Random.org vs actuals

[Turner]

What do people think about random.org numbers?
RX wheel analysis Passes them as if its a balanced wheel

If I gave 280 Random.org numbers to someone who checks wheel balance at the casino, would he say "this wheel is balanced" or "these are random numbers off random.org"

[VLS]

Hi Turner.


In my views, 280 spins aren't enough for claiming a balanced or imbalanced wheel.


You may want to try adding more zeroes to the right of that figure


You know what they say: in the short term "ANYTHING CAN HAPPEN" and 280 spins is considered a rather short term.


Regards,
Vic

[Skakus]

 
RX will pass any amount of random.org numbers.

Thay are random through and through.

[MrJ]

I've said it before........If I lined up 800 actuals and 800 random.org numbers, will you be able to tell me which is which (and why) without guessing?

Ken

[Turner]

RX will pass any amount of random.org numbers.

Thay are random through and through.

So, If I test a system with random.org numbers, how far out am I with my conclusions in the eyes of the "got to test using actuals" crew, and, how far out am I in the real world of random.

I mean, I wouldnt want to use actuals from a wonky wheel would I?

[Turner]

I've said it before........If I lined up 800 actuals and 800 random.org numbers, will you be able to tell me which is which (and why) without guessing?

Ken

This is how i envisaged this post going...and I was hoping "for" random.org as a perfectly acceptable roulette test "actual"

[Skakus]

 
Somewhere on this forum Bayes posted 4400 numbers (horror sessions) made up of 22 x 200 actuals from different real wheel  sessions.  Each of these sessions were supposed to be difficult to win when betting on the specified  EC for each session.  This was basically because the random dispersion in each session went against the specified EC for that session.

If you run each of the 200 number blocks through RX some of them will actually pass, and some will fail. If you run the whole 4400 through RX the result is a fail.

What this means is that even fair wheels will from time to time throw up results that appear heavily bias.

The funny part is that once a wheel has produced a bias large enough to fail the fair test it then needs to compensate with another relatively opposing bias in order to pass the fair test next time. If you split the two you get two fails but combine the two and you may get a pass.

So even though after long sessions using thousands & thousands of numbers the wheel appears fair, the result is likely to be made up of many periods whereby the wheel was not fair, in fact it could almost be made up of entirely not fair sections all the way through  but by virtue of cancelation end up being fair.

  This behaviour also creates the possibility that a real bias could exist on a wheel with long term fair results because of other random factors disguising its existence. This is one reason why bias spotters say number collection alone is not enough to confirm a bias.


For this reason I believe random.org numbers (they also behave in the same manner) are as good as actuals for testing purposes.

Just my opinion.

[ego]

When i have fun i use random org only as if you beat them you beat any actuals.
I use to download 10.000 trails sampels and 100.000 trails a day.

[Bayes]

Nice post Skakus, and I agree. "Random" is a tricky concept which can lead to contradictions because it seems that in order for a number sequence to be random, it must also exhibit some regularity at times (ie; be apparently non-random).

[superman]

For this reason I believe random.org numbers (they also behave in the same manner) are as good as actuals for testing purposes

Totally agree, it doesn't matter where your numbers come from, in the long term the up/downs will be the same.

BBBBBBBRRBBBBBBBBB = unfair bias towards black on a short sample, but what happened before that run, and whats going to happen after it, you'll probably find it equals out eventually.

[frost]

Totally agree, it doesn't matter where your numbers come from, in the long term the up/downs will be the same.

BBBBBBBRRBBBBBBBBB = unfair bias towards black on a short sample, but what happened before that run, and what's going to happen after it, you'll probably find it equals out eventually.

so im curious. if thats the case does that mean we will never find a system the will win continuously?

[Turner]

Totally agree, it doesn't matter where your numbers come from, in the long term the up/downs will be the same.

BBBBBBBRRBBBBBBBBB = unfair bias towards black on a short sample, but what happened before that run, and what's going to happen after it, you'll probably find it equals out eventually.

But doesn't this make you think that if we are accepting random.org as true roulette numbers in essence, then we have to notice that random.org doesn't paint them red/black/green or group them nicely into patterns on a baize table.

BBBBBBBRRBBBBBBBBB could be:
31,28,4,4,15,4,35,1,36,26,29,17,13,13,13,15,29,24
or
28,4,6,2,8,20,20,14,34,17,4,33,22,17,11,15,20,20

We see a pattern in BRB because we love patterns. The 2 sets of numbers don't corralate to each other in any way at all.

Bt the way Skakus, your post was a worthy read!

[Bayes]

so I'm curious. if that's the case does that mean we will never find a system the will win continuously?

There's a principle called the impossibility of a gambling system, which says that it's impossible to select any sub-sequence from a random sequence in a way which will improve the odds of winning in the long run. Note that there is no mathematical proof of this, it seems to just be true in repeated experiments (read the forums!). 

[nitrix]

Here's my little contribution about random:

Imagine you walk by a random generator producing "8, 8, 8, 8, 8, 8...."

Somebody could say "Hey ! This thing isn't random at all !", but in fact it is, and you simply happened to be there in the right moment where those numbers spun.

Actually, most events you'd think are "rare" or very unlikely to happens have their equal opposites with the same odds. I'm also a believer (like said previously above), that it'll balance out in the long run, but in the short run, what you'll see is a multitude of "bias" one after another...

IF you can ever make money off a small streak, well a repetition will hurt you. If you exploit repetitions, streaks will kill you. It's one of the other, you can't really have both.

At least from what I learned so far.

[darrnyf]

as bayes has been one of the best mathematician....i would like to ask..then how cum people say that rng cheats..but in reality it doesnt..its random fucking u ..not rng