True/Fair Randomness

[andreib1986]

A want to start a more filosofical idea, that a fair randomness could be expalined like: "To test a true randomness one must evaluate that all spins are connected with each other in order meet standard variations but there must be no conection beetween the spins" Is this nonsens?

[Still]

A want to start a more filosofical idea, that a fair randomness could be expalined like: "To test a true randomness one must evaluate that all spins are connected with each other in order meet standard variations but there must be no conection beetween the spins" Is this nonsens?

Yah, collect enough spins and we get something called a "bell curve".  While the wheel is said to have no memory of the last spin, the bell curve seems to have a memory.  Question is, is there a way to exploit the bell curve?  Is there any way to exploit reversion to the mean (reversion to the bell curve)?  This is proving very difficult.   Filosofical?  The devil is that which is difficult.

[Steve]

The bell curve is not memory. It is normal distribution. You can no more make use of it than knowing there will be around even reds/blacks.

You cant bet on balance. It changes nothing. The basics still arent being understood.

[Firefox]

No way to exploit a normal distribution.

The standard way of "exploiting" seems to be to wait 10-30  spins and then predict numbers to repeat or catch up.

Over say 50 spins SD is about 1, so a number could ocurrs 4 times or not at all. All within 3SD of mean which is 95% confidence.

Push it up to 1 million spins and hoping for a sleeper to catch up the SD is in 100's so it can easily remain a few hundred behind and that is just normal with reasonably high chance.

So either way, impossible to exploit.

[Anastasius]

If the amount of uniques is say 15 minimum and u miss the first 2 thats 13 definate wins in 37 spins. Cant there just be a progression to ensure ur in + at the end of set.

[Steve]

Anas, that bet selection has exactly the same accuracy as random bets. So what's the point of it?

[zhone]

Anas, that bet selection has exactly the same accuracy as random bets. So what's the point of it?

Tested with 800,000 spins, with more than 4000 qualified spins, we could consistently get an edge of 4% (on top house edge) if we were to bet for a repeat after 18 numbers have been covered. Has this been tested in the past?

[Steve]

So you bet on 4000 spins? Thats nowhere near enough.

[andreib1986]

I've seen a news recently that scienticts are working on quantum computers wich are now in the same stage as early 60-70 computers and they could be able to anticipate patterns and predictions for desies etc. I wonder if they would give a mathematical advanvange for roulette just by memorizing the rulete 'debt' in patterns. You dont need a quantum to do that but...im kinda fascinated by the idea.

[Steve]

If you mean looking for patterns based on whatever variables, even if they seem irrelevant, to see what works, i did that. Would have been a lot easier with a quantum computer.

[andreib1986]

And still no advantage found? That is depressing...Is like nothing makes sense and God has disapeared when he made the world. And what Einstein said that the only way to win in the casino is to steal, might remain true forever.

[Steve]

Yes there was advantage, but not in trng. Only in low quality prng and real spins.

[andreib1986]

Can you give some clues about the advantage?

[Steve]

With rng, flaws are easier to see if you use a visual bitmap chart. Random.org explains it. Also theres free source code elsewhere to test rng quality.

For real wheels, use real physical variables.

Forget nonsense like repeaters, balance, all the typical fallacies etc. Theyre useless.

If you want to play rng, attack the rng integrity.

[Anastasius]

What did u test zhone wait 18 then bet just 18 for rest of set