### Author Topic: Regression toward the mean  (Read 13310 times)

0 Members and 2 Guests are viewing this topic.

#### palestis

• Great Contributor
• Posts: 850
• Thanked: 810 times
##### Re: Regression toward the mean
« Reply #15 on: November 11, 2014, 01:33:39 AM »
"Purely and simply, regressing to the mean does not mean any short term variation gets cancelled out, only that it becomes statistically insignificant in percentage terms.

No, that's not correct. Regression to the mean does not mean that - he's talking about the law of large numbers here; that the ratio will approach the mean as you take more samples. Regression to the mean says that if an extreme event occurs, the next event will not be so extreme, it doesn't lump together the two events and say that the combined event will be closer to the mean than the first, although that is a consequence.

If the first event is 10 reds in 50 spins, and the next 50 spins produces 20 reds, then the ratio of reds is 0.2 in the first event and 0.3 for the combined events (20 + 10)/(50 + 50), but it is 0.4 for the 2nd event taken alone.
It's nice to see all the math geniuses arguing about  advanced probability theories, each one having a valid argument.
However you don't need all that to win in roulette. Thankfully things are a lot simpler  than that. All you need is extreme patience. What happens in 1 million spins and what  will happen in the next billion is unimportant. It's what happens in about 30-50  spins average per betting session that will make the difference.  I started going back to a casino 1.5 hrs. from me and have no problem walking away with \$200 every time. In less than 3 hours. Using basic common sense and patience for conditions to form according to the system's guidelines, is all it takes. I didn't notice  any threat form the HE or the gambler's fallacy.  Because I exercise my right to stop if things seem to become unnatural. Winning is the least of my problems. My problem is the looks that I get form the people inside the pit boss area as well as dealers.
As I stand on the side and observe and write, and I only emerge when the opportunity arises and that's for 2-3 spins.  Where all other players are betting non stop as fast as they can extend their hands. No wonder the casinos never lose.
The good news is that they approved 3 casinos in my state and in 2 years they will be up and running. Having the choice within  a short distance is a tremendous advantage. So that I don't have to look like I'm from another planet when I always go to the same small casino.
My advice to aspiring professional roulette players is to exercise  extreme patience. Wait long enough until conditions seem to be ripe for a series of bets. And always use VIRTUAL BET LOSSES. Its advantage cannot be argued and counter argued unless you see it in action at the table. And it works flawlessly. I read all the negative comments in various forums, backed by advanced probability theories, but I don't see them applying in real life. It seems that they are stuck in the theory stage.  I'm still waiting to run into a situation where gambler's fallacy applies. Could it be that patience, sporadic betting when conditions are just right, and virtual bets in between is all it takes to send the HE and the negative expectation to the cleaners?

#### Mike

• Great Contributor
• Posts: 912
• Thanked: 188 times
##### Re: Regression toward the mean
« Reply #16 on: November 11, 2014, 08:22:59 AM »
palestis,

Can you say what these opportunities are which are worth waiting for? I'm happy to be proved wrong, but I never have been so far. :-)

There are no opportunities or events in roulette which indicate other events with certainty, so your system cannot work "flawlessly".  If it seems to, then you have just been lucky so far.

#### palestis

• Great Contributor
• Posts: 850
• Thanked: 810 times
##### Re: Regression toward the mean
« Reply #17 on: November 11, 2014, 09:35:49 AM »
palestis,

Can you say what these opportunities are which are worth waiting for? I'm happy to be proved wrong, but I never have been so far. :-)

There are no opportunities or events in roulette which indicate other events with certainty, so your system cannot work "flawlessly".  If it seems to, then you have just been lucky so far.
I've heard these excuses many times before. That I was so far lucky. I didn't start playing roulette yesterday. I had the same results in the past, years ago and still get the same results, now  that I decided to drive 2 hrs. to get to the casino. I know when winning is due to coincidence or simply good luck. There are many possibilities that opportunities arise in roulette. Great imbalance in the short run is one of them. If you aim to achieve one and only profitable bet. Not 2 not 3 not 10 in a row bets. Just one. And then go thru the painful waiting process again.  Another is placing virtual bets on a 50% chance and lose virtually 3-4 times in a row. Then place 2-3 real bets. Sorry, but guessing randomly red or black I never lost over 6 times in a row. In testing and in real betting. The 18/37 probability after reach loss doesn't seem to apply. Only in theory. Try it yourself and see if 6 random 50% EC guesses can be annulled by 6 opposite results. It's not going to happen, any time soon.
Noticing a clear trend in repeated 2 double streets or 3-4  streets and betting on the  continuation of this trend for 1 more time is yet another.
Trends simply don't seize to continue every time you start betting aiming to win 1 time only. At least not every time a noticeable trend forms.
If it does, you are having an exceptionally bad luck for the day and it's best to leave the casino for that day.
These are just a few of the opportunities that can arise and you can take advantage of them for one time only. More than one time becomes greed. And that's y most systems fail.
In theory anybody can prove me wrong, but when it comes to the real thing, it's only the theory that is proven wrong. Theories fail to incorporate extreme patience as part of the betting process. They only apply to frequently betting, or non stop betting  players. And there are plenty of them around.
« Last Edit: November 11, 2014, 09:38:55 AM by palestis »

