0 Members and 1 Guest are viewing this topic.

*

Turner

  • 3000+ Posts Member !!!
  • *****
  • 3855
  • Rated: +201
I was actually starting to consider becoming a Canadian Citizen! 
are things really that bad? lol

*

Bigbroben

  • 1500+ posts member!
  • *****
  • 1551
  • Member
  • Rated: +160
Being a Canadian is one of the best situation there can be, you know.  Next to being a porn star.
Life is hard, and then you die.
Mes pensées sont le dernier retranchement de ma liberté.

*

TurboGenius

  • 500+ posts Member
  • *****
  • 896
  • Math beats a math game - Turbogenius@gmail.com
  • Rated: +105
So if there is potentially any difference between betting on say the 5 hottest and coldest numbers this rating method should show it. What do you think?
Comments from Turbo welcome. ;-)

Interesting approach

*

The General

  • 2000+ posts MEMBER!!
  • *****
  • 2034
  • The gambler's fallacy is alive and well.
  • Rated: +125
For there to be a hot number it has to have hit several times in the past, not the future.  This means that we have to use past spins, whether we have bet on them or not doesn't matter.

Not all hot numbers are created equally.  Some hot numbers are biased numbers.  They hit more frequently than they should because of a defect on the wheel.  These are the kind of hot numbers that people want to play because they hit more frequently than probability would predict for a reason.  The accuracy of a prediction has been increased.

Defining a biased number can be achieved through standard deviation testing of a statistically relevant number of spins and by testing the over all fitness of the wheel using chi square.  Now, many people like to point to the law of the third and binomial distribution, but really chi square is a much better way to quantify how far outside the norm a wheel is performing.  Several people are now looking at how many non hit verses hit numbers there are in various samples to define how a wheel is performing.  My suggestion to them is stop using the law of the third nonsense,  collect much larger data sets, and use the chi square instead. 

Hot numbers that aren't biased numbers are just random fluctuations. Meaning they're bobbing up and down due to variance.  Unfortunately, variance/luck is not predictable.  We can define the range of variance that's possible, but that's only because we can define the size of a spin sample.  Meaning we can't have 11 reds in a row if we're only looking at 10 spins.

-The General

Basic probability and The General are your friend.
(Now hiring minions, apply within.)

*

TurboGenius

  • 500+ posts Member
  • *****
  • 896
  • Math beats a math game - Turbogenius@gmail.com
  • Rated: +105
Those are all valid points - that have nothing to do with how I'm playing hot numbers.

*

The General

  • 2000+ posts MEMBER!!
  • *****
  • 2034
  • The gambler's fallacy is alive and well.
  • Rated: +125
Those are all valid points - that have nothing to do with how I'm playing hot numbers.

Oh, are we still talking about you?  Ermm I thought that's what the other threads we're for?

I felt that it was time to post some facts rather than the absurd.  ;D
Basic probability and The General are your friend.
(Now hiring minions, apply within.)

*

TurboGenius

  • 500+ posts Member
  • *****
  • 896
  • Math beats a math game - Turbogenius@gmail.com
  • Rated: +105
My apologies - carry on. I thought all of your posts were about me and contradicting anything I post. My mistake. 11 reds in a row can't happen in 10 spins, got it.


Share via delicious Share via digg Share via facebook Share via linkedin Share via reddit Share via stumble Share via twitter

 

Popular pages:
  • XHTML
  • RSS
  • WAP2