Rng's - Short term trends.
Quote: EvenBobIf you have a lot of experience you can tell the difference between pseudo random and real random on games like roulette.
While there are "amounts of randomness" - see volume 2 of Knuth's The Art of Computer Programming (and I have come across IBM's notorious randu() function - in fact, when I learned programming on an IBM 1130, I was shown a different RNG that got its numbers by "reading from the line printer") - you can never tell with 100% confidence that a sequence of numbers is from an RNG as it is always possible for the non-RNG source to generate the same sequence.
Quote: mickeycrimmI used to have respect for you. Now I think you are the dumbest of the dumb. Take you old ass of wisdom and shove it right up your ass. You don't know jackshit, bitch. You are a clown.
I mentioned to someone else this would
happen, it's like telling a Catholic there is
no Virgin Mary. Certain beliefs are so
entrenched in the gambling community
that they get this kind of almost hysterical
reaction every time.
PROVE IT, they scream, like a color blind
person being shown two shades of green
and then told they aren't the same.
Quote: mickeycrimm
Bob, your cornhole shit will never fly with the American kids.
Which American kids, be specific. How about Asian
kids, or French kids. How many kids on this forum,
lets see a show of hands..
Are you merely trying to state such? It is well known that games that go through a predictable series of outcomes from a pseudo-random number generator can be known to go through predictable states and therefore exploited.
http://www.cigital.com/papers/download/developer_gambling.php
So many times I think many on here just argue without even knowing what the other person is talking about.
I can assure you that I don't know what you are talking about.
But if you want someone to agree with you, I can agree that if you have a known seed to a pseudo-random number generator and you generate pseudo-random numbers, they are in fact predictable for that given seed. That's why they call it a "pseudo-random" generator.
In order to prevent things from being predictable, care has to be taken in the development of any casino application.
At GLI, that's almost everything that they do to go over code to ensure that it meets standards for ensuring that there is no way to predict the outcome of the game and that everything is fair.
http://www.gaminglabs.com
We took a tour of their facility several months ago, and they have tons of games in there all partially disassembled.
I don't know if you're aware of how this stuff works, but basically GLI collects money from just about everyone who makes games (maybe over-generalizing a bit) and they have a little cube farm of guys who just do math who pore over the code to determine if there is any weakness and advise accordingly.
There are all sorts of techniques that are used in order to prevent games from being exploitable. And there are all sorts of stories about various games being exploitable due to bad design.
Just plopping in a pseudo-random number generator and sticking in the same seed each time the game was powered on would never work. Some people have a hard time understanding the lengths that companies go through to ensure that there is enough data being shuffled around internally to make sure that everything is fair. But an RNG is just one tiny little piece of the overall puzzle.
Quote: AhighA pseudo-random number generator is called such for a reason: it's not really random.
Are you merely trying to state such? .
No. I'm saying that in the long run, over millions
of outcomes, RNG's are identical to true random.
In the immediate short term, they are not, and
the difference can be seen. Everybody assumes
they are identical in every way and they are not.
Quote: AxelWolfGive him a chance to prove it.
It's the same nonsense that everyone else says. You see so many more results per hour with machines that you obviously see rare stuff much more frequently. 1 time in a thousand becomes once every few hours instead of once every few days.
Remember that this is coming from the same guy who used to claim that he had a friend who played roulette and never lost, and then changed his story to be that he plays roulette and never loses.
Quote: AxiomOfChoiceYou see so many more results per hour with machines that you obviously see rare stuff much more frequently.
It has nothing to do with results per hour.
You have to compare session to session,
not results per hour. Making 40 bets on
a real wheel and 40 bets on an RNG. Over
a long period of time.
The worst thing about RNG's is, you see
things that statistically don't happen on
a real wheel very often, you see them
all the time on an RNG. It's laughable.
Then the math guys come along and
say, it all falls within the parameters of
random. To which I always reply, good,
you waste your money on them then.
Personally, I think the RNG's are more
statistically accurate than a real wheel.
That's the problem. Real random is not
very statistically accurate in the short
term and RNG's are, because they are
fake random. Real random results are
very hard to produce artificially, as we
all know.