#### Mike

• Great Contributor
• Posts: 912
• Thanked: 188 times
##### Re: Regression toward the mean
« Reply #18 on: November 11, 2014, 10:45:44 AM »
palestis,

The theory is not wrong. Skipping spins or "virtual" betting has no effect on the distribution of outcomes, and this can easily be shown.

The problem is that most gamblers have no understanding of scientific methodology or statistics. They have what seems to them to be a good idea, it may appear to be just common sense, but common sense is often wrong where probability is concerned, that's why it took thousands of years and some very smart people to develop the theory you are dismissing. Basic methodology says that when you're doing an experiment, you need a "control group", so instead of just assuming that when see 3 or 4 reds in a row, it will be to your advantage if you start betting black, you should run a parallel test without waiting for the 3 or 4 reds before betting, in other words just bet randomly.

If your theory is correct, the variance (dispersion or number of losses) will be fewer in the first "experiment" than in the second, but if there is no difference (the variance is the same) it means your theory is incorrect, and you will find this to be the case.

I'm not saying you're not winning, but the reason why is not because of any virtual bets. You can choose not to believe me, of course, but numbers don't lie.

Edit:

Sometimes you may not be sure whether the data shows a difference in variance or not. There are some statistical tests you can use for this, the most common being the F-test, which is available in Excel.
The test is sensitive to Normality (that is, it is assumed that the population the data comes from is Normal, but this will be the case for roulette outcomes, unless you have a biased wheel. ;-))

http://www.statisticshowto.com/how-to-conduct-a-statistical-f-test-to-compare-two-variances/
« Last Edit: November 13, 2014, 03:28:48 PM by kav »

#### palestis

• Great Contributor
• Posts: 850
• Thanked: 810 times
##### Re: Regression toward the mean
« Reply #19 on: November 12, 2014, 05:12:44 AM »
palestis,

The theory is not wrong. Skipping spins or "virtual" betting has no effect on the distribution of outcomes, and this can easily be shown.

The problem is that most gamblers have no understanding of scientific methodology or statistics. They have what seems to them to be a good idea, it may appear to be just common sense, but common sense is often wrong where probability is concerned, that's why it took thousands of years and some very smart people to develop the theory you are dismissing. Basic methodology says that when you're doing an experiment, you need a "control group", so instead of just assuming that when see 3 or 4 reds in a row, it will be to your advantage if you start betting black, you should run a parallel test without waiting for the 3 or 4 reds before betting, in other words just bet randomly.

If your theory is correct, the variance (dispersion or number of losses) will be fewer in the first "experiment" than in the second, but if there is no difference (the variance is the same) it means your theory is incorrect, and you will find this to be the case.

I'm not saying you're not winning, but the reason why is not because of any virtual bets. You can choose not to believe me, of course, but numbers don't lie.

Edit:

Sometimes you may not be sure whether the data shows a difference in variance or not. There are some statistical tests you can use for this, the most common being the F-test, which is available in Excel.
The test is sensitive to Normality (that is, it is assumed that the population the data comes from is Normal, but this will be the case for roulette outcomes, unless you have a biased wheel. ;-))

http://www.statisticshowto.com/how-to-conduct-a-statistical-f-test-to-compare-two-variances/
Again we are getting into advanced probability theories. Based on my results over the years, I never needed any of the theories that members insist on posting every time. Not even variance. Extensive testing under the conditions specific to the way I play have proven to be far more important that standardized theories. There so many other parameters, that you need specific probability applications for each specific parameter. Probability is a general guideline. Probability doesn't know a player's style. A player can stop and leave when he wins \$200, but he can also stop when he lost \$50. Other players will risk \$500 to win \$100. A player can vary his bets from extreme low to extreme high. Probability is blind when it comes to \$ amounts. There is also random noise non other than good  luck or bad luck. These parameters cannot be accurately incorporated in the general probability theory. Too vague.  Only real testing over long period and many thousands of spins can be conclusive. I'm sorry but if those theories were exact for every scenario, I would know it by now. I didn't start roulette yesterday. There was plenty of time for things to fall into place. All I keep confirming each and every time are the results of extensive testing specific to my playing style. Rare exceptions do happen, but in no way they can be catastrophic. Because I don't let them.
Getting into extensive arguments and counterarguments about  probability and how it applies in roulette (with no regard to specific playing styles), it's counterproductive. Only real examples, can put this issue to rest. Maybe a roulette video conference and plenty of time for every one to prove his point.

