I have recently been playing around with a new idea.
I have been using different bet methods like martingale, Alembert, Labouchere, pluscoup, etc... but instead of basing them on each bet, I play a mini-game and each game represents a unit either won or lost in the global bet method.
In other words if I'm playing an even chance, say Red, with a flat bet and when I reach either +5 or -5 that ends the game. That "GAME" becomes the 1st unit won or lost in my global bet method.
So, If I am playing Alembert (+1 on a loss, -1 on a win) as my global bet method and each unit is made up of a game to +5 or -5 and I play 25 spins and end up at -5 in my flat bet even chance mini-game, then that represents -1 in my global method. My next bet will be at 2 units in my mini-game of flat betting an even chance. Now my min-game will be at 2 units per bet and will end when I get to +10 or -10 (this is really my basic unit size times 2. My basic unit size could be $1, $5, $10, $12 etc...)
If I lose my next mini-game, I add 1 to my game size and will be betting my basic unit size times 3. I will play until I am +5 times 3 or - 5 times 3. If I reach +5 times 3 or +15 if my unit size is 1, my next mini-game will be played at 1 times 2, etc...
This idea can be use to give an added element of safety to almost any system that can be played in roulette. It works best with systems that can be played to + or - the same number of units.
Think about it. I think you'll like it.
George
You should not play like that, you are only losing more without gaining anything. If you flat bet 5 units bankroll to get 5 more (so your target is 10 units) with 1 unit bet size, then your probability of winning is only 43.2825%. It is much worse than 48.6486% if you bet five units on even chance.
Look at this table, where it is clear, that flat betting is not worth it:
bankroll | bet | W | L | Pwin |
0 | no bet | - | - | 0.0000% |
1 | 1 unit on 1:1 | 2 | 0 | 7.7466% |
2 | 1 unit on 1:1 | 3 | 1 | 15.9236% |
3 | 1 unit on 1:1 | 4 | 2 | 24.5548% |
4 | 1 unit on 1:1 | 5 | 3 | 33.6656% |
5 | 1 unit on 1:1 | 6 | 4 | 43.2825% |
6 | 1 unit on 1:1 | 7 | 5 | 53.4336% |
7 | 1 unit on 1:1 | 8 | 6 | 64.1488% |
8 | 1 unit on 1:1 | 9 | 7 | 75.4592% |
9 | 1 unit on 1:1 | 10 | 8 | 87.3980% |
10 | no bet | - | - | 100.0000% |
Bankroll is your actual bankroll, you start at 5. If you win (W) you go to 6, on lose (L) to 4. Pwin means what is overall probability you get to 10 before you get to 0.
Rather try this as your minigame:
bankroll | bet | W | L | Pwin |
0 | no bet | - | - | 0.0000% |
1 | 1 unit on 8:1 | 9 | 0 | 9.6347% |
2 | 1 unit on 8:1 | 10 | 1 | 19.4039% |
3 | 1 unit on 5:1 | 8 | 2 | 29.0382% |
4 | 3 units on 2:1 | 10 | 1 | 38.9424% |
5 | 1 unit on 5:1 | 10 | 4 | 48.8436% |
6 | 2 units on 2:1 | 10 | 4 | 58.7449% |
7 | 3 units on 1:1 | 10 | 4 | 68.6461% |
8 | 1 unit on 2:1 | 10 | 7 | 78.8149% |
9 | 1 unit on 1:1 | 10 | 8 | 89.1212% |
10 | no bet | - | - | 100.0000% |
probability of doubling your bankroll is higher than flat betting five units on even chance. You have 5 units and bet 1 unit on line, if it loses your bankroll is 4 and you bet 3 units on dozen, and if that loses, your bankroll is 1, and you bet 1 unit on corner, if it loses, session ends, if it wins, your bankroll is 9, and you bet 1 unit on ec, if it loses then bet 1 unit on dozen, if that loses then bet 3 units on ec, if that loses bankroll is 4 again, and you bet 3 units on dozen and so on until you reach your target or lose 5 units.
Table is simple to follow, W says on what line/bankroll to go if you win, L says on what line to go if you lose. Useful feature of this little system - max. bet size is 3 units.
If you want your unit size to be 1, then this is still better than betting five units on ec, but worse than previous method:
bankroll | bet | W | L | Pwin |
0 | no bet | - | - | 0.0000% |
1 | 1 unit on 8:1 | 9 | 0 | 9.6217% |
2 | 1 unit on 8:1 | 10 | 1 | 19.3924% |
3 | 1 unit on 5:1 | 8 | 2 | 28.9905% |
4 | 1 unit on 5:1 | 9 | 3 | 38.7220% |
5 | 1 unit on 5:1 | 10 | 4 | 48.6589% |
6 | 1 unit on 2:1 | 8 | 5 | 58.3634% |
7 | 1 unit on 2:1 | 9 | 6 | 68.3000% |
8 | 1 unit on 2:1 | 10 | 7 | 78.5810% |
9 | 1 unit on 1:1 | 10 | 8 | 89.0011% |
10 | no bet | - | - | 100.0000% |
Mr. Ore,
Thanks for your very interesting method of trying to turn 5 units into 10 units. I will give this a test since it looks like an interesting way to play.
