Intresting thought about the deviations

[Master_of_pockets]

Marigny system is trying to bet on the correction of a big deviation.
But the main problem with these kind of systems is that we never know IF this deviation that we see in the short term is a real deviation....because this short term deviation that we see ,it may happened because the opposite event had a big deviation before and this new deviation that we see may be is the correction of the previous opposite deviation....this is why this kind of playing isn t so stable as it should be....because we never know what happened in the previous spins.....

So roulette is a random generator that is making deviations all the time in order to correct previous deviations(in my opinion)

So what about instead of betting for the deviation to balance ,to bet for the deviation to grow...in this case we are always on the right track of correcting previous deviations that we don t even need to know when happened.

Opinions?

Also I need to ask ....
Why some ppl think that the deviation in events (like RRR and RRB etc.) have bigger success then just the amount of the Ecs(eg R and B)?

[GARNabby]

Which came first: the wheel, the spin... or "silly guy" Twain?

[Nickmsi]

 I think betting for sequences to continue is betting against the standard deviation and that doesn't make sense.  Perhaps if the number of sequences is low (less than 5) it might be worth a shot,  but as the sequences increase in length(Standard Deviation increases) then no, I don't see it.


Why not use the quantity of even chances rather than their sequences?
It may take thousands of spins to correct a numerical deviation in Red/Black or other even chances.
At spin 100 you may have 60 Blacks and 40 Reds so one would think betting on Red to show up is a good bet. Not so, at spin 1000, you may have 520 Blacks and 480 Reds, which is closer to balance but still not there.


Using sequences to bet, you generally would need only 9-36 consecutive sequences to formulate a bet with a good standard deviation score.


I am not a math person, but that's my two cents.


Nick

[Ralph]

They equal out, but it is difficult if not impossible to use it.

In 100 spins the the chances can have more spreed in absolute numbersand in %.

In 10000 spins the chances even out i % of the share, but it is not so sure  in absolute numbers.

In 10000000 of spins it may be a diff. of 1% but still in absolute number more than in 100 or
1000 spins.

[Bayes]

Marigny system is trying to bet on the correction of a big deviation.
But the main problem with these kind of systems is that we never know IF this deviation that we see in the short term is a real deviation....because this short term deviation that we see ,it may happened because the opposite event had a big deviation before and this new deviation that we see may be is the correction of the previous opposite deviation....this is why this kind of playing isn t so stable as it should be....because we never know what happened in the previous spins.....

By that logic, if you knew the entire history of a wheel you should be able to find out whether the current deviation is reliable or not, except of course that would be silly.

In reality, there is no "correction", that's not the point of looking for a deviation. The most likely outcome in a sequence of X spins is the average which is +/- 1 STD. That will occur whether or not the previous sequence of X spins is 3 STD. The point of the trigger is that it's less likely that you'll encounter the run from hell (because it's just occurred). It doesn't guarantee anything but it's better than "cold" betting, and you need to look for signs of "correction" before actually betting, as Ego has indicated in his many posts on the topic.

[Nickmsi]

Bayes . . 

Help me out here. . .

I read somewhere that if you have a 3 STD that you would have a 15% statistical advantage and if you had a 6 STD you would likely be permanently winning.

Is there such a chart or formula that tells you what statistical advantage you would have as the STD increases?

Thanks for your help . . . Nick

[Colbster]

1 standard deviation will cover 68% of results, 2 SD will cover 95% and 3 will cover 99.7%.  Beyond 3 SD, you are getting into the realm of tracking time spent not justifying the betting opportunities because it would be so rare.  6 SD is just wildly out of the question unless the game is artificially fixed.

[Nickmsi]

Bingo Colby . . . you are right that tracking 3 STD bets would not be justified because they are so few.

In my testing, I have found only about 600-700 betting opportunities per 100,000 spins.

Not for me, but for a bot might be worthwhile, especially if you can get a statistical advantage.

Nick

[Colbster]

That is higher than we would expect.  You should only get 300 opportunities over 100,000 with 3 STD bets.

[Nickmsi]

It's not always possible to get a bet with exactly 3 STD when dealing with sequences, as some of the "series" type sequences get quite long, so there can be a large margin of error.

It's fascinating to study deviations with sequences rather than numbers as it appears they may apply to not only roulette and other even chance games and/or perhaps to any game with a W/L registry than can be sequenced.

Much to do, so little time . . .

Nick

[vladir]

That is higher than we would expect.  You should only get 300 opportunities over 100,000 with 3 STD bets.

That is tracking one kind of event correct? I aask because we can track several different kinds of events... and if for each we get 300 opportunities in 100.000 bets ... With just 10 differnt events, we woudl be playing on average 3% of the time...

[Ralph]

Some events are overlapping

[Bayes]

Bayes . . 

Help me out here. . .

I read somewhere that if you have a 3 STD that you would have a 15% statistical advantage and if you had a 6 STD you would likely be permanently winning.

Is there such a chart or formula that tells you what statistical advantage you would have as the STD increases?

Thanks for your help . . . Nick

Hi Nickmsi,

I'm not aware of any chart or formula, but it wouldn't be too hard to create one. Personally I don't see much use in making that relationship explicit; If you have a true mathematical edge the STD will keep increasing as you make more bets. Suppose instead of the usual -2.7% you have +2%, then every 100 bets you will have made 2 units instead of losing 2.7. You could plug those values into the z-score formula to find the relationship between STD and advantage, or you could work the other way and take the expected value of p (probability), plug that into the z-score formula and then set this equal to the formula with an assumed unknown value of p. Then solve for p and use that to calculate the advantage.

Bit busy at the moment but let me know if you want the details fleshed out.

[Nickmsi]

Thanks Bayes for the information.

I was just curious where the 15% advantage came from.  I ran across it in one of the translations from Italian to English when researching Marigny and often the translation are not clear.

Cheers . .  Nick