+ Register

[Dane]

I must be cutting my hair too short, the last few posts have gone straight over my head.  What are you lot on about here?

Basic fact: The 18  known red numbers is JUST ONE OUT OF MANY
possible combinations with 18 numbers out of 37.
A combination calculator reveals that C(n, r) = 17672631900.
One of those combinations is much too common.
So I don´t think that another combination corresponds to such a long red series.

I AM NOT AFRAID TO LOOK AT THE COLOURS 
And with the Permanenzen from table 3, Spielbank Wiesbaden it is easy to SEE the  coloured series!
Much easier than counting spins in any combination independent of the colours
and jumping to conclusions.
                                                       [smiley]aes/coffee.png[/smiley]

[plolp]

say it another way:
36 times the same EC happens very often, but these series appear in disguise.
18 numbers = 18 numbers
It's a fact .

[Dane]

say it another way:
36 times the same EC happens very often, but these series appear in disguise.
18 numbers = 18 numbers
It's a fact .

I see.
In order to see what could happen
inside one of the known  boxes of Even Chances
we simply have to imagine 17672631894 invisible boxes in disguise.
In this way we have almost TRANSCENDED EC!
In a way we might as well have entered the research of PLEIN (or Straight Up)
and repeated this old question: Why has noone seen all 37 numbers in 37 spins?

[Slacker]

The problem with this approach is that you come up against the ever-receding horizon of probability. For example, suppose I reason like this: if there were 30 reds in a row, I might justifiably think that it would worth betting on black for the next X spins, due to regression to the mean. But since I don't want to hang around for years waiting for 30 reds in a row, and I know that any bet selection is as good as any other, in terms of the distribution of wins and losses, then would it not be reasonable to just look at the last 30 outcomes and bet the opposite?

After all, if my bet selection was the opposite of what the last 30 happened to be, and I was just betting this sequence over and over, then the last 30 outcomes would represent a loss of 30 bets in a row. Wow! that means I'm bound to get a lot of winners (due to regression to the mean) if I bet the opposite, and no waiting required!

Unfortunately, this doesn't work, because probability really has no limits if you look at it in this way. So why hasn't anyone ever seen 50 or 100 reds in a row? Because although it's not impossible, it would require one particular predetermined sequence (50 reds in a row) to occur out of the 2⁵⁰ = 11259 x 10¹⁵ possible permutations of 50 outcomes, and that's really really big number.

The basic idea (and this is something a lot of people have trouble grasping) is that something predetermined is less likely to occur than something which isn't. An example of this is the "law of the third", which most roulette enthusiasts know about. It says that in 37 spins, approximately 12-13 numbers won't show up. But think about this, what's the record "sleep" for a standard dozen on the layout? it's around 35 spins! that means  it's a routine occurrence for at least 12 numbers not to show up in 37 spins! The difference is, you don't know in advance what those numbers will be, but again, there are an awful lot of ways of picking 12-13 numbers out of 37, so although it's extremely unlikely that a particular set of 12 numbers (in this case, a standard dozen) will sleep that long, it's almost guaranteed that some set of 12 numbers will.

So viewed in this light, betting the opposite of the last 30 R/B doesn't seem like such a great idea after all. And anyway, why stop at the last 30? why not pick the last 100 to bet against? hopefully you can see the error more clearly now...

[Drazen]

The problem with this approach is that you come up against the ever-receding horizon of probability. For example, suppose I reason like this: if there were 30 reds in a row, I might justifiably think that it would worth betting on black for the next X spins, due to regression to the mean. But since I don't want to hang around for years waiting for 30 reds in a row, and I know that any bet selection is as good as any other, in terms of the distribution of wins and losses, then would it not be reasonable to just look at the last 30 outcomes and bet the opposite?

After all, if my bet selection was the opposite of what the last 30 happened to be, and I was just betting this sequence over and over, then the last 30 outcomes would represent a loss of 30 bets in a row. Wow! that means I'm bound to get a lot of winners (due to regression to the mean) if I bet the opposite, and no waiting required!

You are wrong here. The error is in your understanding I am afraid. I see you have some statistical knowledge how to calcualte probability, but it is obvious that you don't understand probability of sequences in full.

Having 30 R-s in a row and betting the opposite of the last 30 EC outcomes can't be the same thing. Why?

EC sequnce has actually much more dimensions to it. As we know by the law of probability each length serie is twice less likely as the one before and that must correspond in the long run. So serie of 2 is twice less likely as a single, serie of 3 is twice less likely as serie of 2 and so on... So sequnce isn't made just of R/B but also out of their distribution too. So in other words we can take singles and series as independent bet too and still it will correspond to the same probabilties as just colours... But it is important to observe it correct, so to understand how sequnce is grouped and of course not all sequnces have same statistical value. This is where your error comes from and that way you see using any length series is pointless...

For example we have sequence:

RRRRRRRRRR

If I asked you what is the statistical value of this sequence you would say it is a z-score of 3.0 on a single 0 wheel. 

So now I tell you that we can have same length sequence value but we must use both colours. Possible?