#### Real

• Fighting the war on absurdity one foolish idea at a time.
• Hero Member
• Posts: 1693
• Thanked: 285 times
• Gender:
##### Re: Regression toward the mean
« Reply #20 on: November 12, 2014, 06:47:13 AM »
Palestis,

You shouldn't pretend to be so experienced.  Based on your way of thinking/system you clearly have not been at the game for long.  There's no disputing the gambler's fallacy and basic probability.  Debating it is really only a demonstration of your inexperience.

Take the time to read and learn from others that have already tried everything you are attempting.

Don't waste your time chasing fallacy.

#### dobbelsteen

• Hero Member
• Posts: 1607
• Thanked: 579 times
##### Re: Regression toward the mean
« Reply #21 on: November 12, 2014, 11:33:08 AM »
Why I cannot past my worddocumet?
See the attachments

#### Mike

• Great Contributor
• Posts: 912
• Thanked: 188 times
##### Re: Regression toward the mean
« Reply #22 on: November 12, 2014, 04:02:41 PM »
palestis,

All of your complaints about probability theory are misconceived. There is a vast literature on probability as applied to gambling, going back centuries. Indeed, gambling was the origin of the science.

Quote
Probability is blind when it comes to \$ amounts.

Not at all. Have you ever heard of "Risk of Ruin"? It's something financial traders are well aware of, and can be applied to any form of gambling. http://en.wikipedia.org/wiki/Risk_of_ruin.

I do agree with you though, that there is more to successful gambling than probability and statistics.

@ dobblesteen,

What is the purpose of the excel sheet?
« Last Edit: November 12, 2014, 09:29:44 PM by kav »

#### palestis

• Great Contributor
• Posts: 850
• Thanked: 810 times
##### Re: Regression toward the mean
« Reply #23 on: November 13, 2014, 08:41:18 AM »
palestis,

All of your complaints about probability theory are misconceived. There is a vast literature on probability as applied to gambling, going back centuries. Indeed, gambling was the origin of the science.

Quote
Probability is blind when it comes to \$ amounts.

Not at all. Have you ever heard of "Risk of Ruin"? It's something financial traders are well aware of, and can be applied to any form of gambling. http://en.wikipedia.org/wiki/Risk_of_ruin.

I do agree with you though, that there is more to successful gambling than probability and statistics.

@ dobblesteen,

What is the purpose of the excel sheet?
Oh really? If I was to bet on BLACK \$1 the probability to win is 18/37, but if I was to bet \$1000 what would the probability be? Something else?
That's what I mean by blind to \$ amounts. The probability stays the same, no matter what the amount of the bet is.

#### dobbelsteen

• Hero Member
• Posts: 1607
• Thanked: 579 times
##### Re: Regression toward the mean
« Reply #24 on: November 13, 2014, 01:24:54 PM »
Mike
A graph says more than thousands words.
Nowadays we have PCs to test our ideas, systems and strategies. It is nearly impossible to analyze or compute manual long random sequences.
I did my first research on the commodore 64.
Not everybody has the knowledge to program roulette events. The writing programs takes a lot of time. The results of the programs has learned me much about the features of random rows. On these results I have based my theory of nano and macro sequences.
With excel it is very easy to compute a lot of samples with a single touch of the F9 key. Many members all over the world use my programs. Sometimes I give explanation on internet.

#### kav

• www.Roulette30.com
• Hero Member
• Posts: 2244
• Thanked: 1179 times
• Gender:
##### law of large numbers
« Reply #25 on: December 01, 2014, 11:32:22 PM »

#### dobbelsteen

• Hero Member
• Posts: 1607
• Thanked: 579 times
##### Re: Regression toward the mean
« Reply #26 on: December 05, 2014, 11:24:18 AM »
I have visited the YouTube college about the Law Of the Large Numbers. For mathematicians very interested. It bears out  me in my theory of nano and macro events. See my contribution about the long run theory.

#### weird

• New
• Posts: 109
• Thanked: 12 times
##### Re: Regression toward the mean
« Reply #27 on: June 04, 2015, 03:54:31 AM »
Quote
So waiting for triggers, and all the rest of it, makes no difference to mathematical probabilities pertaining to various patterns etc.

Of course it make a huge difference!

Say, If u see that RED had hit streak for 20red in row!
Will RED hit for 20in row in next 20 spins???
Of course not!

If u see, after 50spins, red only hit 10 times,
will red hit only 10 times in the next 50 spins?
Of course not!

Many try not to think hard, and avoid understand this fact...

THINK! why 100red-in-row, never hit in 100spins?

#### dobbelsteen

• Hero Member
• Posts: 1607
• Thanked: 579 times
##### Re: Regression toward the mean
« Reply #28 on: June 04, 2015, 10:44:43 AM »
The opponents of my short run theory give always unrealistic arguments to attack the theory.