I would be willing to bet he cant look at a string of numbers from a live roulette game VS a RNG and tell the difference.Quote: AxiomOfChoiceIt's the same nonsense that everyone else says. You see so many more results per hour with machines that you obviously see rare stuff much more frequently. 1 time in a thousand becomes once every few hours instead of once every few days.
Remember that this is coming from the same guy who used to claim that he had a friend who played roulette and never lost, and then changed his story to be that he plays roulette and never loses.
Quote: AxelWolfI would be willing to bet he cant look at a string of numbers from a live roulette game VS a RNG and tell the difference.
I'll bet I can't either, it doesn't work
that way.
If you have a lot of experience you can tell theQuote: EvenBobI'll bet I can't either, it doesn't work
that way.
difference between pseudo random and real
random on games like roulette.
EvenBob: you see things happen on
a regular basis that you only see in a real
wheel or shoe once in a blue moon.
EvenBob:In the immediate short term, they are not, and
the difference can be seen
Sorry your many statements confused me, it must be due to you talking out your_____.
Sorry your many statements confused me, it must be due to you talking out your_____.Quote: AxelWolf
You said a 'string' of numbers, I don't deal
in strings, I deal in sessions. It's not looking
at a few, it's playing a couple hundred in a
few sessions.
Quote: EvenBobSorry your many statements confused me, it must be due to you talking out your_____.
You said a 'string' of numbers, I don't deal
in strings, I deal in sessions. It's not looking
at a few, it's playing a couple hundred in a
few sessions.
.Well how many would it take for you to conclude one way or the other?
Quote: AxelWolfWell how many would it take for you to conclude one way or the other?
Why do you care, what's the difference.
I don't do tests, this isn't 8th grade.
You can't prove they're the same in the
short term, I'm not asking you to take
a test.
I never claimed I could tell the diffidence, you did. If you could tell the difference this would be a very big deal. It would mean they are not even close to being random. This would headline and rock the casino world and be a huge scandal.Quote: EvenBobWhy do you care, what's the difference.
I don't do tests, this isn't 8th grade.
You can't prove they're the same in the
short term, I'm not asking you to take
a test.
I never claimed I could tell the diffidence, you did. If you could tell the difference this would be a very big deal. It would mean they are not even close to being random. This would headline and rock the casino world and be a huge scandal.Quote: EvenBobWhy do you care, what's the difference.
I don't do tests, this isn't 8th grade.
You can't prove they're the same in the
short term, I'm not asking you to take
a test.
I am glad to give you mathematical proof of everything I play.
If I have an outcome for an RNG between the values of 1 and 38 and it comes up 29 twice, is that a short term trend?
I will say that studying Pearson's Chi-squared test and running the test on various strings of random outcomes to assert the chance that this string of outcomes adheres to what one would expect for a random series of numbers was very educational.
But even then, the fact is that the inability to come to a conclusion is not, in and of itself, not a significant point for discussion when it comes to "probabilities."
The explanation, "yes, weird stuff happens" does, in fact, hold water!
It's the weird stuff that makes gambling fun, you know!!!
Quote: AxelWolfIt would mean they are not even close to being random. .
lol, no it wouldn't. We already KNOW it's fake random.
What we're talking about is, you say they're identical
and I say they aren't, in the short term. Can you prove
that they're identical in the short term? No, you can't.
You just assume they are.
We know they are not completely 100% random. They are close enough, there is nothing you yourself can detect.Quote: EvenBoblol, no it wouldn't. We already KNOW it's fake random.
What we're talking about is, you say they're identical
and I say they aren't, in the short term. Can you prove
that they're identical in the short term? No, you can't.
You just assume they are.
I know you can't tell the difference between a Live roulette and a Video roulette.
Example If i were to play 10 separate sessions (whatever amount you believe is needed ) on a live roulette and an video roulette then present you with all the data you would not be able to tell the difference like you claim. you would have a 50/50 shot at getting each session right.
