There is a bug or intentional code in POE's random number generator that is causing streakiness.

"

Xendran wrote:

So, Real_Wolf, where is the quote from GGG saying that their RNG system has no influencing factors?

They said they have faith in it working properly. They gave you no details about how it works.
For all we know it could have SIGNIFICANT factors at play, but working correctly.

You know nothing about the RNG system. Nobody in this thread does.

I didn't say GGG claimed their PRNG was perfect. I claimed they said they were happy with the result.

I used basic logic to say GGG wouldn't want a deterministic RNG scenario, so when they state its working as they want it must be doing a good interpretation of randomness.

Whats more having a look at PRNG options out there, they have gotten significantly complex enough that they are very close to true RNG. The issue is true RNG inherently has a streak of 1's that stretches near to infinity, making it seem not random.


I trust GGG when they say they are happy with how it is working. I also would prefer to believe them over anecdotal evidence of one user (or even 10) over a limited time scale.

Unless it is repeatable. If someone can show they can repeat teh same set of actions and get the same result.

At that point I will join you and say RNG is broken

"

Powertrip wrote:

It's actually pretty funny to see one group of people simply giving feedback on their impressions of the rng system based on gameplay time and a second argue that the rng system is pure based on them having never seen the actual code.

Nope, that's another misrepresentation. The OP claimed unequivocally "there is an error in PoE's Random Number Generator", right in the title of the thread. Some of us challenged him to show his evidence, and it was far from persuasive, much less conclusive. In other words, he's full of it.

Pointing out that the OP's brash, unproven claims are in fact, nothing more than anecdotal suspicions is NOT tantamount to professing faith in the "purity" of what he presumed to condemn. As far as this Beta Feedback forum is concerned, what I, the OP, or anyone else may believe about GGG is irrelevant and off-topic.

"

RogueMage wrote:

"

Powertrip wrote:

It's actually pretty funny to see one group of people simply giving feedback on their impressions of the rng system based on gameplay time and a second argue that the rng system is pure based on them having never seen the actual code.

Nope, that's another misrepresentation. The OP claimed unequivocally "there is an error in PoE's Random Number Generator", right in the title of the thread. Some of us challenged him to show his evidence, and it was far from persuasive, much less conclusive. In other words, he's full of it.

Pointing out that the OP's brash, unproven claims are in fact, nothing more than anecdotal suspicions is NOT tantamount to professing faith in the "purity" of what he presumed to condemn. As far as this Beta Feedback forum is concerned, what I, the OP, or anyone else may believe about GGG is irrelevant and off-topic.

Yeah, I'll give you that, title was less conversation starting and more stir things up.

Where's the post from GGG about them having faith in their random number generator? (turn off your sarcasm detectors, I'm legitimately curious and I'm not calling the post into doubt at all) I haven't seen it, would be nice to see. Also, even the most faithful can make mistakes.

It's up to you to decide if my presentation of evidence is persuasive or not, though. Either you're GGG and you'll read the posts in this thread and countless other threads and consider taking a look in your code, or you won't consider it - or you're a player who wants to judge my conclusions. If you want to judge my conclusions, then perform tests yourself, look at player evidence yourself in the forums or in global chat, but for Dominus's sake don't just blindly latch on to ideas of "x is trolling, y doesn't understand probability because I read about it somewhere," etc. etc. and proclaim it without thinking about what you're saying.

Essentially, what I present is two messages to GGG and to the community:
GGG: "You should consider looking at your random number generation code and/or software and/or hardware"
Community: "You should be aware that there might be some streakiness in random number generation in this game. If you can observe patterns and share your experiences, you should, so that GGG is aware of the problem if there is indeed a problem. Otherwise, if you observe that there is no problem, you should share that experience as well."

"

mark1030 wrote:

"

The odds of winning the mega millions jackpot, according to Wikipedia, is "1 in 175,711,536"

Let's make this a real comparison that sounds more like the arguments I'm hearing as far as the game goes.

If over the course of my life, I bought 350 million tickets. How likely would it be that I would win the jackpot twice? I bought twice as many tickets as the odds to win, so I should be guaranteed to have won twice. If I don't win at all, does that mean something is suspicious?

Um, what? That's pretty far from what I was saying. What I'm saying is,

"If over the course of my life, I bought 350 million tickets, and I won the jackpot twice within 12 tickets, doesn't that mean there's something wrong with the ball bouncing thing? I mean, how could I win TWICE within twelve tickets? That doesn't make any sense."

EDIT: Thread title edited to reflect Xendran's conjecture.