I wasn't really suggesting that anyone use my flat bet on Red as a valid means of winning or losing 5 units. I was just using a simple example to try to explain that my real post is the idea of playing mini-games within a larger framework that controls how large or small our unit size in a mini-game should be.
Let's say we play your 1st example of a better way to play and we bet $1 unit size. But we don't just play $1 for mini-game after mini-game. We use the classic Alembert bet method as a control to determine when to change our mini-game unit size.
So we have global units represented by a mini-game and betting units represented by the units in the mini-game. A loss of a mini-game is a -1 unit in our global bet structure. So a lost mini-game = -1 in our global structure and the rules of our Alembert global structure tell us to increase our mini-game unit size by 1 unit. This means that our next mini-game will be played at $2 units. If we win this mini-game we will have recovered the previous $5 loss and be ahead $5 because we have $5X2=$10-$5=+$5.
Had we lost the 2nd mini-game at $2 units, we would have a minus for our global structure and our Alembert rules say that on a loss we add 1 unit to our next bet (mini-game) so our next mini-game will be played at our base unit size times 3 or $1 X 3 = $3. We continue to increase and decrease the size of our mini-game units per Alembert rules until we are at +$5 at which time we reset our global structure to 1, our base bet size.
I'm not suggesting that we use Alembert as our global structure, it's just a common example that people can understand without having to go into a lot of teaching. A safer global structure would be the +2/-2 or +3/-3 where we don't increase the size of our mini-game unit until we have lost 2 or 3 in a row etc...
This is an adaptation of the idea many authors use when they present their systems and they give you 3 levels of play. You play at the basic level as long as you win. On a loss you increase the size of your unit for quicker recovery of the lost units. Their argument is that the odds of losing 2 times in a row with their system is very small. If you do lose two games in a row, they have a third level to go to that raises their unit size much more so a win at the 3rd level will recover from the 2 previously lost games. The odds of losing 3 times in a row are so remote as to be beyond the realm of possibility, so they say.
At first I thought I had come up with a new idea. Now I realize it's just a tweak on an existing idea and not so clever after all. I do think it's an improvement. But that remains to be seen.
Once again, thanks for taking the time to respond. Your input is always logical, mathematical, worthy of consideration and very appreciated.
George
I understood what you meant. But I just want to say, that hardly any method will do any better than the second method I posted there, if you have that five starting units. Classical systems like Labouchere, having only 5 units starting bankroll, have lower chance of doubling them than this, I'm not sure how bad are they, though.
Good roulette system should have probability of winning x units with starting bankroll sb very near (36/37)*sb/(sb+x), and you can hardly have it higher, only by very little small amount like half a percent, not enough to overcome house edge.
Look at this: VLS' 50% Money Management
link:://rouletteforum.cc/vls%27-view-b57/vls-50percent-money-management/msg25947/ (link:://rouletteforum.cc/vls%27-view-b57/vls-50percent-money-management/msg25947/)
It is similar idea with intersession bankrolls. You save half of a won money as your win, and the other half is divided again to two halfs, first is used to increase unit size in a session you will play next, and the second for cushions bankroll with minimal unit size. If you win several sessions in a row, your unit will become bigger, and you get more cushion bankrolls, then if you lose, you use minimal bankroll and play with that, again inflating unit size. If you lose several sessions in a row, you lose them with minimal unit size. It is positive progression. Might work nice with no-zero roulette, on normal roulette you need little more luck with that. You can experiment with kept % and how you split them, maybe after several wins play with cushion bankrolls, and after losing it, play with inflated one.
Well put Mr. Ore. I presumed you knew what I was up to. I just wanted to clarify for the sake of others reading who may not have spent as much time tinkering with roulette.
I appreciate the idea of using Victor's money management method in this situation.
I have been testing this idea with excellent results on the "Very Near Infallible Roulette System" posted in the General Discussion section. The stretched pluscoup progression I recommend is also a positive progression. Playing the VNIRS takes about 40 spins on average to either win or lose 10 units which is a lengthy mini-game, but I have been averaging about 55% wins which keeps me in the positive betting according to the pluscoup.
If my win average stays at this level, a big if, I am going to a more aggressive global bet method to increase profits without increasing risk too much.