RBRBRBRBRB

Both examples above have same value and it will show equal amount of times in the long run. Be in no doubt! [smiley]afb/secret.gif[/smiley]

You also mentioned regression toward mean. Interesting phenomenon in statistics which if used and understood right can give you what we all search(ed) for... [smiley]afb/pray.gif[/smiley]

Cheers

[Slacker]

Hi Drazen,

I think there's some misunderstanding here.

Having 30 R-s in a row and betting the opposite of the last 30 EC outcomes can't be the same thing

It is if you're looking at the sequence, not in terms of the no. of B versus the no. of R, but as a sequence of R/B in a specific order. You can't compare apples with oranges, which is what you're trying to do, I think. And the law of series is irrelevant here because we're talking about sequences of the same length.

This is the point I was trying to make in another post. Sequences of the same length have the same chance of occurring in one respect, but not in another.

RRRRRRRRRR has the same chance of hitting as RBRBRBRBRB when viewed as an ordered sequence, but the latter does not have a z-score of 3.0, because the z-score refers to the number of B vs R. We're talking about two different distributions here, the binomial and the uniform.

The point I was trying to make was that betting the opposite of the last X spins does not give you easy access to a "rare" event, because you have not predetermined that event - the table has chosen it for you. And, strictly speaking, it doesn't matter either way; if you wait for a predetermined event to show up you don't actually have any better chance that the sequence will not repeat than if you were betting the opposite of the last.

[Drazen]

Hi Drazen,
RRRRRRRRRR has the same chance of hitting as RBRBRBRBRB when viewed as an ordered sequence, but the latter does not have a z-score of 3.0, because the z-score refers to the number of B vs R.

Well, still you are wrong and I am right  [smiley]aes/tongue.png[/smiley] what else I can say except explanation I already gave. Series and singles have same correlation like R vs B and that is provable my friend. Famous Marigny de Grilleau has done lot of research on this subject in the past, and respected roulette boards members like Ego, Bayes, Alberto Jonas, and myself studied this for a countless hours too...

There is no "uniform" distribution in this game. This game is all about statistics so only binomial distribution exists in this case.

ONLY point of waiting for the rare events is that they usually don't happen in successive samples, and that is all what regression toward mean says! Waiting for 10, 20 or 30 EC-s in a row doesn't mean that after one point, the reverse must happen to "catch up" that deviation in the short run. It only says that after strong deviation it is less likely that same deviation will repeat immediately again, and that is what we can use to profit with some mild progression and carefull money management.

Waiting for a strong deviation doesn't gives us advantage in terms of higher hit rate then expected, but it certainly lowers down deviations and we can be "safer" to use progression which can't go beyond some point and profit from it. But entry and exit points are crucial too!

Best

Drazen

[Slacker]

Drazen,

I wish I knew what I was wrong about. What exactly are you saying here? that waiting for a rare event is superior to betting the opposite of the last X spins?

We are fortunate that, unlike some forums which discuss politics or religion, in this subject we don't have to rely on good arguments; we can do empirical testing, too. 

respected roulette boards members like Ego, Bayes, Alberto Jonas, and myself studied this for a countless hours too...

Ah yes, Bayes. I have a lot of respect for the man.

There is no "uniform" distribution in this game. This game is all about statistics so only binomial distribution exists in this case.

I beg to differ.

http://en.wikipedia.org/wiki/Uniform_distribution_%28discrete%29

[Leapyfrog]

Ah yes, Bayes. I have a lot of respect for the man.

[smiley]aes/beer.png[/smiley]

[Drazen]

I wish I knew what I was wrong about. What exactly are you saying here? that waiting for a rare event is superior to betting the opposite of the last X spins?

You were wrong when you said that series vs singles don't have same correlation like R vs B.

Waiting for a rare event can be superior in terms of lower variance after some point (not in terms of higher strike rate). And to capitalize that you need good entry and exit points and most of all, good money management and of course corresponding bank.

If you have lot of respect for Mr. Bayes, how come you don't know that he actually made few brilliant pieces of software with which you can test all I stated above. Also he gave excellent tracker, and described his play and mentioned too that series and singles can be same as R v B when observed correct? His newest unpublished tracker actually tracks different length series as bet selections with deviations... And he is very successful player as we know.

I am very sad that he isn't here, things just got spicy when the only forum I know he was active on, closed... I really miss him and some other guys there too.

Cheers

[Turner]

@Slacker
I recon Bayes could have you in a scrap.

[Rolex-Watch]

@Slacker
I recon Bayes could have you in a scrap.

Definitely

[Slacker]

You were wrong when you said that series vs singles don't have same correlation like R vs B.

Forgive me, but I don't think I actually said that. And Bayes is right, for the record.

the only forum I know he was active on, closed...

Yes, shame about that. Hopefully one day Steph will reopen it. It had the makings of a very good forum.

[Dr. Mabuse]

Who is that Bayes cat?   Never heard  of him .

[Rolex-Watch]

Who is that Bayes cat?   Never heard  of him .

Good analogy, a wandering un-neutered alley cat, liable to go missing for days.  I could do with his opinion on something he contributed