Our members are dedicated to PASSION and PURPOSE without drama!

Reality of "Law of large numbers"

Started by Albalaha, June 19, 2013, 02:55:35 AM

Previous topic - Next topic

0 Members and 4 Guests are viewing this topic.

Albalaha

This is a big eye opener for most of the newbies to safeguard them from growing number of fallacies.
http://albalaha.lefora.com/2013/06/19/reality-of-law-of-large-numbers/
Email: earnsumit@gmail.com - Visit my blog: http://albalaha.lefora.com
Can mentor a real, regular and serious player

monaco

In your example Albalaha, wouldn't you rather be the player playing the 2nd 100 spins though, where there are 53 reds & 44 blacks? With a bit of MM you could make a profit from those numbers. 

Playing after a strong deviation is not (only) to play for the catch-up of the least-showing side, but to play in a period of lower variance.

Drazen

Exactly monaco

From my experience it is significantly better to switch triggers more often according to the stats they all have in common, then longer chasing one trigger even if the starting STD on it is huge.

Regards

Drazen
Common sense has become so rare it should be classified as a superpower.

Albalaha

QuotePlaying after a strong deviation is not (only) to play for the catch-up of the least-showing side, but to play in a period of lower variance.


Another fallacy. It can be opposite too. You can not pick a right or wrong time to get a hit on a bet irrespective of its past performance. Only in very long run, you can expect a "correction in the proportion of variance and not in the count".
            For instance, in a 100 spins you may get 45 Blacks and 55 Reds and in next you may see a bigger gap of 35 Blacks and 65 Reds. Gradually and in large number of data the gap shrinks and they tend to go towards their respective mathematical proportions. The initial gap of 10 may never shrink ever and it can go rather larger, although percentage wise, they will come closer to their mean.
                         Wrong interpretation of a correct theory is also a fallacy. I would like Bayes to come forward and clarify this.
Email: earnsumit@gmail.com - Visit my blog: http://albalaha.lefora.com
Can mentor a real, regular and serious player

Albalaha

Quote from: PatternAnalys on June 19, 2013, 06:20:27 AM
dear Allabaha, may u please delete all my stupid kiddie post :P on your respected blog, [i had ask u long ago, but still there]...lest i become laughing stock by others. Thanks in advance! :cheer:


       You requested membership at my blog. Then you started talking your wise martingale (which you are admitting as stupid kiddie posts now). Later after I admonished you, you started verbal abuses which I never tolerate even if you are the wisest man in the world. I can delete them only if you apologize for your harsh and undue words.
Email: earnsumit@gmail.com - Visit my blog: http://albalaha.lefora.com
Can mentor a real, regular and serious player

Bayes


The law of large numbers and regression to the mean aren't the same thing. This extract from the wiki entry on Regression to the Mean might help:

QuoteRegression toward the mean simply says that, following an extreme random event, the next random event is likely to be less extreme. In no sense does the future event "compensate for" or "even out" the previous event, though this is assumed in the gambler's fallacy (and variant law of averages). Similarly, the law of large numbers states that in the long term, the average will tend towards the expected value, but makes no statement about individual trials. For example, following a run of 10 heads on a flip of a fair coin (a rare, extreme event), regression to the mean states that the next run of heads will likely be less than 10, while the law of large numbers states that in the long term, this event will likely average out, and the average fraction of heads will tend to 1/2. By contrast, the gambler's fallacy incorrectly assumes that the coin is now "due" for a run of tails, to balance out.

Albalaha

QuoteIn no sense does the future event "compensate for" or "even out" the previous event,


Thanks Bayes. I highly appreciate your inputs. Does it mean that if previously a bet had bad strike rate, it is not necessary that some future event will "even out" the gap?
                In my personal observations and studies, if a bet is going bad now, it does not necessarily even out, it may rather go worse too.  Number 3 of zumma is a leading example of this.
Email: earnsumit@gmail.com - Visit my blog: http://albalaha.lefora.com
Can mentor a real, regular and serious player

Drazen

Of course it is not necessary in some shorter future. Regression to the mean is not causal phenomenon...