I will look at how to most effectively apply Victor's method to the VNIRS.
Thanks,
George
Mr.Ore
Bankroll:5 units
Win goal: + 5 units
According with all mathematicians and experts of the web,the highest probability to get it,is playing 5 units in one spin on an EC.
I could give you a lot of links (in French,Italian,German)
Are you sure you are not mistaken with your "second" method?
I am sure I am not mistaken. It works because on average you bet less than five units to reach your target, so they are not exposed to house edge. It is same idea as instead of betting two dozens, bet 1 unit on ec and if it lose, bet 1 unit on dozen. If you win, you are 1 unit up, if you lose, you are two units down. Probability of winning is 18/37 + 19/37*12/37 = 0.65303140978816654492.
If you bet on two dozens, each of them 1 unit, probability of winning is 24/37 = 0.64864864864864864865.
0.65303140978816654492-0.64864864864864864865=0.00438276113951789627
That is a difference of 0.4%, not much, but it is there.
We have two units, want to have 3 units. If we avoid betting one unit sometimes, probability of winning is increased. You do not bet on dozen, if you win on even chance. If you bet two dozens, you ALWAYS lose one unit while winning two.
Another example:
Let's have two methods to win one unit with five units bankroll, both are based on probability:
method 1: bet 1 unit on five lines, we cover 5*6 numbers.
Probability to win is p(method 1 wins) = 5*6/37=0.810810.
method 2: bet 1 unit on even chance, if it loses bet 1 unit on dozen, if that also loses bet last 3 units on even chance. Anytime it wins, profit will be 1 unit, while still risking 5 units bankroll like with method 1. Probability is different:
p(method 2 wins) = 18/37+19/37*12/37+19/37*25/37*18/37 = 0.821826
method 2 is better than method 1
Difference is 0.821826 - 0.810810 = 0.011016. That is 1.1% difference in probability. There are those little nasty tricks in gambling. Similar trick can be used in craps, for example. You will still lose in the end, and the difference is actually negligible. But you will play more spins this way for same price :)
Another example:
bankroll | bet | W | L | Pwin |
0 | no bet | - | - | 0.0000% |
1 | 1 unit on 35:1 | 36 | 0 | 1.5907% |
2 | 1 unit on 35:1 | 37 | 1 | 3.1838% |
3 | 3 unit(s) on 17:1 | 54 | 0 | 4.8236% |
4 | 3 unit(s) on 17:1 | 55 | 1 | 6.4227% |
5 | 5 unit(s) on 11:1 | 60 | 0 | 8.1081% |
6 | 1 unit on 35:1 | 41 | 5 | 9.7049% |
7 | 1 unit on 35:1 | 42 | 6 | 11.3037% |
8 | 3 unit(s) on 17:1 | 59 | 5 | 12.9716% |
9 | 3 unit(s) on 17:1 | 60 | 6 | 14.5857% |
10 | 10 unit(s) on 5:1 | 60 | 0 | 16.2162% |
11 | 6 unit(s) on 8:1 | 59 | 5 | 17.8350% |
12 | 6 unit(s) on 8:1 | 60 | 6 | 19.4665% |
13 | 1 unit on 35:1 | 48 | 12 | 21.0744% |
14 | 4 unit(s) on 11:1 | 58 | 10 | 22.7067% |
15 | 9 unit(s) on 5:1 | 60 | 6 | 24.3474% |
16 | 4 unit(s) on 11:1 | 60 | 12 | 25.9963% |
17 | 1 unit on 35:1 | 52 | 16 | 27.6124% |
18 | 1 unit on 35:1 | 53 | 17 | 29.2307% |
19 | 5 unit(s) on 8:1 | 59 | 14 | 30.8554% |
20 | 8 unit(s) on 5:1 | 60 | 12 | 32.5260% |
21 | 1 unit on 35:1 | 56 | 20 | 34.1531% |
22 | 1 unit on 35:1 | 57 | 21 | 35.7833% |
23 | 1 unit on 35:1 | 58 | 22 | 37.4180% |
24 | 1 unit on 35:1 | 59 | 23 | 39.0576% |
25 | 7 unit(s) on 5:1 | 60 | 18 | 40.7068% |
26 | 2 unit(s) on 17:1 | 60 | 24 | 42.3517% |
27 | 3 unit(s) on 11:1 | 60 | 24 | 43.9988% |
28 | 4 unit(s) on 8:1 | 60 | 24 | 45.6459% |
29 | 1 unit on 17:1 | 46 | 28 | 47.2638% |
30 | 6 unit(s) on 5:1 | 60 | 24 | 48.9401% |
31 | 1 unit on 17:1 | 48 | 30 | 50.5628% |
32 | 2 unit(s) on 11:1 | 54 | 30 | 52.2074% |
33 | 3 unit(s) on 8:1 | 57 | 30 | 53.8626% |
34 | 13 unit(s) on 2:1 | 60 | 21 | 55.5088% |
