law of probability not true

[simon]

...or if true, it is meaningless.  there is a law of probability which states that there will be half as many runs of 2 as runs of 1, and half as many runs of 3 as runs of 2, and half as many runs of 4 as runs of 3, etc.  if this were true you could beat roulette or baccarat with a very simple system... all you would have to do is watch any even chance, and when it changes for example from black to red, next bet black, and if red hits again you wait till the run of red ends (black shows again), then bet red, if red wins next bet black, etc.  Anotherwords always bet on the chop (a single) and if lose to a double or more, wait for that run to end and then bet for the chop again.

if the distribution of series is true, then you would win in the long run by just betting on chops, and losing ONCE to every series of 2 or more (certainly at baccarat or a no zeroe wheel, or make small insurance bets on the zeroes.)  I have tried to find out how many spins are required to get this ideal distribution, but have been unable to get this answer.  if this expected distribution of runs is true, then I would like to know why it does not apply to roulette (my test results of hundreds of spins show that you can not count on more chops then runs, in fact often the opposite happens.)

[winkel]

There is a simple flat-bet strategy you can win with.

It is called Coup de Lion (it is also known with other names)

Try it flat bet with just one bet. Stop when your at +1 units.

br
winkel

[hoper35]

hmmm, something that doesn't happen as often as it should. 

[rjeaton1]

Hopefully I can help on this one...as I have actually been playing around with a similar idea in RXtreme...

The likelyhood of a chop sequence occuring (which requires two numbers...one red and one black) is just as likely as two numbers of the same color coming out.  What I mean is this:

There is a 50/50 shot of red coming out on any given spin (I'm doing the math as if there is no zero...just to keep it simple).

So, the likelyhood of two spins being this - Red, Red are simply the probability of it happening once times the probability of it happening once...like this:

0.5 X 0.5 = .25 X 100 = 25% chance you'll see a run of Red, Red.  Now, the probability of you seeing a 2 step chop sequence (because a chop only qualifies if it is at least two steps long) would be calculated like this:

Red then Black

0.5 X 0.5 (as they both have a 50/50 shot of coming out) = .25 X 100 = 25% chance of you seeing a two run chop sequence.

When you look at a sequence as more than one spin, you have to take into account the mathematical probability of each event.

The point is, a two run chop sequence has the same probability as a run of 2 reds.

Here is an example (a real world one...not theory)

I wrote a script for RXtreme that keeps track of two sets of "patterns".  The first being: r,b,r,b  the next being r,r,b,b

How you would calculate BOTH of those runs would be like this:

0.5 X 0.5 X 0.5 X 0.5 = 0.0625 X 100 = 6.25% chance I'll see a run of r,b,r,b OR r,r,b,b


Now, after 5,843 spins, my results were as follows:

r,b,r,b happened 349 times or (in percentages) 349/5,843 = 0.0597 X 100 = 5.97% of the time r,b,r,b came out


Using those same 5,843 spins I did the same for r,r,b,b,  Here is what I got

r,r,b,b happened 345 times or (in percentages) 345/5,843 = 0.590 X 100 = 5.90% of the time r,r,b,b came out


Hopefully this all makes sense and clears things up.




The reason you'll see streaks of one twice as many times as you'll see streaks of two is because of this (and this is what makes the "law of probability" annoying)

For you to analyze data saying "how many times did a streak of one happen and how many times did a streak of two happen" you have to look at the data TWO TIMES looking for streaks of ONE (in which case you can ONLY TAKE INTO ACCOUNT SPINS AS INDIVIDUAL OCCURENCES) and streaks of two (IN WHICH CASE YOU CAN ONLY LOOK AT SPINS IN CHUNKS OF TWO AS INDIVIDUAL OCCURENCES)

So, let's say we have four spins to analyze...we've got red, red, black, black

First, let's analyze for streaks of one...

1st spin - Red - streak of one
2nd spin - red - streak of one
3rd spin - black - streak of one
4th spin - black - streak of one

Now, let's analyze of for streaks of two...


