Bet selections - are they really helpful?

[mr.ore]

I can't understand, whether bet selections are good for something.  It is something to be decided.  While I understand why well made progression in both bet size and risk can improve probability of winning, I can't see, how to construct bet selection and prove that it affects game.  I suppose perfectly random game.

If I want to have a nice session, I usually play this way:

I only play with one unit, never increase bet size, but instead of that when I want to win more I play options with higher payout, and if not winning for too long time, I play options with lower payout and higher probability of winning, in order to be more secure.  So I usually start with a sixline or a dozen, and if I lose I play squares, streets and splits, if too down then I play a number for some time, and then return to sixline and when win then try to play a number for a few spins again.  I'm trying to hover around my bankroll size and from time to time to go a little up.  I understand that this is not mathematically optimal way to maximize probability of reaching some fixed target, but I'm trying to balance between possible win, not losing and having some fun.  It sometimes works like a magic, and I would never play again a classical progressions, where you can bet in one spins as much units as you could play with for say sixty spins, and results are often almost the same.

I have observed that dozen I play on have from time to time a concentration of hits, where I try to play aggressively on streets and splits, and then is some time when the other two dozens are alternating.  It seems that dozens "tends" to alternate for some time, then all three are hitting, and then again only two of them.  It is logical, I wrote all possible combinations of dozens in three spins, and there are less of those where all three hits.  My strategy is based on attempt to survive negative dispersion with playing dozens and sixlines, and when good time comes to recoup on that concentration of hits.  The advantage of playing always with one unit is that I can have fun much longer, and I don't have to decide whether to increase the bet or not, the biggest mistake I can do is to stay too long on a number or a split.  I have no fixed set of rules, it seems better to play by "feel", it is also much more fun.

Now I'm thinking - a lot of people use some kind of "bet selection", which should help to select right place to play on.  I have experienced that best is to stay on one location and try to survive, because when "following" the game it is really easy to miss.  From time to time I play repeaters, it seems to work sometimes.

I think repeaters partially works because if probability of hitting is p, then probability of hitting after n spins is ((1-p)^(n-1))*p and if plotted most hits are concentrated in the beginning of the graph.  So there I can see at least some logic in playing repeaters, but I'm not really sure and might be wrong.  In the long run situations when number does not repeat often enough will more than balance out situations when it is helpful.  Could this be enough to state that "random events tends to repeat"?

I used to believe, that bet selection is useful, like Wrangler selection seems to eliminate chops and filter out big fluctuations on either red or right side.  But in the long turn it does not matter, and even in the short run it can be fatal.  I believed that such a bet selection can force bankroll graph to more copy expected value, I mean "units already bet"*HE, but now I'm not sure.

So I would like someone to PROVE that some bet selection can ASSURE that bankroll will be always as near expected value as possible, that the probability is higher this way that this would happen.  Right now I believe that no bet selection can in any way affect anything at all, and that player has to be hit by wave of negative dispersion and have a strategy to survive it.  Even if someone PROVED that bet selection actually do something at all, even if that would be negative for the player, would be enough.  Please give me someone an example of a really bad bet selection.

Can someone recommend me a bet selection on one dozen and somehow explain why it is more useful than just staying on one dozen? I'm willing to accept the fact that bet selections actually works.

[VLS]

Thanks for the post mr.ore,

I must say when doing this sort of "parachuting" (opening and reducing the layout coverage) it is advisable to  hold the bets for a full cycle.

I.e. you hold that double-street bet for 6 spins, the Dozen bet for 3 spins, the corner for 9 spins, etc.

I agree with the concept of decrementing the coverage and hunting more payout on concentration of hits and opening more when in dispersion.

Starting your bets on the double-streets is spot-on; reminds me of my "floating point" method for the layout.

It used a pluscoup +1 on a hit with the goal of breaking even, which means if you were doing 4 units on the corner and got a hit which put your balance 7 units behind form high point, you would go to the dozen with your unit size, since a win on a dozen with 4 units nets you +8, enough to get your bank up and less risking than holding. In case of a hit you would have netted more, but your primary goal is keeping the balance even, any win coming from a hit that nets more than highest point is a plus.

I actually liked the "going with the wheel" selection. When a session exhausted its cycle and I had to "go down" and use more numbers or after a win which means picking less coverage, I always picked the one with the most recent hits. No problem with dozens, hi-low and streets, but for some locations like corner bets and splits I found myself picking them looking for the latest-hit numbers hit.

For instance if I got a hit on the double-street 1-6 with number 4 and now I had to go to the corner bet, I could pick corner 1/5 or 4/8.. and it depended on whether which 1-2 or 7-8 were more recent in the marquee.

Cheers!
Victor