35 | 5 unit(s) on 5:1 | 60 | 30 | 57.2201% |
36 | 1 unit on 17:1 | 53 | 35 | 58.8563% |
37 | 2 unit(s) on 11:1 | 59 | 35 | 60.5332% |
38 | 2 unit(s) on 11:1 | 60 | 36 | 62.1923% |
39 | 4 unit(s) on 5:1 | 59 | 35 | 63.8464% |
40 | 4 unit(s) on 5:1 | 60 | 36 | 65.5283% |
41 | 1 unit on 17:1 | 58 | 40 | 67.1897% |
42 | 1 unit on 17:1 | 59 | 41 | 68.8596% |
43 | 1 unit on 17:1 | 60 | 42 | 70.5429% |
44 | 2 unit(s) on 8:1 | 60 | 42 | 72.2261% |
45 | 3 unit(s) on 5:1 | 60 | 42 | 73.9094% |
46 | 1 unit on 11:1 | 57 | 45 | 75.5767% |
47 | 1 unit on 11:1 | 58 | 46 | 77.2541% |
48 | 6 unit(s) on 2:1 | 60 | 42 | 78.9592% |
49 | 1 unit on 11:1 | 60 | 48 | 80.6652% |
50 | 2 unit(s) on 5:1 | 60 | 48 | 82.3712% |
51 | 1 unit on 8:1 | 59 | 50 | 84.0697% |
52 | 1 unit on 8:1 | 60 | 51 | 85.7919% |
53 | 1 unit on 5:1 | 58 | 52 | 87.4903% |
54 | 3 unit(s) on 2:1 | 60 | 51 | 89.2363% |
55 | 1 unit on 5:1 | 60 | 54 | 90.9817% |
56 | 2 unit(s) on 2:1 | 60 | 54 | 92.7272% |
57 | 3 unit(s) on 1:1 | 60 | 54 | 94.4727% |
58 | 1 unit on 2:1 | 60 | 57 | 96.2653% |
59 | 1 unit on 1:1 | 60 | 58 | 98.0822% |
60 | no bet | - | - | 100.0000% |
To double 30 units to 60, probability is 48.9401%. If you look at this operation as on a "flat bet 30 units on even chance", then you could say, that house edge is -2*48.9401%+100%=2.1198%. But it is not true, because you do not usually bet all 30 units before you win.
Thanks,very clear.
I'm playing roulette looking at these tricks too.
Do you know the "Bold strattegy" of Dubins and Savage?
If Yes,do you agree?
If not,could you tell me the best way to increase your bankroll by 10%(Bankroll 10 units,win goal + 1 unit)?
Thanks in advance
Mr.Ore
I'm happy to have finally met someone that "speaks my language" playing roulette.
1)Your example 30 units to 60 units is not clear.
What the first bet?What the second if you lose or win the first bet and so on...?
2)Your approch is clear:less you are exposed to house hedge,better it is.
But in ALL your examples ,if you have a long streak of W and L,without reaching your goal yet,the house hedge could be higher and higher because you are obliged to play a lot of spins.
Am I wrong?
Yes, I know what a bold policy is. Those are simple bold policy. Simple because there is always bet only on one location.
Policy to win 11 units:
bankroll | bet | W | L | Pwin |
0 | no bet | - | - | 0.0000% |
1 | 1 unit on 8:1 | 9 | 0 | 8.7203% |
2 | 1 unit on 8:1 | 10 | 1 | 17.5148% |
3 | 1 unit on 8:1 | 11 | 2 | 26.4321% |
4 | 1 unit on 5:1 | 9 | 3 | 35.2262% |
5 | 3 unit(s) on 2:1 | 11 | 2 | 44.2668% |
6 | 1 unit on 5:1 | 11 | 5 | 53.3046% |
7 | 2 unit(s) on 2:1 | 11 | 5 | 62.3424% |
8 | 3 unit(s) on 1:1 | 11 | 5 | 71.3802% |
9 | 1 unit on 2:1 | 11 | 8 | 80.6623% |
10 | 1 unit on 1:1 | 11 | 9 | 90.0698% |
11 | no bet | - | - | 100.0000% |
For comparison, even chance only version - anti+Martingale:
bankroll | bet | W | L | Pwin |
0 | no bet | - | - | 0.0000% |
1 | 1 unit on 1:1 | 2 | 0 | 8.1882% |
2 | 2 unit(s) on 1:1 | 4 | 0 | 16.8313% |
3 | 3 unit(s) on 1:1 | 6 | 0 | 25.7125% |
4 | 4 unit(s) on 1:1 | 8 | 0 | 34.5976% |
5 | 5 unit(s) on 1:1 | 10 | 0 | 43.7549% |
6 | 5 unit(s) on 1:1 | 11 | 1 | 52.8534% |
7 | 4 unit(s) on 1:1 | 11 | 3 | 61.8523% |
8 | 3 unit(s) on 1:1 | 11 | 5 | 71.1174% |
9 | 2 unit(s) on 1:1 | 11 | 7 | 80.4107% |
10 | 1 unit on 1:1 | 11 | 9 | 89.9406% |
11 | no bet | - | - | 100.0000% |
You start on line with same number as your starting bankroll and according to W,L columns you continue. If you win, go on line in W column, if you lose, go to line with L column. At any point, you know your probability of winning. In this case, start on line 10 to win 11 units.
