Thanks to Bayes

[ozon]

A few days ago in another forum appeared post of Bayes, about a stable bet selection.
Indeed, this bet selection is very stable, unfortunately, I played with the progression of + 1 / -1 reset when reach new high. After making about 100units progression lost bankrol 200 units.
So I decided to use a different staking, which entered Bayes long ago
It is rather the whole strategy
Used bet selection with Bayes MM target profit of 100 units pers session stoplose 200 units.
We Know That MM will lose in some point, then if you lose  2 sessions  from 3 sessions the stakes rise by 1 unit, and so on if we win on level two, 2 sessions or more from 3 reduce the stakes, now we need Bankroll of 3000-3200 units. But i think that this strategy is very difficult to lose.

Bayes  posts

My preferred bet selection for the even chances involves betting against one of a "family" of patterns, but in a dynamic kind of way. I'll give an example in a moment but the family of patterns all contain a single, a series of 2, and series of 3. E.g.

(B) R BB RRR (B)
(R) BB R BBB (R)
(B) RRR B RR (B)

Note that the first and last outcomes (in brackets) are just "markers" or boundaries which tell me where in the pattern I am (this will become clearer in the example).

So each pattern (including markers) is 8 outcomes long. In any 8 spin sequence there are 28 = 256 possible permutations but only 12 of them form a pattern in which there is a 1,2,3. They are:

Starting with RED

1,2,3 = (B) R BB RRR (B)
1,3,2 = (B) R BBB RR (B)
2,1,3 = (B) RR B RRR (B)
2,3,1 = (B) RR BBB R (B)
3,1,2 = (B) RRR B RR (B)
3,2,1 = (B) RRR BB R (B)

Starting with BLACK (these are just mirror images of the above)

1,2,3 = (R) B RR BBB (R)
1,3,2 = (R) B RRR BB (R)
2,1,3 = (R) BB R BBB (R)
2,3,1 = (R) BB RRR B (R)
3,1,2 = (R) BBB R BB (R)
3,2,1 = (R) BBB RR B (R)

It doesn't matter which pattern you choose; just pick one and switch to another randomly (I generally switch when the current pattern isn't "performing" well, but you could just do it mechanically every fixed number of spins).

Now, say you have chosen the first pattern (starting with R). You are betting against R BB RRR. So your first bet would be on B.

If this wins, continue to bet on B until a loss.

If it loses, so you have R, the next bet will be on R (against BB forming).

If the bet on R loses the pattern is R B so you bet on R to prevent R BB.

And so on. This is straightforward enough when you have a series of losses, but where the method differs from the usual "anti pattern" system is what you do after a win. You don't recommence betting against the pattern from the first element but continue betting on the same side until a loss. This ensures that you're always on any trend which may occur.

E.g. Suppose I get 6 straight losses because R BB RRR has occurred. The pattern is not completed because it hasn't been "bounded" by a B. My next bet is therefore on R. If this wins I continue to bet R until a loss, then I bet on the B to continue. When R occurs I'm inside the first marker and so bet against the pattern as described earlier. It's simple enough but can be a bit confusing if you're only familiar with the "static" way of using this bet selection.

Example:

first bet on B

1.  R L (next bet on R)
2.  B L (next bet on R)
3.  R W (next bet on R)
4.  B L (next bet on R)
5.  B L (next bet on B)
6.  R L (next bet on B)
7.  B W (next bet on B)
8.  R L (next bet on R)
9.  R W (next bet on R)
10. R W (next bet on R)
11. R W (next bet on R)
12. B L (next bet on B)
13. B W (next bet on B)
14. B W (next bet on B)
15. R L (next bet on R)
16. B L (next bet on R)
17. R W (next bet on R)
18. R W (next bet on R)

I like this bet selection because it's simple, mechanical, and gives good results. I've experimented with other patterns involving longer streaks but the 1,2,3 seems to offer the best returns because you get some nice winning runs and losses are usually quite "choppy" (good for D'Alembert).

Of course there are times when the pattern you're betting against can repeat itself partially, in which case the losses can be running at 2-1 or more, but these sequences don't last long and are almost followed by a counter balance or nice win streak.


Quote from Bayes:

The progression is based on starting from a base of 1 unit and multiplying by around 10% - 12% to get the value of the next stake. Here is a version I just coded which is very similar to the Holloway progression using a percentage of 11.5% which gives approximately the same values and the same number of steps.

The error in the Holloway progression comes from rounding. You can see the same thing here: The middle column is the calculation unrounded and in the right column the numbers have been round DOWN to the nearest whole number. Notice that from step 27 (18u) to step 28 (21u) there is an increase of 3 units, but from step 28 to step 29 (23u) there is an increase of only 2 units!

Depending on how the rounding is done there will usually be some error like this, but it won't make a lot of difference. You could always change the values in the Holloway progression slightly to make the steps between values consistent.

1 1.115 1
2 1.243 1
3 1.386 1
4 1.546 1
5 1.723 1
6 1.922 1
7 2.143 2
8 2.389 2
9 2.664 2
10 2.970 2
11 3.311 3
12 3.692 3
13 4.117 4
14 4.590 4
15 5.118 5
16 5.707 5
17 6.363 6
18 7.095 7
19 7.911 7
20 8.821 8
21 9.835 9
22 10.966 10
23 12.227 12
24 13.633 13
25 15.201 15
26 16.949 16
27 18.898 18
28 21.072 21 << 3 from 18 to 21
29 23.495 23 << 2 from 21 to 23
30 26.197 26
31 29.209 29
32 32.568 32
33 36.314 36
34 40.490 40
35 45.146 45
36 50.338 50
37 56.127 56
38 62.581 62
39 69.778 69
40 77.803 77
41 86.750 86
42 96.726 96
43 107.850 107
44 120.253 120
45 134.082 134
46 149.501 149

[Priyanka]

Love it

[tuddilue]

Thanks,  interesting!
Do you know what Bayes calls it?
-Tuddilue

[ozon]

It does not have a name, a post about the bet selection is new, post about MM is old, progression and modification of sessions is from another system.

[beretta28]

Why your pattern should give you an advantage?
If it's a way to be discipled OK,but all patterns are the same