IF you yourself or anyone you know could tell the difference, this would mean they are not even CLOSE to being random at all. Don't downplay how this would be significant. I'm sure someone more mathematical or someone who has dealt with RNG's could explain how shocking it would be if you could accurately tell the diffidence.
The fact is they are random enough no not make a diffidence in the short or long run.
Edit: No I cant prove it, because I cant tell the difference, you are the one claiming you can. The proof falls on you.
Quote: AxelWolfIF you yourself or anyone you know could tell the difference, this would mean they are not even CLOSE to being random at all.
I'm going to type very slowly now, because you
obviously don't know what this is about. Nobody
is disputing they are both random. They are two
different KINDS of random. Pseudo and true.
They are BOTH random! That's a fact. The
argument is, they are identical, which they
are not. In the short term.
This is obviously all very new to you, take
a breath and relax.
Quote: EvenBobI'm going to type very slowly now, because you
obviously don't know what this is about. Nobody
is disputing they are both random. They are two
different KINDS of random. Pseudo and true.
They are BOTH random! That's a fact. The
argument is, they are identical, which they
are not. In the short term.
This is obviously all very new to you, take
a breath and relax.
And again if they are not identical in the short term as you claim then you should easily be able to tell them apart but you refuse to show this to anyone. If you could do this it could lead to easy exploitation of the game and making tons of money yet you refuse to do that. Is the reason you refuse to do that because what you are claiming is hogwash.
He will just claim it does not work that way.Quote: TwirdmanAnd again if they are not identical in the short term as you claim then you should easily be able to tell them apart but you refuse to show this to anyone. If you could do this it could lead to easy exploitation of the game and making tons of money yet you refuse to do that. Is the reason you refuse to do that because what you are claiming is hogwash.
Quote: Twirdmaneasily be able to tell them apart but you refuse to show this to anyone. If you could do this it could lead to easy exploitation of the game and making tons of money
Really? Easily? You think it's easy, go ahead
and try it. And explain how telling them apart
leads to making money. So I can tell one random
from another, they are still random. How does
seeing that lead to money.
Quote: EvenBobReally? Easily? You think it's easy, go ahead
and try it. And explain how telling them apart
leads to making money. So I can tell one random
from another, they are still random. How does
seeing that lead to money.
I'm saying its basically impossible to tell them apart but you are saying they are different so you should be able to tell them apart. Also if they are different there has to be some reason they are different this difference from true random can lead to exploitability. Are there more streaks bet on that, are long streaks less like to appear bet on that, there has to be something differentiating the 2 otherwise you couldn't tell them apart.
Quote: IbeatyouracesMaybe it's why when I'm playing VP, I'm consistently getting a certain pair dealt to me over and over again compared to other pairs. The last few days it's been 9's.
I play Vegas Star Single Zero Roulette (PRNG) at my local casino
Starting in April this year I started keeping a record of spins that I saw on my
weekend visits to the casino, not continuous, but as I observed when I was there.
I collected 1782 spins over a period of 6 weeks and entered them
into RX and did an analysis. IT FAILED the tests.
Results: 1782 spins, Chi-Square 72.7755, Test % 0.03, Mean 48.16, SD 6.85
The PRNG produced very similar numbers for 6 weeks on end, easily
enough to make significant returns just by playing the top 12 numbers
The strike rate was well in excess of the expected rate for 12 numbers
Then after 6 weeks, the numbers I observed changed, and now they pass
the tests, at least for the moment
This does seem to support the theory that sometimes the outcomes are not random
He found that after receiving a final hand of four-of-a-kind (quads) in Deuces Wild, the frequency of a final hand of quads was less than the expected frequency of one in about fifteen hands. After Mr. Paymar published his observation, many readers reported observing the same phenomenon. Mr. Paymar then kept statistics on 160,000 hands of Deuces Wild video poker and found that the freqency of quads-after-quads was one in twelve hands. That is a more than six standard deviations from a random distribution.
