This is a topic that I have been working on for a while now.
I had a thought, I thought, "Is there anything ordered about random numbers, any patterns that I could take advantage of.
Im not talking about repeaters or sleepers, I am talking about maths.
Is there anything about the randomness of the numbers that come out of a roulette session that I could definately say was ordered?
I thought "If there is then I can use it to create a method, and I beleive I have.
Firstly, What is the pattern.
The only pattern I could see that had its roots in maths was the binomial distribution.
Even though the individual numbers were coming out randomly in the short term, if we build a picture over the long term, it becomes more of a patern.
So in order to get order from radomness, we would have to build up a statistical picture over many many spins.
This of course is something which to the roulette player is useless, because we want to bet on the next spin or series of spins.
So I set about applying this theory to a system.
Here is what I produced.
I tracked for 10 spins, but not the number,the street.
So the series 1, 34, 23, 12 would be 1, 6, 4, 2
with 0 tracked as 0.
for each spin I then calculated how often in the last 10 spins had that street appeared.
For example, If a 3 came out, I would write down if it appeared 1 spin ago, 2 spins ago, 3 spins ago . . . .. and so on up to 10 spins.
Once I had done this for many numbers, using excel and randon.org, I worked out the amount of times in 1000 spins as a percentage that the street had been out once , twice, three times and so on, in the last 10 spins.
What you get is very reproducable from 1 set of 1000 numbers to another.
The numbers looked like this.
The street just spun had not appeared in the last 10 spins 18% of the time, once 32% of the time, twice 28% of the time, three times 16% of the time and 4 or mor times 6% of the time.
This is a kind of binomial distribution and is remarkably reproducable.
So the system is this.
Track 10 spins as streets like this
1-6 = 1
8-12 = 2. . . . . .you get the drift
then write down how many times each has appeared.
so you might get this
1=1
2=2
3=2
4=2
5=1
6=2
so the total will add up to 10.
We bet 5 streets when we bet, so from now on we will be selecting which street to leave out, not which to bet.
So, how do we select which to leave out.
We leave out the street that has appeared 3 or more times, and is the one that has appeared the most.
It must be alone, not equal to another.
So if we tracked 1
1=1
2=1
3=3
4=2
5=2
6=2
we will leave out street 3 and bet on 1,2,4,5,and 6
We do this with a 2 step progression as follows:
Its a Grand Martingale,
1,1,1,1,1
7,7,7,7,7
If we lose 2 on the trot, we take the hit and start again, If we win on the first bet we start again at the beginning.
Every spin has 10 preceding spins, but we only bet when one street is dominant and has appeared 3 times or more
QUESTIONS??????
GINGER
interesting concept.
1 question though.. by streets, you mean double streets right?
gingermolloy.
Thanks for explaining the selections.
i presume 1 loss every 30 wins would bring us in profit or at least even. And from your stats there is a 6% chance of a double street hitting 4 times or more in 10 spins. Thanks ginger :thumbsup:
tested this on 25,000 actual spins.
1386 wins
46 losses
( average 1 loss every 30)
without progression (playing only once):
1179wins
249 losses
(average 1 loss every 4.7)
Quote from: rss on June 01, 2009, 12:21:29 PM
interesting concept.
1 question though.. by streets, you mean double streets right?
sorry double streets yeah
An example:
after these spins
23
34
12
14
3
11
31
1
22
23
we have
4
6
2
3
1
2
6
1
4
4
1=2
2=2
3=1
4=3
5=0
6=2
so the bet would be
1 unit on all except double street 4.
however if we had this
1=1
2=2
3=1
4=3
5=0
6=3
we would not bet because double streets 4 and 6 have equal spins.
We would wait for one to dominate before betting.
I think this would be hard to track in the casino, but with excel its easy.
hope this is clearer
ginger
i did a sample test with double streets (see results above).
rss
could you explain you results a little better
cheers
ginger
sure,
tracked 10 spins, and bet all double streets, excluding the 1 that was most hit. If there were 2 streets that had most and equal hits each, tracked another group of 10 spins.
bet 2 step progression.
(as u explained)
I used the excel to calculate. maybe the only small difference that I made was.. that I had to skip 2 spins every time there was no bet. (I had to do this as it would have been too complex to program it in excel)
eg. track 10 spins.. if 2 streets have 3 hits each, restarted tracking at spin 13 instead of spin 11. I dont really think it makes much difference in the long term.
Ps: I tried to attach file, but it is 5Mb
Quote from: rss on June 01, 2009, 02:19:19 PM
tested this on 25,000 actual spins.
