do present spins have any influence on the probability of past spins?

[hanshuckebein]

hi folks,

it's been discussed over and over again if past spins influence the outcome of future spins.

in the book "entropy demystified" the author kind of turns the question around an asks:

"do present events affect  the probability of past events?" 

I have attached the relevant pages to this post and hope this doesn't break any copywrite laws.

unfortunately  for a detailed solution to this question the author refers to another book which I haven't read, yet. but he states that present events DO influence past events probability wise.

I'd really like to hear your opinions on this issue.

cheers

hans

[hanshuckebein]

of course I mean copyRIGHT not copyWRITE *lol*.

unfortunately I can't find the "modify button". 

[Colbster]

On a 1-on-1 consideration, the previous spins have absolutely no correlation to the future spins.  However, the next spin is 1 second away from being past history, and a part of the entirety of statistics and it will conform with the statistics.  Over the enormously huge long term, we can expect that there are only a certain number of 32s to be spun, as they will eventually only make up 1/37 of all the spins.

I could be swayed to acknowledge that spins are related in the super-macro sense, but not in any way that we can use in betting contexts.

[Gizmotron]

Consider this: let's say that in the past 100 spins 60 reds, 40 blacks, and 0 greens occurred. Did that segment of 100 spins have its own statistics? Let's  also stipulate that at spin 50 of this same segment there were 30 reds and 20 blacks. Do the temporary odds of the first half of this segment confirm a state of continuation in the second half?

Is there a type or study in math that regards temporary states or a kind of temporary  condition of statistics? If not then why not?

[MauiSunset]

The odds, in Roulette, are fixed - they don't change.


Maybe in an alternate universe Roulette is different but past spins mean nothing to future spins and future spins mean nothing to past spins in our universe.

[Gizmotron]

MS, with that logic if 100 spins happens then all the spins that ever happend  on earth either never happend or the 100 spins that just happend never happend. The good news is that you are in a great situation where you will never have a clue how to use randomness or trends.

[reddwarf]

Hi Hans,

the answer is in the book: conditional probabilities should, in general, not be confused with cause and effect. In the example the sample space decreases with every draw, so the probability that I drew a white ball indeed changes under the condition that I draw a white ball now!! This has nothing to do with present spins influencing previous spins.

There are therefor 2 reasons why this text does not apply to roulette at all:
1. the sample space of roulette does not change, there are always 37 or 38 numbers to choose from
2. RNG's or roulette tables are Software or Mechanical implementations of a random generator, to their best possibilities these are random in the sense that:

a. numbers are unpredictable (BM: true, in the past Shannon et all succeeded using a roulette computer, 1961 if I'm not mistaken, RNG: true in the past there were really lousy SW algorithms)

b. numbers are independent (again: if the croupier would be a robot or if we have a really lousy RNG, than there might be dependence)

c. numbers are unbiased (=uniform distribution)

This implies that, if all is OK, conditional probabilities are meaningless (P(1/2)=P(1)*P(2))

reddwarf

[reddwarf]

Hi Gizmotron,

It all depends: i could build an RNG that shows biases on the short term, but is unbiased on the long run. However, this would imply that, as a feature of short term, that it would be possible to identify short term blocks of numbers! This then violates randomness, so actually this is a lousy RNG, or stated in technical terms, the dimensionality is very low.

The question is solved for RNG's (if we assume they are not hampered with): the current algorithms are so sophisticated that every person on the planet has to test the RNG till the universe is supposed to end before they will have a hint of a significant pattern

For roulette wheels the questions remains open: if the croupier is a robot, if the starting speed is constant and if the intervals between spins are exactly identical, then yes, you probably have a lousy embodiment of a random generator.

But apart from this ramble of mine, if your question is: couldn't there be an underlying process that forms pools of correlated spins in the sea of random numbers. If that is your underlying question: my answer is - I do not know, actually nobody knows. hell the Chinese thought there was (I Tjing), Princeton did some research (You might want to read this)

reddwarf

[Bayes]

Consider this: let's say that in the past 100 spins 60 reds, 40 blacks, and 0 greens occurred. Did that segment of 100 spins have its own statistics? Let's  also stipulate that at spin 50 of this same segment there were 30 reds and 20 blacks. Do the temporary odds of the first half of this segment confirm a state of continuation in the second half?

Is there a type or study in math that regards temporary states or a kind of temporary  condition of statistics? If not then why not?

Gizmo, regression to the mean suggests that if the percentages are as you say at spin 50, then it's more likely than not that in the second half the proportions of R/B will be more nearly equal. Note this is NOT gambler's fallacy - the odds are still fixed and it doesn't mean that black will try to "catch up" somehow, it just means that rare events are more likely to be followed by "average" events. The rarer the event, the stronger the regression effect.

@ reddwarf, that article you linked to seems like new age gobbledegook to me. It's one of my pet peeves that quantum physics is often invoked to apparently justify all kinds of nonsense.

[reddwarf]

Hi Bayes,

i have no opinion about that article, to be honest, i did not even read it. I just added the link because some might find it interesting, and it illustrates that more people ask themselves if RNG is really RNG.

I am a strong believer in Occam's razor.

greetings reddwarf

[Bayes]

On the other hand, if you don't know the probability distribution, and assume nothing (the wheel may be biased for all you know), then you should go with the empirical evidence. Thus, if you see 10 reds in a row, the smart bet is on red. If you see 30 reds in the first 50 spins, you should bet red for the next 50 (or until the empirical evidence suggests otherwise).

[Bayes]

I am a strong believer in Occam's razor.

greetings reddwarf

Me too, but note "Controversial  aspects of the razor".

[reddwarf]

Hi bayes,

Yeah: Occam's razor is a tool to be used within a paradigm, everything else being equal and stuff.

By the way, this reminds me of a story I once read and it goes something like this: there was this maths PhD and a successful business man who only learned to read and write. Both sat in a casino at the same table, after 100 reds in a row a guy asked both the question what he should bet. The maths guy said: it doesn't matter (blah blah you know the drill). The business guy shook his head and said: red of course, this wheel is surely biased.

reddwarf

[Bayes]

It's a good example of how sometimes the "mathboyz" can be completely blinkered in their thinking. 

[Bayes]

Your views on randomness are just insane - sorry to inform you.....


You proved you can't debate in another forum here when you deleted my posts.


Can't take the heat can you......

Maui, I don't have any argument with you, but do you use any kind of bet selection process in the way you play, or do you rely purely on money management?

If you do have a method of selecting your bets, then why do you believe that this method is superior to betting the same location spin after spin, or just randomly?

On the other hand, if you DO bet randomly, and rely on pure MM, surely you know that it's mathematically impossible to win consistently just by manipulating your stakes?

Either way, I think you're being inconsistent and/or hypocritical, given that you've said (on the VLS forum) that you ARE winning consistently.