I just experiment with repeats using the principal of 1/3.
I use the formation of a dice with four patterns/formations.

XXX
XOX
XXO
OXX

I did never get 4 loses during 300 trails with today's random org file ...
That is pretty cool.

2    2    1
2    1    2
2    1    2 W
1    2    2 L
2    1    1 W
1    1    1 L
2    2    1 L
2    1    1 W
1    2    2 W
1    1    1 W
2    2    2 W
2    2    2 W
1    1    1 W
1    1    2 L
2    1    2 L
2    1    2 W
2    1    2 W
2    2    2 L
1    1    1 W
2    2    1 L
2    2    1 W
1    1    2 W
2    1    2 W
1    1    2 W
2    2    1 W
1    1    1 L
1    2    1 W
1    1    2 L
2    2    1 W
1    1    1 L
1    1    1 W
1    1    2 L
2    2    1 W
2    1    1 L
2    2    2 L
2    2    2 W
1    2    2 W
1    1    2 L
2    2    2 L
1    1    1 W
1    1    2 L
2    1    1 W
2    1    2 L
2    2    1 W
1    1    1 L
2    2    2 W
2    1    1 L
1    1    1 W
2    1    1 W
1    1    2 L
1    1    1 L
1    2    1 L
2    2    2 W
2    1    1 L
2    2    2 W
2    2    1 L
2    1    1 W
2    2    1 W
1    2    1 L
2    1    2 W
2    1    1 L
1    1    2 W
2    2    2 L
2    1    1 W
2    2    1 L
1    1    2 W
1    1    2 W
1    2    2 W
1    2    1 L
2    2    1 W
1    2    2 L
1    2    1 W
2    2    2 L
1    1    2 L
2    1    2 W
2    2    2 W
2    2    2 W
2    2    1 L
2    1    1 L
2    2    2 L
2    2    2 W
2    1    2 L
1    2    1 W
2    1    2 W
2    2    2 W
1    1    2 L
2    2    1 W
1    2    1 W
2    1    1 L
2    1    2 L
1    1    2 W
2    2    1 W
1    1    1 L
1    1    1 W
1    1    1 W
2    2    2 W
2    2    1 L
1    1    1 L
2    2    1 L
1    2    1 W

Nice for you.
But 300 trials is a very small sample.

We can use a method in 1000 trials and be very +.
And then in an other trial of only 200 spins we can be very -.

It was just for fun, no need to test it for any larger sample, as all bet selection lose in the long run.
That i old thinking, you can not beat the game based upon larger sample when testing bet selections.

Quote from: Sputnik on March 02, 2013, 11:44:45 AM
It was just for fun, no need to test it for any larger sample, as all bet selection lose in the long run.
That i old thinking, you can not beat the game based upon larger sample when testing bet selections.

Then the question is :
WHY are you testing bet selections if you already know that all bet selection lose in the long run.

Good question.

I find out is not so much about where you bet that is crucial - it is not that who result in a winning session or a losing session.

I start testing John Patrick's and Brett Morton's Money Management strategy's with out any specific system, just testing Follow The Last and avoid playing when the table is choppy.

I win two out of three and four out of five sessions.
With different set ups of MM plans.

Some is harder and some are more easy to achieve.
Its about finding the cut point between win goal targets and stop loss amount.

I think that you are contradicting yourself.And I am saying this to you with good intention. . I mean no war my friend.
In my point of view if roulette is beatable ,the way is to increase the accuracy of the predictions and play FLAT bet ONLY.

Its fine by me and you can't not get plus one to overcome the amount of attempts doing so.
But with regression you capitalize on dominance ...