Even though roulette is a perfect independent results' game, there are some interesting long term features that could be easily tested by everyone.

The strategy was conducted over 1.500.000 real spins (single zero).

I mean some events are more likely than others. Unfortunately zero tax and some other practical features will lower a lot the value of such aknowledge.

The trigger we are looking for is really simple: we take note of the last number produced then we bet all the 3 EC belonging to this number. And this procedure is made per every last number sorted out.

For example number 33 sorted out, next spin we want to bet black, odd and high.

Obviously such betting will get 4 different outcomes (zero ignored):

- winning all 3 EC (sorting of 29,31,33,35): +3

- winning just one unit (sorting of 11,13,15,17,19,21,22,23,24,25,26,27,28): +1

- losing just one unit (sorting of 1,2,3,4,5,6,7,8,9,10,20,30,32,34,36): -1

- losing all 3 EC (sorting of 12,14,16,18): -3

Well, in the long run the number of spots winning all 3 bets are greater than the number of spots losing all 3 bets and it will increase the more the hands are played.

Of course to try getting the best of it from this finding needs also to take into account the spots where we'll either win or lose 1 unit.

Despite of what many may think, there are no better numbers or worse numbers to spot as triggers.

True, red-odd-low numbers or black-even-high numbers should have the theorical best probability to match the same EC on the next spin but this wasn't the case, at least on our quite long sample.

In a word and transferring the plan on the statistical field, we'll expect to have more single total different EC outcomes than streaks of different EC outcomes, more double different EC outcomes than 2+ streaks of different EC outcomes and so on. And the reverse is also true regarding the same EC situations (more streaks than single, more 2+ than doubles, etc).

The variance and the weight of zero could be quite high, yet the final result will be sure.

as.