Number of spins is on average low enough, so house dos not eat so much, that's what I think. While my policy improvement algorithm is not always that good and needs some tuning, computed probabilities are correct, it is just a matrix multiplied many times. So if algorithm finds a policy, it might not be absolutely optimal, but if I see that computed probabilities are higher than Marty, then I know it is very near.
Also remember that those are just fancy versions of parachute...
Thanks
I'm afraid in your EC version there is a mistake.
In line 8, L is 4 and not 5,if I have well understood.
As far as your policy to win 11 units is concerned,I'm still convinced that the house hedge,for all bets at 2,70% level, instead of 1,35%(all bets EC version),offset the advantage you mention(90,0698% vs 89,9406%) in particular,long and unlucky sessions!
I must calculate how often these bad sessions occur.
There is no mistake - on line for bankroll 8 we bet 3 units. After L you have 5 units, so L is 5. There cannot be any mistake, tables are computer generated, it is faster ;)
If you have access to a roulette with imprison/la partage option, then you should probably play only even chances. The tables are for single zero wheel without imprison or la partage option, outsides are also exposed to 2.7% house edge.
Greatest differences with those policies are on american roulette.
Same table for american wheel
bankroll | bet | W | L | Pwin |
0 | no bet | - | - | 0.0000% |
1 | 1 unit on 8:1 | 9 | 0 | 8.3705% |
2 | 1 unit on 8:1 | 10 | 1 | 16.8811% |
3 | 1 unit on 8:1 | 11 | 2 | 25.6305% |
4 | 1 unit on 5:1 | 9 | 3 | 34.1394% |
5 | 3 unit(s) on 2:1 | 11 | 2 | 43.1292% |
6 | 1 unit on 5:1 | 11 | 5 | 52.1088% |
7 | 2 unit(s) on 2:1 | 11 | 5 | 61.0884% |
8 | 3 unit(s) on 1:1 | 11 | 5 | 70.0680% |
9 | 2 unit(s) on 1:1 | 11 | 7 | 79.5202% |
10 | 1 unit on 1:1 | 11 | 9 | 89.2212% |
11 | no bet | - | - | 100.0000% |
Marty on american wheel:
bankroll | bet | W | L | Pwin |
0 | no bet | - | - | 0.0000% |
1 | 1 unit on 1:1 | 2 | 0 | 7.3917% |
2 | 2 unit(s) on 1:1 | 4 | 0 | 15.6047% |
3 | 3 unit(s) on 1:1 | 6 | 0 | 24.2805% |
4 | 4 unit(s) on 1:1 | 8 | 0 | 32.9433% |
5 | 5 unit(s) on 1:1 | 10 | 0 | 42.1392% |
6 | 5 unit(s) on 1:1 | 11 | 1 | 51.2588% |
7 | 4 unit(s) on 1:1 | 11 | 3 | 60.1476% |
8 | 3 unit(s) on 1:1 | 11 | 5 | 69.5470% |
9 | 2 unit(s) on 1:1 | 11 | 7 | 79.0251% |
10 | 1 unit on 1:1 | 11 | 9 | 88.9606% |
11 | no bet | - | - | 100.0000% |
Some simulations of that 10 bankroll to win 10% system.
Sorry,I miscalculated.Your table is perfect.I'm not stronger than a computer......
In fact all my comments or doubts were based on the fact that I'm in Europe and I know and play only roulette,one zero,partage rule.
That's why your tables seemed to me a bit strange,mainly bankroll 10 units win goal + 1 unit.
In the book of Dubins and Savage(How to play ,if you must),that take into consideration only European roulette,partage rule,the best strategy for 10 units to 11 units is "to play the minimum amount of units on EC,that allows to reach the goal in the lowest number of spins,that is:
1 unit,if lose 2 units,if lose 4 units,if lose 3 units.If lose again,game over.If win last 3 units bet,than bet 5 units and so on with the same principle that Dubins and Savage call "the brave and risky strategy" or something like that.
