to do tw with good quality. With the recent patches it has become almost impossible to do tw and countless amount of people receive the dreaded stoplight constantly. Sometimes my client simply has to be restarted due to a memory leak.
  • Michael_Dark - Lost City
    Michael_Dark - Lost City Posts: 9,091 Arc User
    edited April 2010
    Alasen - Heavens Tear wrote: »
    If you tell that hamster in your head to stop moving for 2 secs and think.. you'll realize you started this for no reason at all. how bout u just shut up and move on little boy

    You should carefully re-read the thread from the beginning and see who looks like they fell off the first step of the short bus and busted their head on the curb. Isn't me, I'll tell you that much.

    b:pleased
    I post in forums. This one and others. That's why I post.
  • Klizzahrd - Lost City
    Klizzahrd - Lost City Posts: 91 Arc User
    edited April 2010
    ok so back to the subject.....

    i actually am more interested in learning about the "blocking" idea, because it would explain alot if its true....like it would explain how I can go into HH3-1 during 2x and come out with only 1 mirage (same for bh, ect.) Best one so far was a dude who got all 3 molds and 3star from bh boss. Even after he dropped one, he still kept getting it until he'd dropped it enough times. It feels like the random generator get stuck one/two people in squad, or totally forgets others.

    And I've noticed the fail/ success trends while grinding as well. It's very obvious during 2x. You will go for a while with totally nothing, and then suddenly DQ items are dropping like crazy, and then suddenly you're getting nothing again...for a long time...


    So would waiting a while before picking up items/killing mobs help at all with this? is there some trik to make this stuff actually random?
  • Untamedf - Lost City
    Untamedf - Lost City Posts: 304 Arc User
    edited April 2010
    Michael_dark - Lost City wrote: »
    You should carefully re-read the thread from the beginning and see who looks like they fell off the first step of the short bus and busted their head on the curb. Isn't me, I'll tell you that much.

    b:pleased


    yep this war well never end... b:embarrass
    [SIGPIC][/SIGPIC] hmm hi?
  • Delia - Harshlands
    Delia - Harshlands Posts: 227 Arc User
    edited April 2010
    To the people derailing this thread: Some of us actually care about the topic, could you please take it elsewhere? Thanks :)
    Thelas_Carr - Sanctuary wrote: »
    There's also the question about the often-observed "cluster" of success values... in particular of getting a string of critical hits all in sequence or quickly picking up drops and the first 4-6 all go to the same person in a random distribution squad. It's as if some bias in the pseudom-random sequence is being experienced. Or perhaps there is, as another person suggested, a global cached value for "the next" random number that only gets updated once a second or once every 3 seconds or on some heart-beat interval.

    This is the one of the few things that makes me believe PWI-random may be somewhat less random than your standard pseudo-random generator. Interesting idea about the global - worth looking into.
    Elenacostel - Heavens Tear wrote: »
    It is quite possible that the developers fell victim to a programming fallacy: reseeding the generator.

    I wouldn't rule that out.
    Devoted - Lost City wrote: »
    I never said it did, that's why I had a separate paragraph. The second was just to point out to the op that randomization isn't as simple as a sin wave. Rand() is just the real time clock afaik. b:surrender

    Technically only based on the system time if they seed it that way. I wouldn't put it past PWI to have some other (probably less random) way of generating it.



    The thing is, most of the random-fails they could do wouldn't result in something easily detectable, like getting lots of low/high randoms in a row. So if there is something to have reason to believe is off, I think Thelas' idea sounds one of the more plausible.

    Of course, humans tend to view random sequences of events as non random. Humans actually expect a lot more variation in random streams than there actually is. So I'm not opposed to thinking this is just the case.

    And while I think there might be reason to believe otherwise, I will continue to tell the people who complain that random is random. Although it may be bugged time wise, there's no evidence to show that some people have "lucky genies" or "lucky accounts" or "lucky with drops" or that standing closer to the boss makes you more likely to get drops, or anything else.

