RNG vs Real Wheel (RW)

[Steve]

Start with link:s://en.m.wikipedia.org/wiki/Randomness_tests

Theres much more to it specific to roulette. Identifying statistical anomalies is exactly what my non-computer system does, but part of my approach is cross referencing as it allows better analysis with less data. Ive been doing that for roughly 15 years. You also need to know the right variables, and it helps to understand their physical relationship to each other, so you know what to look for.

If you're working with large volume data, it's easy. Use a bitmap generator like random.org explains. But it's not practical for low volume data. Still if you have enough data for the same wheel and ball, and use the right variables, even low volume bitmaps wil show anomalies.

You're more on the right track than anyone has been in a while

[Badger]

You get different types of random number generators.
The following algorithms are pseudorandom number generators:
•   Blum Blum Shub
•   Blumâ€"Micali algorithm
•   ChaCha20 â€" some implementations of Google's TLS
•   Complementary-multiply-with-carry
•   Counter-based random number generator (CBRNG)
•   Fortuna (PRNG) â€" Yarrow refinement
•   Inversive congruential generator
•   ISAAC (cipher)
•   KISS (algorithm)
•   Lagged Fibonacci generator
•   Linear congruential generator - of historical importance
•   Linear feedback shift register
•   Maximal periodic reciprocals
•   Mersenne Twister
•   Middle-square method
•   MIXMAX generator
•   Multiply-with-carry
•   Naorâ€"Reingold pseudorandom function
•   Parkâ€"Miller random number generator
•   permuted congruential generator â€" small/fast; passes BigCrush with just 36 bits of state
•   RC4 PRGA
•   Well equidistributed long-period linear
•   Wichmannâ€"Hill
•   Xorshift
•   Xoroshiro128+
•   Yarrow

(Sorry. I copied and pasted from Wikipedia.)
A roulette wheel would fall under the type, linear congruential generator.

The generator is defined by  Xn+1=(aXn+c) mod m

where X is the sequence of random values, and
         mâ€" the modulus
         aâ€" the "multiplier"
         câ€" the "increment"
         Xnâ€" the "seed" or "start value"

I am no expert in the above, but think of Xn+1 as the next random number,
Xn as the last number spun.
a as the force with which the ball is launched
Not sure about c
Modm is a mathematical function that keeps the numbers in the 0-36 range


I don’t think that the generation of a number is a problem, but the program associated with RNG can be used to cheat a roulette player, or manipulate RNG numbers to play against a player.

[Joe]

Thanks for the replies, guys.

You also need to know the right variables,

Any hints on what these right variables are? I assume your system only applies to real wheels?

Maybe the OP's question doesn't mean much since there are so many kinds of RNG, but let's assume that an online casino is going to use a high quality RNG because it's in their interest for it not to be biased. To keep it concrete, suppose I provide you with 2 files containing say 5000 spins each, one RNG and one from an unbiased wheel. Would you be able to say which was which? and what would you look at in order to find out?
This just looking at the raw numbers and you don't have any other information about dealer changes, spin direction or any physics based variables.
Hopefully the Generally will turn up and give us the benefit of his experience too.

[Joe]

Hopefully the Generally will turn up

Generally?  I mean The General.

You're more on the right track than anyone has been in a while

Glad to hear it.

[The General]

Live verses RNG

The differences will sometimes show in as little as a few thousand spins.  By 50k spins most people can differentiate the two wheels a high percentage of the time when viewing graphs representing the total number of hits for each number as they lay on the wheel.

As the two data sets grow, the chi squares diverge from each other. 
On the live wheel after several thousand spins the chi could be well over a hundred or even a few hundred.
On the RNG the chi may move up and down a little but it will usually remain below 50 or 60 on the high side.

Few people realize the above because very few people have ever viewed such spin counts from live wheels. 

[RouletteGhost]

What does that mean exactly ?

I mean I know it proves RNG is different than real life but what does what you say mean exactly

[The General]

In slang, using incorrect terms and an oxymoron...

Live wheels aren't random.  All wheels are biased to some degree.

And RNG wheels are toooo random. 

Most system players have no idea as to much difference exists because...
1. System players often have a very short attention span and are distracted by shiny objects.
2. None of them have ever tracked a live wheel to a statistically relevant number of spins and live spin data just isn't readily available.
3. They usually have no comprehension as to what's statistically relevant and what's not.