"

adghar wrote:

What I'm saying is,

"If over the course of my life, I bought 350 million tickets, and I won the jackpot twice within 12 tickets, doesn't that mean there's something wrong with the ball bouncing thing? I mean, how could I win TWICE within twelve tickets? That doesn't make any sense."

LOL, every hypothetical example you concoct demonstrates how you fail once again, to comprehend how random processes actually work.

"

RogueMage wrote:

LOL, every hypothetical example you concoct demonstrates how you fail once again, to comprehend how random processes actually work.

Are you sure that's so, or are you missing my point?

Yes, I do know that getting 2 lottery wins within 12 tickets is possible, and may make it into the papers as a strange occurrence. "How is this possible?" is perhaps a dumb question to ask in that case.

What about 2 lottery wins within 6 tickets?
2 lottery wins within 2 tickets?

6 lottery wins all in a row?

When one person wins 12 lottery rolls all in a row, do you continue to accept that the process is random, or do you begin to suspect that there is another phenomenon causing the process to become less random?

What about 24 lottery rolls all in a row?

It is up to the individual player to decide where the number of cases of streakiness observed within the community falls on the spectrum of analogy (just 2 people winning the lottery in history, or 24 lottery wins all in a row on one person, or anywhere in between). I'm not sure if it's of the most benefit to them to simply read people's personal attacks on other people with no arguments to back it up, or for them to think about probability and consider whether there is the possibility of something other than effective random number generation happening in POE.

I've already said it, but I'm going to say it again, because apparently I like babbling on about things of little importance:

It doesn't matter to me if YOU, the reader, is convinced that POE's random number generator is not sufficiently random.

The only thing that matters to me is that IF there is a problem, GGG should fix it after other priorities are taken care of. And I am personally convinced that there is a problem. That's all there is to it.

Whether you, the reader, are convinced or not, well, that may influence GGG's decision. But you, the reader, should be performing as a rational individual with reasoning for your actions instead of ONLY listening to hate-slinging and repeated generalisms. The repeated generalisms are good as a starting point, yes, but generalisms are general, to phrase it in a popular fashion.

I don't feel motivated to do science/statistics to convince GGG. But I offer my experience and the experience of others in feedback so that IF there is a problem (and I am convinced there is one), then it should be fixed when other priorities are handled. Bottom line.

RogueMage, I can't quite remember if you provided posts with actual counter-evidence (and I'm too lazy/would rather play than check), but if you did, then thanks for providing constructive posts at that time.

"

adghar wrote:

"

RogueMage wrote:

LOL, every hypothetical example you concoct demonstrates how you fail once again, to comprehend how random processes actually work.

Are you sure that's so, or are you missing my point?

I've already said it, but I'm going to say it again, because apparently I like babbling on about things of little importance:

It doesn't matter to me if YOU, the reader, is convinced that POE's random number generator is not sufficiently random.

Thanks for making it abundantly clear that you could care less what anyone else says, you're just gonna...

Keep on Trollin'

http://www.youtube.com/watch?v=zYDl-HwsnaM

I was grinding,
Playing in the streets, I was unknowing;
I didn't know my fate.
Playing
The game of PoE, but never really scoring;
I thought my luck would break.

I was a young man and sure to go astray.
You walked right over my drops and told me RNGesus would find a way
To keep on trollin', keep on trollin', keep on trollin'
Yeah, yeah, yeah.

I was standing,
Looking in the face of one who PK'ed me,
Feeling so ashamed.
Hoping,
And praying, that rogue exile could understand me,
But then I read her name.

She took my hand in hers and told me I was wrong.
She said, you're gonna be all right boy, oh just as long
As you keep on trollin', keep on trollin', keep on trollin'
Yeah, yeah, yeah.

Maybe,
Someday baby, who knows where or when,
Just you wait and see.
We'll be face-rollin',
Together hand in hand, alone forever;
Shavronne just you and me.

That RNG has gotta change, Lord, I know it's true;
Our luck is gonna keep on glowing and growing and it's all we gotta do.
Keep on trollin', gotta keep on trollin', gotta keep on trollin'
Yeah, yeah, yeah.

In retrospect, assembling a team to try to find out IF there is a problem in random number generation is probably too much of a hassle with only a 43-person team, much less finding the problem inside the algorithm itself and fixing it. So, my thread probably would have been better served posted much later, in, say, patch 1.2.1, or something like that. Oh well. It's nice for people to be made aware of the possibility, though.