1st chunk - red, red - streak of two
2nd chunk - red, black - no streak
3rd chunk - black, black - streak of two


So, same data analyzed and here are the results - 4 streaks of one, two streaks of two...holding up exactly to the "theory of probabilitys"

[rjeaton1]

Oh, and for anybody interested in the rx script, it actually keeps track of a lot more...here is a screenshot of what I've programmed it to track. 

If you want it, let me know.  It doesn't place bets, it only tracks stats.  So, if you want to see if your spins hold up to the "math" of the table, (and trust me they will...it sucks I know...) you can import your spins and run them and then look at the data.

[rjeaton1]

Well, it actually does place bets...er, it used to...until I deleted the "betting" part from it.

Here was my idea...which worked...then stopped working...haha

Things happen in streaks...meaning those percentages will be right AFTER thousands of spins....in the short term it is kind of all over the place.

You might see r,b,r happen just as many times (100%) as r,b,r,b which you shouldn't...when looking at it mathematically.

Or you might see r,b,r happen 100 times and r,b,r,b happen only 10% of the time (which is FAR behind what it should be)

So, I set it up to bet when BOTH of these stipulations were accurate:

r,b,r,b happened less than 5.00% of the time in the amount of spins tracked and the difference between how often r,b,r happened and r,b,r,b happened as less than 30% (it should be at or around 50%).  If both of those were accurate it would bet on black every time r,b,r happened.

I got all the way to 1,000 bets placed with a win to loss ratio of +16%  (16% above the expected 50%)...then...if all went down hill...

[rjeaton1]

I'm presently using that script to TRY to determine when a streak is beginning or when a streak is coming to an end...and then bet accordingly...I'll let you know how it goes.

[rjeaton1]

Below, I'll show you a "cut" portion of the BR balance trend mentioned above...and then what happened afterwords (Oh, and the only reason I'm posting this here is it is directly related...in fact..the same...to using or talking about the theory of probablility.

Anyway, the first picture you'll see means I just kept on getting lucky with at what point in a "streak" I was betting in.  Meaning, everytime it bet it was after it was much lower than it should have been, and was on it's way to an upswing (correction) of that abnormally low percentage.

The second picture is when it all caught up to me and I caught WAY more downswings than upswings.  Meaning, it was abnormally low when I started betting and then either stayed low or went even lower.

(it is flat betting by the way...the first picture had a win to loss ration of 16%...then it all went down hill...at one point at its highest, it had a win to loss ration even after 3 or 4 hundred bets that was 24%...that is 24% over the expected 50%...so, a 74% strike rate flat betting on only red and black)

[Spike]

You're talking about Law of Series, and its not a law, its an observation. You can't use it to beat random outcomes because it happens randomly, you can't predict it, only observe it after it happens.

[rjeaton1]

You're talking about Law of Series, and its not a law, its an observation. You can't use it to beat random outcomes because it happens randomly, you can't predict it, only observe it after it happens.

I don't know what the name of it is, but if it is based in math and always happens and always will happen, I'm pretty sure those are all of the qualifications of a law.  And technically speaking, anything that can be proven with math can be predicted to happen...just not necessarily when...

For instance, you can predict that on a single zero wheel (using mathematical laws) that I will lose 2.70% of my bets.

[simon]

thank you rjeaton for your excellent explanation-- I have to re-read it and think about what you wrote.  sounds right, thanks!  (seems like there might be a better way of betting on chops, maybe not...)

[rjeaton1]

thank you rjeaton for your excellent explanation-- I have to re-read it and think about what you wrote.  sounds right, thanks!  (seems like there might be a better way of betting on chops, maybe not...)

No problem at all Simon, glad I could help!   

[winkel]

For instance, you can predict that on a single zero wheel (using mathematical laws) that I will lose 2.70% of my bets.

You can´t because it is only 1,35%

[rjeaton1]

You can´t because it is only 1,35%

It's 1.35% if the La Partage rule or En Prison rule are in use at the (single zero) table you're playing on...if not the house edge (single zero wheel) is 2.70%

[Spike]

pretty sure those are all of the qualifications of a law. >>>

Its not a law. Just like Law of the Third is not a law, Law of Series is not a math law. Look it up.

[Spike]

-