Trying to convince advantage players (AP) that there can be nonrandom anomalies in pseudorandom number generator (PRNG) outputs is even more difficult than trying to convince a recreational player (RP) that a skilled AP can attain a long-term positive expected value (EV+) by selectively playing EV+ casino games or than trying to teach a pig to sing. I have no desire to proselytize AP's to shed the scales from their eyes and go forth and find and exploit EV+ PRNG's. Nor do I want to encourage unskilled RP's to find nine out of every zero EV+ anomalies.
tl;dr However, I would love to have a high signal-to-noise discussion with those interested parties who are interested in generating more light than heat
on this interesting subject.
If you can accurately detect and tell the differences between a normal gaming machine RNG apart from live random, you will make money from this in a variety of ways. Best of all you will be be considered a absolute genus(opposite what people think now), people will be busting down your door to talk to you. I'm sure some companies would want to work with you. This would be highly valuable to help find a solutions to creating a better RNG's( If Bob can spot the"RNG" it sucks.)Quote: EvenBobReally? Easily? You think it's easy, go ahead
and try it. And explain how telling them apart
leads to making money. So I can tell one random
from another, they are still random. How does
seeing that lead to money.
Quote: PandoThis does seem to support the theory that sometimes the outcomes are not random
I can believe that people have implemented games which do not have random results.
That's not the claim though. The claim is that pRNGs produce results that are distinguishable in a short term from RNGs, but both are still pass tests for random in the long term.
(pRNG - psuedo random number generators - such as the use of the Mersenne Twister).
MathExtremist would be far better able to explain it why the claim makes no sense, but not seen him around for a while.
''Quote: AhighA pseudo-random number generator is called such for a reason: it's not really random.
Are you merely trying to state such? It is well known that games that go through a predictable series of outcomes from a pseudo-random number generator can be known to go through predictable states and therefore exploited.
http://www.cigital.com/papers/download/developer_gambling.php
So many times I think many on here just argue without even knowing what the other person is talking about.
I can assure you that I don't know what you are talking about.
But if you want someone to agree with you, I can agree that if you have a known seed to a pseudo-random number generator and you generate pseudo-random numbers, they are in fact predictable for that given seed. That's why they call it a "pseudo-random" generator.
In order to prevent things from being predictable, care has to be taken in the development of any casino application.
At GLI, that's almost everything that they do to go over code to ensure that it meets standards for ensuring that there is no way to predict the outcome of the game and that everything is fair.
http://www.gaminglabs.com
We took a tour of their facility several months ago, and they have tons of games in there all partially disassembled.
I don't know if you're aware of how this stuff works, but basically GLI collects money from just about everyone who makes games (maybe over-generalizing a bit) and they have a little cube farm of guys who just do math who pore over the code to determine if there is any weakness and advise accordingly.
There are all sorts of techniques that are used in order to prevent games from being exploitable. And there are all sorts of stories about various games being exploitable due to bad design.
Just plopping in a pseudo-random number generator and sticking in the same seed each time the game was powered on would never work. Some people have a hard time understanding the lengths that companies go through to ensure that there is enough data being shuffled around internally to make sure that everything is fair. But an RNG is just one tiny little piece of the overall puzzle.
Easily, Ahigh, this is your best post. Ever.
Machines and electronic shufflers are pseudo-random. (what other kind of random is there for an electronic RNG?)
Hand shuffles and actual wheels are supposed to be random.
Quote: boymimboBob's logic:
Machines and electronic shufflers are pseudo-random. (what other kind of random is there for an electronic RNG?)
You can use a seed from a source of entropy. Use a typical pRNG. Generate X numbers. Have a source of entropy that generates 0/1's. Take bits of of entropy source, and use that to select which of the X numbers to take.
For example. I have no idea how easy it is to make an entropy source generate 0/1's effectively. But this would add an external source rather than a predictable system. By predictable, meaning if I know state X, I can find state X+1.
Or would that still be a pRNG?
Quote:
Hand shuffles and actual wheels are supposed to be random.
At a guess, and it's a wild arsed one I have done no research on (trying to find a reference paper I read on this, to do my own research), roulette wheels and hand shuffles are more likely to show bias than a pRNG.
Quote: TwirdmanAlso if they are different there has to be some reason they are different this difference from true random can lead to exploitability.