What I don't understand is that playing this way the % of success is 89,9406,exactly the same % you mention with NO partage rule in your EC version.(???!!!)
Quote from: beretta28 on Mar 28, 09:31 AM 2011
Sorry,I miscalculated.Your table is perfect.I'm not stronger than a computer......
In fact all my comments or doubts were based on the fact that I'm in Europe and I know and play only roulette,one zero,partage rule.
That's why your tables seemed to me a bit strange,mainly bankroll 10 units win goal + 1 unit.
In the book of Dubins and Savage(How to play ,if you must),that take into consideration only European roulette,partage rule,the best strategy for 10 units to 11 units is "to play the minimum amount of units on EC,that allows to reach the goal in the lowest number of spins,that is:
1 unit,if lose 2 units,if lose 4 units,if lose 3 units.If lose again,game over.If win last 3 units bet,than bet 5 units and so on with the same principle that Dubins and Savage call "the brave and risky strategy" or something like that.
What I don't understand is that playing this way the % of success,according to the book and my calculations and experience, is 89,9406,exactly the same % you mention with NO partage rule in your EC version.(???!!!)
To make thinks clear, my software policy generator cannot use partage rule right now. I still wonder if with big bankrolls and targets it would always prefer even chance, just because there is lower house edge. In theory we should only use the options with lowest house edge, but I have seen in some article someone combined two worse bets on craps and obtained better probability than with don't pass, so who knows.
Look there: link:://wtwii.wordpress.com/2006/08/24/debunking-a-craps-system/ (link:://wtwii.wordpress.com/2006/08/24/debunking-a-craps-system/)
Quote
More interesting is a case (which can also be called a system) in which playing two bets simultaneously in a certain manner results in a lower house advantage. A place bet on the six or on the eight is well known for having a vig of 1.52% (calculated as ((5/11)*7+(6/11)*(-6))/6). However, when you play a place bet on the eight simultaneously and meticulously put both bets up at the same time and remove them both when either wins, the house advantage is only 1.04% (calculated as ((10/16)*7+(6/16)*(-12))/12)!
Anything but Seven
Another similar case occurs with the so-called Mensa Anything But Seven system. You place a one-roll bet for $22 ($6 each on the place 6 and 8, $5 on the field, and $5 on the place 5). The house advantage on this bet (verified here - link:://wizardofodds.com/askthewizard/109 (link:://wizardofodds.com/askthewizard/109)) is only 1.136%, less than any of the individual bets placed separately.
By the way, what happens with imprison rule, if zero(es) hits, bets are frozen, and next spin another zero hits? Are they lost or imprisoned again, until something other than zero(es) hits? On win they are returned if I remember correctly, right?
They are imprisoned again and again....
That's why is better to ask the dealer to give half bet back,at the first zero!
Does imprison affects also dozens and columns? They are also outside bets.
I heard also something like that,concerning roulette,but I didn't found anything about that.
I know that is only valable if you play a low number of spins(how many?who knows?) and if the ratio win goal/bankroll is a certain % that I don't know too,but also very low.
I'm wondering if is not called "Parrondo paradox".
That's why I 'm so interested in your tables.I thought that was what I've been looking for several months.
I'm still wondering why % of 89,9406 for EC Marty is the same in your tables(NO partage option,you said-sure?-) and in Dubin and Savage book(partage option taken into account)
Your example for craps is very interesting.
Unluckily I don't know craps very well(in Europe it doesn'exist partically)
NO,partage rule exists only for EC.
Dozens and colums are like lines,streets,squares and so on...no partage option!
Maybe they did not use that rule, maybe the difference is so negligible, that it is not visible in a few first digits of computed probability number.
An additional point for roulette is:
-if 1000 players play at least 20000 spins on EC and 1000 players play 20000 spins on an given straight number,both flat bet,at the end of 20 thousand spins ZERO EC players will have a positive balance(all losers),but at least 5 players of straight number will still earn money.
That's why EC are not the best bet in certain situations of BKR,wingoal,etc
It is obvious, singles have higher variance. After 20000 spins with en prison rule, you have no chance to be positive flat betting. On a number, the chance exists, maybe even on american roulette.
For a nearly optimal play strategy, just do a progression on a single, or with lower targets make a parachute. Nearly optimal strategy is to win at least your target bankroll size in one hit with as few units as possible. If there are more possibilities, just select the one which is as near your target as possible, it's the one which wins minimal amount of units. You can also cap your target bankroll size to force probability stay near optimal strategy. It is that simple.
