TurboGenius The Fallacy of Gambler's Fallacy - 2020

Nope. You're making a rookie mistake. It's just more gambler's fallacy. That is not how regression to the mean works. Before you attempt to play the game for real money you should take the time to study the basics. A good place to start is the wizard of odds. All newbies should study that site. The expectation moving forward is always just expectation and not that an imbalance should even out. After all, why should it since there are the same number of pockets on the wheel from one spin to the next?



Best of luck,

Dr. Sir Anyone Anyone.

 

Absolutely wrong. The Law of Large Numbers and every expert will disagree with the above.

 

JBS and Siranyone,
Do not read my statement in bits and pieces. I have written a separate topic on how RTM is supposed to help us practically, avoiding the worst possible.
If there is an event that is at 5 SD, the same can not repeat again and again. Have you seen 25 or more losses followed by a single win coming twice, thrice or more in succession in any EC? Do not argue only for the sake of an argument. Talk practically and logically.
@Siranyone,
Wizard of odd is meant for those who are mathematically illiterate and can not comprehend the basic principles of statistics and probability. Even for such fellows, buying a basic book on probability and statistics will help more than surfing around Wizard of Odds.
P.S: Those who are believers of absolute randomness and those who will even 100 black numbers will show some day without any hint of any red number, may God help you. There is a virtual limit of randomness. It is not tending towards infinite. RTM helps sane people only and ignorant may abuse it as a fallacy.

 

In the short term there are more sets where RTM can work in your favour simply because the number of sets that have for example 100 reds in a row are rare. But you can not be 100% certain you won’t get more reds than that. So in theory you could be winning with your system for 10 years! And then one day your system stops working and that very rare event with 101 reds or 200 reds occurs.

That is why it is a fallacy not because it can’t help you win in a hit and run strategy.

This is all common sense. I can’t believe after 20 years of these forums people are still debating these basic ideas.

 

In theory you should not gamble or even talk of gambling. Even 101 or 500 reds in a row will not harm us even a little, using my RTM principles as I illustrated in a separate topic already. By the way, this is not just rare but next to impossible to happen, if you see the sequential probability of such events. getting 30 reds in a row could be said to be rare.
I have gone through over 50 million real spin data from various sources, there is no incident of even 25 or more losses followed by a win being repeated immediately, even once.

 

You're RTM principles only work when you already know how many times your numbers will hit in a confined amount of data. In other words, when you can cheat and travel forwards and then backwards in time to bet. In the real world of gambling, your newbie ideas won't work because you won't have the luxury of looking forwards and backwards in the data.



The probability of a number hitting or not hitting is not influenced by the past spins. After all, what would be the physical cause if past spins could influence future spins? Magic maybe?

If you naively believe that past spins influence future spins, then simply look down and count the pockets to see if any of them have disappeared since the previous spin.

 

I m talking of regression towards mean which is a statistical reality and these self proclaimed experts are talking of past spins and predictions. Who the hell is saying that it is possible to predict future with past spins? Stop this bloody rant.

 

What is the difference between RTM and gambler's fallacy?

 

Fallacy lies in believing that after a bad patch, "good" and "compensatory" wins are due to happen and vice versa. RTM says that after an extremely bad patch, further patches are more likely to "eventually" get closer to the mean. RTM doesn't expect compensatory wins.

 

anyone who believes that numbers or sections of a roulette wheel are due to come would have to believe the wheel has a mind, and that it is very conscious and concerned about the imbalance of its past distributions

and that the mind of a roulette wheel does not like the imbalance that has occurred in the past
and the wheel will consciously make efforts to rectify that past imbalance by controlling what it distributes in the future

this is the only way a roulette wheel can be happy - if its distributions even out - and they must even out very quickly - not in the course of millions of spins - the wheel will force the evening out with just a few hundred or even less - maybe just a few dozen spins

once the past imbalance is corrected the roulette wheel becomes content

and once again begins random distributions

until another imbalance occurs - and the wheel again decides to control its upcoming distribution to regain its balance

the wheel will not tolerate an imbalance for very long - this makes it upset

it must act to correct a past imbalance

 

This is a nonsense post, but instead of removing it - here is the common sense argument





You can clearly see that "random" stays within limits, a lower limit and a higher limit.
It isn't because anyone believes the wheel is conscious or has some intention of staying within limits...
It's called math. Random performs within predictable boundaries.

So to rightfully assume that something at the far edges of these limits are "due" to either slow down
or speed up in appearance is absolutely correct and easy to prove/demonstrate.
Being ignorant to this simply displays that you don't understand math or random, or both.

Sure enough, numbers well above average "slow down" and numbers well below average "speed up". To argue this would be nonsense.

