Tiers et Tout + Reverse LabouchÄ—re

[Bigbroben]

I've been thinking lately about those two positive progressions.

Did a few rough drafts of progs that would include both aspects of these two.

Anyone thought of that before?

TnT allows for a loss before a win, whereas RL scraps 2 wins on a loss.

There must be a way to mix these in a safer but less volatile prog setup?

[The General]

The reverse labby can be a lot of fun to play, especially when you're out with friends and are willing to let it ride!

[Bigbroben]

Sure, but I suspect the table maximum would keep the thrill at a lower level, unless online.

Here, outside bets are minimum 10, max 100, so no way to let ride too high, especially with reverse labby's steep rise.  Hence the mix with tiers et tout: less destructive on a loss (or more than one), and a rise reduced to a third.

Real quick:
Say this starting line, where playing top and bottom 1st nrs, like revlab, and giving a second chance with the top and bottom 2nd column:

1   2
2   4
3   6
4   8
5   10  Total: 45

Let's start with a win ( all wins here mean a win at 1st or 2nd attempt, like tnt).
So, 1+5: +6u.
Write 6 in a tnt way: 2-4.
Let's wait for 2 wins before to modify the wagers.

After 2w:
1   2
2   4
3   6
4   8
5   10
---------
2     4
2     4
--------

If we wait to have a group of 2 rows before to rise the stakes, then adding only one of the 2 rows to the stake, next bet would be: again the top and bottom early lines (1+5, then 2+10) and a 2-4, so: 8-16.
Let'S pretend it's another 2 wins:
1   2
2   4
3   6
4   8
5   10
---------
2     4
2     4
--------
2    4  reserve 2
2    4  res. 2
-------

This was a great game, next will also be 2 wins.  We'll take 5+1 then one line of each group, so 5+1+2+2 (10) and their doubles to make it tiers et tout:
1   2
2   4
3   6
4   8
5   10
---------
2     4
2     4
--------
2    4  reserve 2
2    4  res. 2
-------
3    6  res. 1
3   6  res.1
----


...so on:
1   2
2   4
3   6
4   8
5   10
---------
2     4
2     4
--------
2    4  reserve 2
2    4  res. 2
-------
3   6 r1
3   6  r1
-----
4   8  r1
4   8  r1
-----
5   10  r2
5  10  r2

...so on.

I now must think over how to manage losses, what to play after a loss.

Did not calculate or compare odds & payout with revlab or tiers et tout, could end up the same, I suspect it is safer but less spectacular?

To be dug further...

[Ricky]

I am currently testing different moneny management and devised a way to use Labouchere and Reverse Labouchere on the inside bets. In testing Splits I found that Reverse Labouchere performs increadibly well as it allows winning streaks to be capitalized on and losing streaks significantly reduce the bet size and subsequent losses.

I have used RX to model long term a bet based on Dyksexlic's repeat in x spins. In this case it is repeat of Split in 19 spins. I use the Reverse Labby method to size my bets each spin. (I am thinking of modifying this to update the labby sequence after each cycle)
h
Here is one chart I generated. Although there was a very short periods of large drawdowns due to table limits being reached, I was amazed when it finally recovered and went on to make new highs. Obvously we would not play long term like this but the concept proves to have some merit long term.

[Bigbroben]

The down spikes suggest a Labouchere and not RevLab...

Just saying.
Thanks for reviving the thread.
I had worked on this a bit then got carried away by life..

[Ricky]

The down spikes suggest a Labouchere and not RevLab...

Just saying.
Thanks for reviving the thread.
I had worked on this a bit then got carried away by life..

Hi Bibbroben,
I have just reviewed the rules on the Reverse Labouchere, and I realize I am not implementing exactly the same process but the result is an amazing new progression. Let me just briefly explain the process I am following (which I now realize is not a Reverse Labouchere). I have applied this to Splits so it is not on an Even Chances Bet which I believe where the uniqueness lies

1. I define the amount of money I wish to make eg 200 units
2. I define an initial element size of my sequence eg 50 units
3. I create a sequence of size 50 units to get a total of 200 units

50 + 50 + 50 + 50

Now, originally I defined a Labouchere type bet and performed the following

Labouchere Rules
--------------------------
1. Betting on Splits paying 17:1 I added the first and last numbers.
           50 + 50 = 100 units
2. Since a payout on splits is 17:1 I divided the total by 17
                 100 /17 = 5.88 units
                 round to 6 units
3. As I am betting more than 1 split I need to take this amount into account.
          So If I am betting 5 splits, my cost will be 6 units x 5 = 30 units.
4. Add the cost of the bet to the original amount
                100 + 30 = 130 units
5. Now recalculate the unit size for a win
                130 / 17 = 7.65 units
                 round to 8