Yes,the problem is that the 995 losing players on straight numbers,lose a huge amount of money,much more than each EC players.
The nearly optimal strategy you mention is what I play in Montecarlo Casino.
My win goal is very low(no greed),15% at the most of my bankroll.
Progression:1,2,4,8,10,12,14,16,18,20:105 units(Marty 4 terms and D'Alembert 6 terms).
Risk of ruin/bankruptcy 1,2% or 1 every 83 attacks
(you are expert in Computers,can you check these % as risk of ruin?Thanks)
I don't trust in bet selection:IMHO it doesn't exist.
An alternative is "the bold strategy".
Parachute is dangerous,but I use it just for fun when 15% of bkr is achieved.
I must study your tables and reflect on them.......
If some good ideas i ÃÆ'Ã,¬ll write to you again.
If you have 105 units for your progression, then your target = 105*1.15=120 units. You can try this strategy instead, start on a line with bankroll 105. It is should be very near optimal method.
bankroll | bet | W | L | Pwin |
0 | no bet | - | - | 0.0000% |
1 | 1 unit on 35:1 | 36 | 0 | 0.7902% |
2 | 2 unit(s) on 35:1 | 72 | 0 | 1.5918% |
3 | 3 unit(s) on 35:1 | 108 | 0 | 2.4124% |
4 | 3 unit(s) on 35:1 | 109 | 1 | 3.2048% |
5 | 3 unit(s) on 35:1 | 110 | 2 | 4.0083% |
6 | 3 unit(s) on 35:1 | 111 | 3 | 4.8302% |
7 | 3 unit(s) on 35:1 | 112 | 4 | 5.6249% |
8 | 1 unit on 2:1 | 10 | 7 | 6.4303% |
9 | 3 unit(s) on 35:1 | 114 | 6 | 7.2539% |
10 | 10 unit(s) on 11:1 | 120 | 0 | 8.1081% |
11 | 1 unit on 35:1 | 46 | 10 | 8.9005% |
12 | 2 unit(s) on 35:1 | 82 | 10 | 9.7058% |
13 | 3 unit(s) on 35:1 | 118 | 10 | 10.5402% |
14 | 3 unit(s) on 35:1 | 119 | 11 | 11.3362% |
15 | 3 unit(s) on 35:1 | 120 | 12 | 12.1462% |
16 | 6 unit(s) on 17:1 | 118 | 10 | 12.9722% |
17 | 6 unit(s) on 17:1 | 119 | 11 | 13.7719% |
18 | 6 unit(s) on 17:1 | 120 | 12 | 14.5866% |
19 | 9 unit(s) on 11:1 | 118 | 10 | 15.4042% |
20 | 20 unit(s) on 5:1 | 120 | 0 | 16.2162% |
21 | 9 unit(s) on 11:1 | 120 | 12 | 17.0270% |
22 | 12 unit(s) on 8:1 | 118 | 10 | 17.8363% |
23 | 12 unit(s) on 8:1 | 119 | 11 | 18.6433% |
24 | 12 unit(s) on 8:1 | 120 | 12 | 19.4674% |
25 | 1 unit on 35:1 | 60 | 24 | 20.2644% |
26 | 2 unit(s) on 35:1 | 96 | 24 | 21.0763% |
27 | 8 unit(s) on 11:1 | 115 | 19 | 21.8901% |
28 | 8 unit(s) on 11:1 | 116 | 20 | 22.7085% |
29 | 8 unit(s) on 11:1 | 117 | 21 | 23.5258% |
30 | 18 unit(s) on 5:1 | 120 | 12 | 24.3481% |
31 | 11 unit(s) on 8:1 | 119 | 20 | 25.1681% |
32 | 8 unit(s) on 11:1 | 120 | 24 | 25.9970% |
33 | 1 unit on 35:1 | 68 | 32 | 26.7966% |
34 | 2 unit(s) on 35:1 | 104 | 32 | 27.6140% |
35 | 5 unit(s) on 17:1 | 120 | 30 | 28.4374% |
36 | 1 unit on 35:1 | 71 | 35 | 29.2381% |
37 | 2 unit(s) on 35:1 | 107 | 35 | 30.0578% |
38 | 7 unit(s) on 11:1 | 115 | 31 | 30.8623% |
39 | 7 unit(s) on 11:1 | 116 | 32 | 31.6963% |