 

That sounds like the same thing. Do you believe you use that definition to win on the next spins?

 

I will remove each and every post that I wish to - when that post does not have to do with the topic of the
thread or is nonsense/trolling. Understood ? (As much of this thread is exactly that).
If there is a dispute concerning what I posted, lets discuss it.
If there is none and people want to post nonsense - expect it to be removed.
Feel free to use the PUBLIC forum for your nonsense, or come up with reasonable debate
and discussion on the topic.
You can always start a topic in the public section and not troll in my section ? See how it works.
I'm not the moderator that you're used to - I don't tolerate BS so expect to see routine cleaning.

 

RTM is not meant to deliver you great moments where you can expect to win more and lose less autopilot. It could just help one to get rid of the worst possible.

 

https://stats.stackexchange.com/questions/204397/regression-to-the-mean-vs-gamblers-fallacy

Best answer :

@ albalaha,

So your statement that "it is safer to presume that now Tail should behave relatively better than earlier since it was too harsh earlier."
is an example of the gambler's fallacy, because you're referring to previous history. Regardless of what happened earlier, the expectation for tails is whatever the expectation is for the number of tosses. e.g. if 20 tosses the expectation is 10 tails. But the expectation for 10 tails doesn't change depending on what happened in the previous 20 tosses. So waiting for a sequence of 20 tosses in which tails only hit once or not at all doesn't mean it is now more likely that you'll get 10 tails in the next 20. The whole "virtual losses" concept is part of the gambler's fallacy.

Any statement of the form "Given X outcomes have occurred, then Y outcomes will occur" only applies to games where there is dependency between outcomes. RTM is a true statistical phenomenon, but it has nothing to do with independence/dependence of outcomes; the "best guess" is always the expectation regardless of what happened previously. For some reason this seems to be a tricky concept to grasp for system players.

If you want to test whether virtual losses helps, you should not just look for confirming instances (ie counting the number of wins after a virtual loss), but also count the number of wins after virtual wins, or an average sequence. You'll see that there is no difference on average and that waiting for virtual losses is a waste of time.

 

Turbo,

But assuming your results are significant, doesn't this contradict your theory that you're more likely to win betting on hot numbers/repeaters?

Both tests seem to indicate this. Betting on cold numbers resulted in not one failure, but betting on hot numbers did.

 

I tested this below are the results. To be fair I only saved the ones when there were 5 numbers left just as in the example turbo gave.







total result was -2022 units.

I did the same test with 5 random numbers. The result was -1746 units.

I did another test with wat ever the last numbers were without a 0 (so the average was indeed 5) that result was -2322

So this post is not about trolling but @TurboGenius could you please tell us why your method in the example you gave in the first post is better than random?

p.s. The numbers in red are the one without a win in the cycle.

 

I'm not sure what the problem is - or how your data is somehow different other than you experienced 0's appearing.

I started a new chart this morning but frankly the results are the same as before and I don't see the need to waste time
doing thousands of spins again by hand to end with the same result.

In this test, I ran off spins until 5 numbers remained - then I bet on those 5 numbers only flat bet and
recorded how many wins were achieved per future 37 spin cycle.
I still haven't seen a 0 - but it could happen of course, the odds of it happening are small.





The chart was just flat betting $5.00 per number - granted it won't always produce a win
because some cycles of 37 spins have 1 win, some have 8, etc.

My point of the post/thread was to show that once a rare event has happened - those specific 5 numbers
not showing together as a group for x amount of spins - they won't repeat that, therefore they produce
predictable wins based on past spins. This makes the fallacy not a fallacy at all.
If I can predict that at least 1 number of 5 will show in the next 37 spins based on past spin data - there's
no reason why any player can't do the same.

 

Thanks for doing the test also! I really appreciate that. So if our results are not the same there must be an error somewhere. Just to make things clear. Once you have your 5 numbers you play those number for 22 cycles right? Or do you pick 5 different numbers each new cycle?

I did a new test with 5 different numbers each cycle for 200 cycles (just the 5 numbers) and indeed there were no zeros. However I cant win flatbet. See the screenshot.



So the test you did with RX was just luck. I had those graphs as well, but when doing more and more spins it always go down. As you can see in the screenshot the loss per number per cycle of 37 spins was 1.288 per number. Playing more cycles it would go down to roughly 1 unit per cycle just as the math says it should.

So the question remains (and I have asked this a few times) It can not win flatbet can it? We have to use a progression? Dont get me wrong I'm fine with a progression as long as you know a certain event happens a certain amount of spins. But winning this flatbet is just impossible, I wish I was wrong :-(