[RouletteGhost]

1. System players often have a very short attention span and are distracted by shiny objects.
2. None of them have ever tracked a live wheel to a statistically relevant number of spins and live spin data just isn't readily available.
3. They usually have no comprehension as to what's statistically relevant and what's not.

your points here, labeled 1, 2, and 3 did not help to answer my question at all and is mainly irrelevant to this thread

moving on now, so live wheels all have some sort of bias due to being man made where as RNG has different patterns because its "too" random

are the RNGs rigged?

[The General]

your points here, labeled 1, 2, and 3 did not help to answer my question at all and is mainly irrelevant to this thread

moving on now, so live wheels all have some sort of bias due to being man made where as RNG has different patterns because its "too" random

are the RNGs rigged?

Ghost,

You shouldn't worry whether it is or is not rigged because for you it wouldn't matter.

[RouletteGhost]

so using the chi square you find that RNG behaves different from live wheels

which I have been saying for years....your data shows it

this is great information

[The General]

so using the chi square you find that RNG behaves different from live wheels

which I have been saying for years....your data shows it

this is great information

Ghost,

How do you feel that it will in any way benefit you???

Be careful with chi square.  You'll put your eye out if you're not careful.

[Steve]

Any hints on what these right variables are?

Example: You might consider one half of the wheel wins above norm. While this can happen by coincidence with RNG, its more likely to occur with a real wheel if there's a legitimate bias. There are lots of other things you can look at that have nothing to do with bias, and what you'll find is some things are much more likely to happen on real wheels. You can never be 100% certain, although you don't need to be.

To keep it concrete, suppose I provide you with 2 files containing say 5000 spins each, one RNG and one from an unbiased wheel. Would you be able to say which was which?

Yes probably. But to rely on only spins, you need lots of them. And they need to be from the same wheel and same ball type, or it gets more complicated. To make it practical, you need much fewer spins (hundreds), and more surrounding data. If you don't know how to correlate the data and what to look for, you wont find it unless you run huge tests with permutations (like picking an electronic lock), which takes a long time. My earlier software did billions of calculations for around 300 spins which took a long time. But you can take shortcuts if you model the relationships of variables, the physics etc.

This just looking at the raw numbers and you don't have any other information about dealer changes, spin direction or any physics based variables.

Those are the obvious variables. There are many more but if you include too many, you may increase accuracy but at the cost of a lot more analysis needed.

[Joe]

Thanks Steve & General. I have a large file of spins all from the same table so I will do some analysis and report back soon with my results. I will use spins from random.org for comparison.

[Joe]

Live wheels aren't random.  All wheels are biased to some degree.

Biased to a degree that you can take advantage but not so biased that the casino notices?
BTW, was this you, or perhaps Steve?

In the early 90s, Hobart casino was hit by biased wheel players. The result: roulette virtually closed down for a time, with a $10 MAXIMUM bet on a number, believe it or not. All wheels in Hobart and Launceston were soon replaced with Huxley Mk. 6 adjustable wheels.

link:://glind.customer.netspace.net.au/gambling/biased.html

[The General]

Biased to a degree that you can take advantage but not so biased that the casino notices?
BTW, was this you, or perhaps Steve?
link:://glind.customer.netspace.net.au/gambling/biased.html

Most are not biased enough, but some are biased enough to get a good edge.

Some casinos notice, some don't.  Contrary to popular believe casinos don't remove wheels just because they're biased.  Wheels are expensive, and biased wheels usually make just as much as the more random wheels.  Most people on this board wouldn't be able to beat a biased wheel if you told them the wheel was biased.   If a wheel is very biased, it's sometimes just moved around the floor more frequently or kept in the "evening and weekend" positions.   Periodically rotating wheels around the casinos floor is a common practice and when it happens the data download boards usually stay in position even though the wheels have been moved.  This corrupts the data download because it mixes the data with spins from the newest wheel.   

If you're planning on using data downloads from Roulette Extreme you should know that the data is also corrupt.  For example, the Wiesbaden  table one data is a mix of several different wheels over a period of time.  You won't have any way of knowing when the wheels were changed by just looking at the data.

I may post a biased wheel number stream in segments on the forum for people to try and play, just so they can see how well their systems perform.