Are you comparing winning the lottery 6 times in a row to getting the same unique to drop 3 times in an hour? Because that's the extent of this streakiness that people are saying proves that RNG is broken. I think the drop rates are a little higher than lottery odds.

Streaks are just part of randomness. Random doesn't mean you will always have different outcomes, which seems to be what people think the RNG should deliver. Rolling a 6-sided die 6 times will more than likely not roll each number once. Rolling the number 3 twice or three times in a row does not mean it isn't random. That's the extent of this streakiness people are claiming. I haven't seen any that say "I found 6 of the same uniques in a row". Three is the most I've seen.

"

mark1030 wrote:

Are you comparing winning the lottery 6 times in a row to getting the same unique to drop 3 times in an hour? Because that's the extent of this streakiness that people are saying proves that RNG is broken. I think the drop rates are a little higher than lottery odds.

Streaks are just part of randomness. Random doesn't mean you will always have different outcomes, which seems to be what people think the RNG should deliver. Rolling a 6-sided die 6 times will more than likely not roll each number once. Rolling the number 3 twice or three times in a row does not mean it isn't random. That's the extent of this streakiness people are claiming. I haven't seen any that say "I found 6 of the same uniques in a row". Three is the most I've seen.

The thing about that is a dice has 6 sides. There are 150 uniques, which isn't quite the same. Their drop chance isn't all equally weighted, though. And to make matters worse, it's tied to clear speed, since the more you clear, the more chances you have to get a unique, which may or may not be the same. And clear speed varies. So it's a huge brainsore to do a full, scientific analysis of things.

The problem currently in question is "the expected number of streaks in a certain time frame" compared to "the number of streaks that occur in a certain time frame." What I've observed in my own play is streak after streak after streak, always on some different aspect, sometimes taking a break, sometimes simultaneously with other streaks. And look at the NUMBER of people who had "3 uniques in a row," in a certain time frame. I don't even want to count that up at the moment, since I'm tired and stuff (I like typing, but don't really like digging up old posts). But is it really expected that THAT many streaks should happen? Streaks have been occurring even since I just started playing. Maybe my account is bugged like all those "weeks without wisdom scrolls." But other players are noting their experience of high levels of streaks as well.

One point - I'm looking at a bigger picture. Rolling a 6-sided die 6 times with 3 3's in a row is pretty normal. Several people rolling 6-sided dice and getting 3- and 4-streaks (like, say, 85 out of 100 people) is grounds for suspicion.

Second point - drop rates or monster spawn rates are on a completely different scale from dice rolls. Do you think every hour contains 1/6 chance of a certain unique dropping? No, of course not, there are 150 uniques, the drop rate is likely much less, and has many competitors.

Also - let's return to the lottery example. Certainly, one can't compare the streakiness currently being observed, to winning the lottery 6 times in a row. Not quite. If we wanted a more perfect analogy, it'd be something like several people buying mega millions tickets, and winning $150 several times in a row, but getting no wins from other lotteries, or something like that. At this point, using analogies becomes less useful.

The point of "6 times in a row" was just to point out that scale counts. Suspicious outcomes of a large scale are more likely to be non-random than "suspicious" outcomes of a small scale. 3 3's in a row on a six-sided dice is "suspicious," but gives close to 0 reason for thinking it's anything other than random. 3 42's in a row on a 150-sided dice is "suspicious," but very much within the realm of possibility - in fact, it's predicted to happen multiple times in a game with tens of millions of players, so you can't really take it as evidence that something non-random is happening. However, when you take into account that a "150-sided dice" roll isn't even guaranteed for an hour or even a day of playing, things start to get a little muddled up, and extremely hard to analyze with numbers - but certain players can at least feel that something is off, that the frequency of streaks is just a little bit too high for comfort. And for me, it's not just a little bit - it's an insanely high amount.

Sorry if I'm rephrasing myself too much, but let's consider this: your first sentence speaks as if my original post contains a single, solitary example of a single person getting 3 of the same unique in a row. But that's not the case. Countless people have spoken up about their streaks, the frequency of streaks they've seen, and I myself have provided several examples and can think of several more (for funsies, here they are: the LAST time I got high-level currency, I got two exalts within 2 days. Pretty much no exalts before that or months after that. Also, I just ran into 2-elemental-reflects [NO other rares in between], and later, another 2-elemental-reflects [NO other rares in between]. Think of all the monster mods that can be rolled. Think about this streak happening in close proximity to all the other streaks. Think about the frequency of streaks that has been observed. Think about the frequency of streaks that is expected. There is an incongruency).