    My motto for figuring out how PWI works under the hood is like this: The laziest approach to get it done is probably the way it actually works. This has served me well understanding how instance ownership and the teleport timer works, among other things. This doesn't mean there aren't bugs (of course there are), but it means that I would find it highly surprising if PWI is doing anything to intentionally make random events less random, like having a sine wave multiplier.
  • Devoted - Lost City
    Devoted - Lost City Posts: 3,634 Arc User
    edited April 2010
    Klizzahrd - Lost City wrote: »
    ok so back to the subject.....

    i actually am more interested in learning about the "blocking" idea, because it would explain alot if its true....like it would explain how I can go into HH3-1 during 2x and come out with only 1 mirage (same for bh, ect.) Best one so far was a dude who got all 3 molds and 3star from bh boss. Even after he dropped one, he still kept getting it until he'd dropped it enough times. It feels like the random generator get stuck one/two people in squad, or totally forgets others.

    And I've noticed the fail/ success trends while grinding as well. It's very obvious during 2x. You will go for a while with totally nothing, and then suddenly DQ items are dropping like crazy, and then suddenly you're getting nothing again...for a long time...


    So would waiting a while before picking up items/killing mobs help at all with this? is there some trik to make this stuff actually random?

    Honestly, it's just human nature to notice something out of the ordinary more than something common. That's somewhat obvious...

    What about all those times where something wasn't an extreme case?

    6 green mats drop: "wow awsome! i got lucky!"

    0 green mats drop: "wow that sucks, i'm really unlucky!"

    3 green mats drop: "wow normal outcomes are so memorable!!"
  • Devoted - Lost City
    Devoted - Lost City Posts: 3,634 Arc User
    edited April 2010
    Devoted - Lost City wrote: »
    I wouldn't put it past them. I'm pretty sure the devs that currently work on the code don't care much for optimization of previous code. Really as long as it's playable it's profitable, I have a strange feeling making the game run smoothly is as important as fixing jungle ruins. The game isn't very graphically intensive compared to games of it's time yet it requires a beefy processor to be able to do tw with good quality. With the recent patches it has become almost impossible to do tw and countless amount of people receive the dreaded stoplight constantly. Sometimes my client simply has to be restarted due to a memory leak.
    Delia - Harshlands wrote: »
    My motto for figuring out how PWI works under the hood is like this: The laziest approach to get it done is probably the way it actually works. This has served me well understanding how instance ownership and the teleport timer works, among other things. This doesn't mean there aren't bugs (of course there are), but it means that I would find it highly surprising if PWI is doing anything to intentionally make random events less random, like having a sine wave multiplier.

    Yay, someone that understands. I'm sure the code pw runs takes drastic performance hits to improve the cost. The line "The laziest approach to get it done is probably the way it actually works" is exactly how I feel but replace laziest with cheapest.
  • Alasen - Heavens Tear
    Alasen - Heavens Tear Posts: 1,874 Arc User
    edited April 2010
    Michael_dark - Lost City wrote: »
    You should carefully re-read the thread from the beginning and see who looks like they fell off the first step of the short bus and busted their head on the curb. Isn't me, I'll tell you that much.

    b:pleased

    LOL I NEED to go back and read? I said something perfectly related to the topic that i ACTUALLY BOTHERED READING, and you come at me saying i'm a idiot and basically can't form a thought process. Which is kinda of funny considering i pretty much quoted what i was referring to from what i could remember. Go search the forums for "mirage refining", and look for a thread called Refining for Cheap i think. READ THE DAMN THREAD and find the post where the guy mentions what i did. Until then keep your idiotic mouth shut. By the way, my friend on the phone as well has half my faction has read this and they're all saying you're a total f'ing idiot who needs to go get a life and grow up (to put it blunt and succintly). How old are you like 12? This is coming from people like me in their 30's, seriously you had no call to insult me. But do you honestly think i'm going to sit here and take it? Get a life jackass
  • Alasen - Heavens Tear
    Alasen - Heavens Tear Posts: 1,874 Arc User
    edited April 2010
    Delia - Harshlands wrote: »
    To the people derailing this thread: Some of us actually care about the topic, could you please take it elsewhere? Thanks :)

    Tell that to michael, i WAS talking about the topic..... b:surrender
  • Elenacostel - Heavens Tear
    Elenacostel - Heavens Tear Posts: 2,822 Arc User
    edited April 2010
    Before we continue to theorize about the correctness of the random number generator(s) used in the game, it should first be measured.