1386 wins
46 losses
( average 1 loss every 30)
without progression (playing only once):
1179wins
249 losses
(average 1 loss every 4.7)
in other words:
with 2 step progression, losses would have been 247 units
with 1 step flat betting 66 units.
which is not a lot considering that the sample was of 25000 spins. Perhaps with a little tweaking, better results can be achieved.
maybe a 3 step progression. 1-3-3 .. any ideas?
Quote from: rss on June 01, 2009, 04:21:36 PM
sure,
tracked 10 spins, and bet all double streets, excluding the 1 that was most hit. If there were 2 streets that had most and equal hits each, tracked another group of 10 spins.
bet 2 step progression.
(as u explained)
I used the excel to calculate. maybe the only small difference that I made was.. that I had to skip 2 spins every time there was no bet. (I had to do this as it would have been too complex to program it in excel)
eg. track 10 spins.. if 2 streets have 3 hits each, restarted tracking at spin 13 instead of spin 11. I dont really think it makes much difference in the long term.
Ps: I tried to attach file, but it is 5Mb
The tracking can work like that but I prefer to be t more spins, so the way i do it is to track the 10 previous spins for every spin, you still dont bet every spin but you bet a whole lot more than waiting 10 more spins everytime you dont get a single double street to leave out.
I have set up an excel sheet to do the tracking and have imported loads of numbers from random.org.
I have attached the spreadsheet
I hope it is clear.
ginger
just a couple of foot notes for the above data.
zero is techbically its own double street, but because it is the dominant one so little of the time, i.e. none of it. It is left out.
For that reason, the 10 spins after a 0 appears the numbers will add up to nine not ten, and in the unlikely event of multiple 0's in quick succsesion the numbers will be less than ten for a wile.
for this system i have decieded to just ignor this fact and concentrate on the 1-6 d streets.
AND
for the bankroll graph. I have only plotted a loss when the 2 step progression has failed. single losses are still plotted as a single uint win because the progression is grand martingale.
hope this is clear.
I will however awnser any questions.
ginger
times in last 10 frequency %RSD
0 537 4.22%
1 937 5.73%
2 829 5.89%
3 471 7.64%
4 144 21.14%
5 35 57.07%
6 5 124.90%
7 0
8 0
9 0
10 0
total spins = 2958
Above is the data that shows order in randomness.
Note the relative standard deveations (%error)
The numbers are derived from 3 sets of around 1000 spins.
The sets of data obtained are within 5% of each other for the low numbers.
The % RSD gets bigger when the frequency gets smaller, but we are concentrating on the highr frequency numbers.
The occurences of zero are taken into account in this data by giving it its own d-street.
I have assumed this has little affect on the data because as a percentage of spins it is neglagable. (2.7%)
I would hope the 2 point progression will eliminate it.
ginger
Its tanking
see attached
ginger
Quote from: gingermolloy on June 02, 2009, 04:34:03 AM
The tracking can work like that but I prefer to be t more spins, so the way I do it is to track the 10 previous spins for every spin, you still dont bet every spin but you bet a whole lot more than waiting 10 more spins everytime you dont get a single double street to leave out.
I have set up an excel sheet to do the tracking and have imported loads of numbers from random.org.
I have attached the spreadsheet
I hope it is clear.
ginger
I don't really think it makes much difference if you bet more, cos the average would still be 1 loss for every 30 wins. and the progression is to heavy
where do you get this 1 in 30 number from???
ginger
1386/46
(wins divided by losses)
+ 1 loss for every 30 wins
Is that 30 spins or 30 progressions???
So, overall how much profit was there made when you did and didn't double up?
Ok.. maybe this is simpler:
out of 25,000 spins,
won 1179 sessions on 1st step of progression
won 203 sessions on 2nd step of progression
lost 46 times
Thus: won on 1 st step of progression 1179 units
won on 2 nd step of progression 207*2= 414 units
Total wins= 1179 + 414= 1593 units
Losses= 46*40=1840 units
Thus balance = 1593-1840= -247 units
Hope this is clear enough..
Thanks Rss, thats very clear.
I think you are right, this has got potential, but needs tweaking.
How about flat betting all the d-streets that have appeared 0, 1 or 2 times in the last 10.
Nice and simple.
How would that look.
By the way, how did you do that 25000 spin test, my testing ststem in excel is crued and time consuming!!!
ginger
I don't think it will make much difference. (cos of the zero)
I am attaching a smaller file (with another sample of around 2000 spins) maybe you can check if it can be tweaked.
Top spreadsheet mate
I learnt some stuff there
cheers
ginger
np mate.
Those spins i sent you would have resulted in +15 units if flat betting for 1 spin was used. I really think that with some sort of tweaking, this has a lot of potential. Later on i will try using the 1-3-3 progression. cos wins are frequent and if, for example, there would be a sequence like llw, the damage would only be -17 instead of -40.
sounds good mate