So now I bet 8 units on all splits selected

Possible Outcomes
----------------------------
1a. Win Split
            8 units x 17 - (5 splits x 8units)
           = 136 - 40 units
           = 96 units
1b Update sequence by REMOVING 96 units from the last and first numbers
          50 + 50 + 50 + 50
but the two number add up to 100 units not 96 units. So we have 4 units unaccounted for. To handle this I cross off as much of the 96 units as I can starting from the last number in the sequence. my New sequence is
           4+50 + 50 + 50 + 50
           4 + 50 + 50

2a Lose Split - add the loss to the sequence
        (5 splits x 8units)
           =  40 units
         
2b Update sequence by ADDING 40 units
          50+ 50 + 50 + 50 + 40

Reverse Labouchere Rules
--------------------------------------
Now this is where what I am doing is not exactly Reverse Labouchere

Use same rules as above  to define your sequence and get your bet size for a split

Possible Outcomes
----------------------------
1a. Win Split
            8 units x 17 - (5 splits x 8 units)
           = 136 - 40 units
           = 96 units
1b Update sequence by ADDING won units to the sequence
         50+ 50 + 50 + 50 + 96

2a Lose Split - Update sequence by REMOVING units lost from the last and first numbers
        = 5 splits x 8 units
        = 40 units
       
2b Update sequence by REMOVING 40 units
        50 + 50 + 50 + 50
but the two number add up to 100 units not 40 units. So we have 60 units unaccounted for. To handle this I cross off as much of the 40 units as I can starting from the last number in the sequence. my New sequence is
         50 + 50 + 50 + 50 +10
        new sequence is 50 + 50 + 50 + 10

Now because this is a reverse labouchere where a winning streak will result in the sequence growing we need to place a limit on how large the sequence can grow. So once this sequence has grown to a fixed number of elements or a fixed sum we reset the sequence

Now what I have done above does not make mathematical sence in that the end profit I will get once crossing off  the numbers does not equal my original goal but what I have realized is that you can control your bet size on say splits while getting a reqard of 17:1 for each win. This can be done for other payouts like straights paying 35:1 or streets paring 11:1

If anyone can improve on the idea or improve the rules we may help to come up with a new progression not yet thought of

Cheers,
Ricky

[Bigbroben]

Here is a little comparison on Tiers et Tout and Reverse Labouchère:

T&T:
Pros: 1) 73,6% chance of winning a step.  2) Compounding winning streaks at 30% or so.
Cons: BR lost after 2 consecutive losses.

RevLab:
Pros: 1)  Compounding at 60% or so on a win
Cons: Require a long range minimal ratio of 2W/1L to keep the run going.


So considering RevLab's high EC winrate requirement, and noting TnT's high success rate, I thought there might be a way to integrate both systems to avoid: 1) TnT's two-strike-you're-out, 2) RevLab's melting BR in a few losses.  This will come obviously at a cost: lower compounding effect.

So let's say the stash looks like this, for 30u total in this example, if the starting units are 1234:

T1:  1 2 3 4    We play this first RevLab way: 1&4= 5u.   
T2:  2 4 6 8    On a loss, we then try this lower row.  2&8=10u.

In both cases  of a win it's up +5.  The 5u are to be split in a TnT manner.  A second loss sees the 1/2 and 4/8 nrs striped.1 2 3 4
                                                                                                                                                                                                                      2 4 6 8
So a win can be written  like this:  1 2 3 4 1         or so: 1 1 2 3 4                The +2 is of course the leftover of 5u removed of 1+2u.
                                                            2 4 6 8 2  +2              2 2 4 6 8  +2          It can be banked in of left for later use.

The second play will then play  6 nrs:  1 1 2 3 4 , so 6units.  A win gets: 2 1 1 2 3 4
                                                                   2 2 4 6 8                                           4 2 2 4 6 8

So the third would play: 2 1 1 2 3 4 for 7u.  A win goes: 2 2 1 1 2 3 4
                                           4 2 2 4 6 8                                    4 4 2 2 4 6 8 +1
                Or simply put, since we won 18u (5+6+7): 1 2 3 4 5
                                                                                          2 4 6 8 10 +3
                An all-in for higher compound would include the +3 reserve: 1 2 3 4 6
                                                                                                                            2 4 6 8 12


Real quick, it seems it would compound at 17%  (5/30, 6/35, 7/41...)
Just an idea, I'd have to try it.
But I bet it would be fun to physically play with those piles of token by the wheel.  Don't need pen and paper either.

This is a first way of doing it.

I'll think of another one and be back.