Pattern recognition, sequences in random numbers.

[Chance]

I wanted to start a discussion on a way i discovered how to recognize if the random numbers are falling in a truely random order.  And, sequences that you will always see in a random stream. How you can use this to your advantage. This will take some time to explain. I am not so conceited to think you will stop what you are doing to learn this, but hope you will consider what i write. This will take patience, but it is simple to track with pencil and paper. More in a bit.

Chance

[Chance]

Very few things are random.  Most rngs are really prngs even rolling dice is now obsolete  thanks to the combined effort of two universities. Most rngs use methods of things like time between key strokes x 1000 to come up with a number added to some seed. But, what about craps, cards, and my favorite roulette. Lets first look at what is required to make a random number random. There are really three things in my book. Unpredictability of the following number. Even spread of the numbers (yes i know about the law of the 3rd). The numbers must land in a non sequential pattern (random numbers must land in a random fashion within the random stream). This last item is what we can exploit to our advantage, but to do this we have to look at the first 2 items. I know your saying cut to the chase but please bear with me..........more in a bit

[Chance]

Random numbers have to have an equal spread (yes, i know this is so basic for most everyone). In six rolles if a dice the numbers 1 through 6 should land equally, But the next item is why they don't and we have things like the law of the 3rd and other things. The next says that for every n samples (spins, tosses, cards played) an exponent will be a double, an eponent of that will be a triple and so on.

Etc

....

[Chance]

These runs or series make a stream random,  and can be played to your advantage, but like i found out the hard way many years ago, they are not enough by themselves. I wanted to spot if a dice was loaded, wheel was biased or the person spinning was locked into a rhythm that could be seen easily, but there was no short term ranom test. Here is what i mean by the random numbers have to happen randomly. That sounds like a oxymoron at best. Lets take a dice roll of.

3
2
5
4
3
1
2
3
1
5
5
4
3

Although this appears random it is so much not so that the next roll is obvious, if the sequence keeps repeating......care to guess?

[Drazen]

This suddenly sparked my curiosity.

Just dont ask me why 

[Chance]

Drazen,

Thanks for dropping in. Yes to me this is a fascinating topic and it has a lot of potential not just in gambling. 

[nottophammer]

#1

[Chance]

I understan why you said 1 nottophammer. Let me throw another set at you and i will explain this step by step. Remember we are trying to get: what side of a die is heavier, what pocket or pockets of the wheel has a bias, what cards tend to come up, or if the person has a rhythm going that we can exploit. Here is a harder set of numbers.

5
3
2
4
1
6
2
4
2
4
3
1
4
3
1

At first glance our mind works out that there is a pattern, but who knows what it is. So we run an algorithm , (its simple) i say simple not that you cant handle difficult but at a table you can't divert your attention writing while paying a fast game. After the alg  the numbers above turn into.

3
3
5
5
2
2
4
4
1

So what is the next number in the pattern? Obvios if it continues a 1.

Explanation...

[Chance]

Before i go any further, questions? I know this is very basic. But, this is the first generation i came up with about 20 years ago......once again your a patient lot.

Chance

[Chance]

5 was the first thrown 1 was the last. Btw i do not play craps, only roulette, to adapt this to roulette divide the wheel in 4 sections. There is a better way I'll explain latter. For now this simple example using 1 dice.

You are looking for the first repeat. So you are looking at 532416 no repeats so far but the next one is a 2. First repeat is a 2 beside it you mark the position of its position which is a 3 and cross that off so your sheet looks like this

5
3
2 x
4
1
6
2    3

Next you add your new number, you have already cheked off the 2, next on the list is the 4. The 4 is above the 2 so mark it off and beside the 4 place the position.

5
3
2 x
4 x
1
6
2   3
4   3

Then add the next number. Place your x and its poition.

5
3
2 x
4 x
1
6
2 x 3
4   3
2   5

Continue this with all the rest till you have

5 x
3 x
2 x
4 x
1 x
6
2 x 3
4 x 3
2    5
4 x 5
3 x 2
1    2
4    4
3    4
1    1

So now we will either bet the series will continue or bust. Let's say we know the die will continue and expect the next sieries in our alg will be a 1. To work backward is easy just look at the first number that does nit have an x...which is a 6....

Questions so far?

Actually a real die would not give you a sieries like this if it were out of ballance  or a wheel. I'll give real examples to help you turn this into a working method of play to assist your game.

[plolp]

Then add the next number. Place your x and its poition.

5
3
2 x
4 x
1
6
2 x 3
4   3.............why 3

2   5  ...................why 5 ?

I do not understand the numbers you write next .

[Steve]

Thanks for sharing but the problem with this approach is you can never know WHICH numbers will be repeats, sleepers or whatever.

Its the same kind of problem with law of a third. You might know almost always 1/3rd of numbers will be repeated in 37 spins. But you dont know which numbers with any better than 1 in 37 accuracy.

It's like saying that over 1,000 spins there will be roughly an even spread between red and black. If you wait for an imbalance of say 50 red and 20 black, then what if the last 1000 spins were 600 black and 400 red? But either way, the imbalance tells us nothing, except for the possibility of bias (but that's another matter).

Dont get me wrong. I agree with parts you said. RNG is definitely not random. Nothing is ever random. I DO believe it is possible to predict RNG one way or another, but it cant be done with methods like law of a third.

Ages ago I had software created that automatically checked for discrepancies in spins. Basically to highlight anything that was not expected for truly "random" spins. It looked at things like:

1. If red won for 10 times in a row, what are the odds that red or black spin next

2. If a number repeated in the last 1,2,3,4,5 etc spins, what are the chances that it will spin again in the next 1,2,3,4,5 etc spins.

The program does just basic checks, but over a large scale. What I ultimately found was:

1. There are indeed patterns in RNG that are predictable. But they are predictable in the long term. I could not find any way to change the odds on the next spin, OR any group of spins.

I havent had time to fully understand what I found. But others find much the same thing. Here is an example:

2. The patterns that I found could be used existed only in real spins, and only when I had specific data included with it (not just last winning numbers)

Can can we beat RNG? I think so yes, but it cant be done with methods like law of a third - at least in the manner most players try to use it.

[nottophammer]

I do not understand the numbers you write next .

I'm with plop, unsure, but always worth a look

[Steve]

And something else..

The static on an old TV set is the kind of "random" used by a regulated online casino. The kind of "random" used by random.org.

But the static on a tv set is not random. And if you film and process video, looking at key frames with particular intervals, you see fractals everywhere.

I have no doubt that the real key to "predicting random" is fractals.

Any advantage player would say it's all "cause and effect". Yes that's right. But a fractal is also cause and effect.

Again I haven't had time to pursue it further. My focus is on more traditional approaches, but also incorporating a similar "pattern/anomaly-finding" algorithm for real wheels.

My gut feeling with the RNG fractal stuff is you will probably only be able to marginally reduce the casino's edge against you, but not overcome it and get a positive edge.

[Chance]

I understand fractals Steve and how they apply, but by themselves i think will lead to failure. True it is so so much more than repeats  runs, sleepers, and hot numbers. Mt best success has been with quickly identifying, sequences, and biases. The odds of a fractal happening is exactly equal to the sum of the smaller fractal that have happened its a 50 50 shot. By themselves i can't see success.