If it can be, I'm not aware of it.
Quote: PandoThen after 6 weeks, the numbers I observed changed, and now they pass
the tests, at least for the moment
Those type of machines use an algorithm
that is much different than even pseudo
random. A couple repair guys figured
out the number pattern of Roulette Evolution
a couple years ago and won $300K before
the casino shut them down.
Quote: AxelWolfIf you can accurately detect and tell the differences between a normal gaming machine RNG apart from live random, you will make money from this in a variety of ways.
You keep saying that but have yet to name
one. Go ahead, list just one from the variety.
Quote: thecesspitAt a guess, and it's a wild arsed one I have done no research on (trying to find a reference paper I read on this, to do my own research), roulette wheels and hand shuffles are more likely to show bias than a pRNG.
http://statweb.stanford.edu/~cgates/PERSI/papers/aldous86.pdf
This is very well researched, and you're correct about bias being very possible in a shuffle that is less than enough to ensure randomness.
Gambling with an edge talking about biased wheels.
Quote: thecesspitroulette wheels and hand shuffles are more likely to show bias than a pRNG.
That's what I said earlier. pRNG seems to be
more statistically accurate in the short term
than real random.
Quote: EvenBobYou keep saying that but have yet to name
one. Go ahead, list just one from the variety.
Its impossible to say how to exploit it unless you tell us how you are differentiating them.
Quote: EvenBobThat's what I said earlier. pRNG seems to be
more statistically accurate in the short term
than real random.
Listen to the you tube video about roulette bias. Are you saying if you define "real" random to be a roulette wheel that might have bias and pRNG random is more statistically accurate?
pRNG's don't get biased over time. That's one difference between a roulette wheel and a pRNG.
Quote: Ahigh
pRNG's don't get biased over time. That's one difference between a roulette wheel and a pRNG.
I'm not taking about bias. Real honest random against
pRNG. Real random is not as statistically accurate in
the short term as pRNG is. This is why the old timers
rejected it when the virtual machines came out.
Quote: EvenBobThat's what I said earlier. pRNG seems to be
more statistically accurate in the short term
than real random.
Oh and since you just wanted an example. You mentioned statistically accurate in the short term. I have absolutely no idea exactly what this is supposed to mean but I infer from the way you are talking that you think there is not the level of streaks and things are bunching up more towards expected number of outcomes then you would see in a real roulette wheel. This is easy enough to exploit if you see that one number is fairly underrepresented in the short term you bet those numbers. According to you the pRNG is compensating for their low values so they should be more likely to come up.
number of times eventually. Law of Series (yes,
it's not really a law) holds up in the long run
also. In the short term, it's all over the place.
You get three 14's in a row, number 34 doesn't
show up for 140 spins, you get 18 black
numbers in a row. You'll see this on random
and pRNG results. But if you play long enough,
pRNG seems to behave itself better. Nothing
is predictable or exploitable, just different.
It's all in the nuance. Like a pro golfer can
tell how far the ball went just by the sound
of the club hitting the ball. If they're good,
that is.
Quote: EvenBobStatistically, all the numbers show up an equal
number of times eventually. Law of Series (yes,
it's not really a law) holds up in the long run
also. In the short term, it's all over the place.
You get three 14's in a row, number 34 doesn't
show up for 140 spins, you get 18 black
numbers in a row. You'll see this on random
and pRNG results. But if you play long enough,
pRNG seems to behave itself better. Nothing
is predictable or exploitable, just different.
It's all in the nuance. Like a pro golfer can
tell how far the ball went just by the sound
of the club hitting the ball. If they're good,
that is.
So your peddling unquantifiable gut feeling. Yeah pardon me for being a bit incredulous but I'm going to continue saying you cannot tell pRNG from roulette wheel unless you can offer some more proof then "seems to behave itself better". And again you wouldn't need to know with certainty just be slightly more sure than chance.
Quote: TwirdmanYeah pardon me for being a bit incredulous.
Yet you can say they're the same with zero
proof. You just 'know' they are. Prove it.