    1) Punch the Ancient Sea Dragon a few thousand times. Record all Hits, and Crits, IN ORDER. Do not discard the sequence of hits and crits. You can discard the misses.

    2) Record your crit % as seen through an Eye of Revealing or whatever it's called. There are a few skills that do not show up in your char menu (game bug), which is why you need to use this item.

    3) Take those thousands of Hits and Crits and feed the expected crit % and the entire sequence into a statistical pattern recognition program.

    4) The program will be able to tell you how close the sequence of Hits and Crits is to being actually "(pseudo)random", based on the input data. The expectation is that it should be as "(pseudo)random" as the standard random number generators you'd expect in a programming language (such as C/C++).

    Until a test like this one (or better) is performed, we really have no hard evidence to say whether or not PWI is correctly (pseudo)random.

    Note that I am not a statistician. The above experiment was created in a few minutes within my head. I am certain there are better ways to test the randomness of PWI. My main point is that a test needs to be done before we can appropriately call out any bug or unfairness of the game.
  • Klizzahrd - Lost City
    Klizzahrd - Lost City Posts: 91 Arc User
    edited April 2010
    Devoted - Lost City wrote: »
    Honestly, it's just human nature to notice something out of the ordinary more than something common. That's somewhat obvious...

    What about all those times where something wasn't an extreme case?

    6 green mats drop: "wow awsome! i got lucky!"

    0 green mats drop: "wow that sucks, i'm really unlucky!"

    3 green mats drop: "wow normal outcomes are so memorable!!"

    rofl the sad thing is that I DO notice when random is working well....cause everybody gets something during the bh, which at least for my experiences is odd since I notice that normally a person is left out or just gets one or two really crappy items (and I usually feel bad for them :/ )

    Im not complaining (we all have lucky runs)...I'm more curious as to how PWI's random generator works. Those chains happen a bit too often for me to think it's just a coincidence.

    The only other thing I can think of is that a bh/HH just isn't long enough for everything to even out. If you have a coin and flip it 10 times it often will look like it favors 1 side over the other. Flip 100...200...1000 times and the results are more even. (Except a squad would be more like a die) It's possible that this is happening, although I see no way to test it since you can't make the bosses respawn over and over in a bh/hh without reforming squad, and i doubt anybody wants to grind in-squad for days on end recording every single drop.
  • skyxiii
    skyxiii Posts: 0 Arc User
    edited April 2010
    Why do crits and procs cluster?b:puzzled
    Pretty wonky if you ask me.
  • Asterelle - Sanctuary_1381265973
    Asterelle - Sanctuary_1381265973 Posts: 7,881 Arc User
    edited April 2010
    skyxiii wrote: »
    Why do crits and procs cluster?b:puzzled
    Pretty wonky if you ask me.

    If there were no clusters it wouldn't be random.

    When most people think of random they think like this:
    quasi_random_scatter.gif
    With numbers distributed to fill up the space of possibilities evenly.

    A uniform randomness however looks like this:
    uniform_random_scatter.gif

    See the clusters and empty spaces? Not only do clusters necessarily exist with random numbers, they are also more memorable.
    [SIGPIC][/SIGPIC]
    Refining Simulator - aster.ohmydays.net/pw/refiningsimulator.html (don't use IE)
    Genie Calculator - aster.ohmydays.net/pw/geniecalculator.html - (don't use IE)
    Socket Calculator - aster.ohmydays.net/pw/socketcalculator.html
  • Aadi - Lost City
    Aadi - Lost City Posts: 4,449 Arc User
    edited April 2010
    Statistical pattern recognition programs...graphs of quasi- and uniform random scatter...

    ...Archers b:surrender
    [SIGPIC][/SIGPIC]
    "aadi is a forum ninja, always there, skirting thru the shadows... striking with quick posts while you are distracted by your own" -Alexeno(kin)
    "We talk about you because you're fab. b:cute" -Chillum
    "You live for forums. Like seriously."

Categories