Drazen
Common sense has become so rare it should be classified as a superpower.

monaco

Quote


following an extreme random event, the next random event is likely to be less extreme



If a bet had an extremely bad strike rate (say, -3STD), it is less likely to continue getting worse at that extremely bad rate.


'Evening out the gap' between 2 sides is different & not really the aim.
Like Drazen implies, 'evening-out' & 'due' are the fallacies.

Albalaha

Movement of number 3 of zumma is a burning example against this "even out" theorem. It doesn't matter whether you start playing after 3.0 STD or 5.0 STD of any bet. You are as much likely to face a bad phase as someone who is playing since beginning. Playing a bet after a very bad time doesn't fetch you further wins, flat bet. I can see this fallacy being mongered in many of the so-called "systems" with pride.
                              Flat bet can never win, in a systematic manner in a game of house edge. It can only win if u get streaks of wins, by chance. The best way of trying your luck in a game of chance is use of mild negative progression and patience.
Email: earnsumit@gmail.com - Visit my blog: http://albalaha.lefora.com
Can mentor a real, regular and serious player

Bayes

Quote from: Albalaha on June 20, 2013, 01:02:49 PM
It doesn't matter whether you start playing after 3.0 STD or 5.0 STD of any bet. You are as much likely to face a bad phase as someone who is playing since beginning. Playing a bet after a very bad time doesn't fetch you further wins, flat bet.


It may not get you "further" wins, in the sense of more wins than average, but it's not true to say that you are as likely to hit a bad run after 3.0 or 5.0 SD as anyone who doesn't use those events as a trigger. A 3 SD event is less likely to be followed by another 3 SD even - that's what Regression to the mean says. Think of it like this: Suppose you divide spins into sequences of 10, and each sequence can either include at least one red or not. An extreme event in this case would be no reds in a sequence, which has probability of 1 in 1024. Now suppose you bet red (a) randomly, or (b) only after seeing 10 blacks in a row. While it's true that in the case of (a) you will rarely hit a sequence without at least one red, it will happen more often than if you only bet (b). Neither the gambler's fallacy nor the law of large numbers applies in this case, but simply the fact that the most likely scenario for any sequence is the average, that is, between 4-6 reds in the sequence.




Albalaha

Will regression towards mean help anyways in winning with flat bet? Does it offer more winning opportunities to play after 3.0 STD than not? This is my last query.
Email: earnsumit@gmail.com - Visit my blog: http://albalaha.lefora.com
Can mentor a real, regular and serious player

GreatGrampa

Quote from: Albalaha on June 21, 2013, 03:19:21 AM
Will regression towards mean help anyways in winning with flat bet? Does it offer more winning opportunities to play after 3.0 STD than not? This is my last query.
Answer is no for both
Greatgrampa - Your friend and mine

Bayes

Quote from: GreatGrampa on June 21, 2013, 06:40:22 AM
Answer is no for both


Agreed. Don't read more into RTM than what it says:
QuoteRegression toward the mean simply says that, following an extreme random event, the next random event is likely to be less extreme.

Albalaha


Debate closed with a conclusion that "law of large numbers" and "regression towards mean" both talks of things going smoother in long run than short run in terms of "proportion" and not in terms of "count". These principals are applicable to all short of gambling which has fixed odds and specially useful for the game called "roulette".
                                  Never forget that nobody has so far defined "long run" so playing after 3.0 STD or 9.0 STD has no difference. When we play only after so much dispersion, we always miss all early hits that come in good times and at small or regular intervals. So play after certain number of losses is not a winner concept in itself. After a push of negative progression, although, it may do well.
Email: earnsumit@gmail.com - Visit my blog: http://albalaha.lefora.com
Can mentor a real, regular and serious player