40 | 16 unit(s) on 5:1 | 120 | 24 | 32.5267% |
41 | 1 unit on 35:1 | 76 | 40 | 33.3294% |
42 | 7 unit(s) on 11:1 | 119 | 35 | 34.1604% |
43 | 7 unit(s) on 11:1 | 120 | 36 | 34.9755% |
44 | 2 unit(s) on 35:1 | 114 | 42 | 35.7914% |
45 | 15 unit(s) on 5:1 | 120 | 30 | 36.6160% |
46 | 2 unit(s) on 35:1 | 116 | 44 | 37.4264% |
47 | 2 unit(s) on 35:1 | 117 | 45 | 38.2528% |
48 | 9 unit(s) on 8:1 | 120 | 39 | 39.0805% |
49 | 2 unit(s) on 35:1 | 119 | 47 | 39.8952% |
50 | 2 unit(s) on 35:1 | 120 | 48 | 40.7269% |
51 | 4 unit(s) on 17:1 | 119 | 47 | 41.5376% |
52 | 4 unit(s) on 17:1 | 120 | 48 | 42.3734% |
53 | 1 unit on 35:1 | 88 | 52 | 43.1811% |
54 | 6 unit(s) on 11:1 | 120 | 48 | 44.0199% |
55 | 13 unit(s) on 5:1 | 120 | 42 | 44.8371% |
56 | 8 unit(s) on 8:1 | 120 | 48 | 45.6664% |
57 | 1 unit on 35:1 | 92 | 56 | 46.4756% |
58 | 1 unit on 35:1 | 93 | 57 | 47.2857% |
59 | 3 unit(s) on 17:1 | 110 | 56 | 48.1169% |
60 | 12 unit(s) on 5:1 | 120 | 48 | 48.9593% |
61 | 1 unit on 35:1 | 96 | 60 | 49.7712% |
62 | 7 unit(s) on 8:1 | 118 | 55 | 50.5946% |
63 | 7 unit(s) on 8:1 | 119 | 56 | 51.4345% |
64 | 7 unit(s) on 8:1 | 120 | 57 | 52.2620% |
65 | 11 unit(s) on 5:1 | 120 | 54 | 53.0977% |
66 | 1 unit on 35:1 | 101 | 65 | 53.9128% |
67 | 3 unit(s) on 17:1 | 118 | 64 | 54.7394% |
68 | 3 unit(s) on 17:1 | 119 | 65 | 55.5801% |
69 | 3 unit(s) on 17:1 | 120 | 66 | 56.4040% |
70 | 10 unit(s) on 5:1 | 120 | 60 | 57.2362% |
71 | 6 unit(s) on 8:1 | 119 | 65 | 58.0624% |
72 | 6 unit(s) on 8:1 | 120 | 66 | 58.8952% |
73 | 1 unit on 35:1 | 108 | 72 | 59.7159% |
74 | 4 unit(s) on 11:1 | 118 | 70 | 60.5490% |
75 | 9 unit(s) on 5:1 | 120 | 66 | 61.3864% |
76 | 4 unit(s) on 11:1 | 120 | 72 | 62.2280% |
77 | 1 unit on 35:1 | 112 | 76 | 63.0529% |
78 | 1 unit on 35:1 | 113 | 77 | 63.8789% |
79 | 5 unit(s) on 8:1 | 119 | 74 | 64.7081% |
80 | 5 unit(s) on 8:1 | 120 | 75 | 65.5608% |
81 | 1 unit on 35:1 | 116 | 80 | 66.3913% |
82 | 1 unit on 35:1 | 117 | 81 | 67.2234% |
83 | 1 unit on 35:1 | 118 | 82 | 68.0577% |
84 | 1 unit on 35:1 | 119 | 83 | 68.8946% |
85 | 7 unit(s) on 5:1 | 120 | 78 | 69.7363% |
86 | 2 unit(s) on 17:1 | 120 | 84 | 70.5759% |
87 | 3 unit(s) on 11:1 | 120 | 84 | 71.4166% |
88 | 4 unit(s) on 8:1 | 120 | 84 | 72.2573% |
89 | 1 unit on 17:1 | 106 | 88 | 73.0831% |
90 | 6 unit(s) on 5:1 | 120 | 84 | 73.9387% |
91 | 1 unit on 17:1 | 108 | 90 | 74.7669% |
92 | 2 unit(s) on 11:1 | 114 | 90 | 75.6063% |
93 | 3 unit(s) on 8:1 | 117 | 90 | 76.4511% |
94 | 13 unit(s) on 2:1 | 120 | 81 | 77.2914% |
95 | 5 unit(s) on 5:1 | 120 | 90 | 78.1648% |
96 | 1 unit on 17:1 | 113 | 95 | 79.0000% |
97 | 2 unit(s) on 11:1 | 119 | 95 | 79.8559% |
98 | 2 unit(s) on 11:1 | 120 | 96 | 80.7027% |
99 | 4 unit(s) on 5:1 | 119 | 95 | 81.5469% |
100 | 4 unit(s) on 5:1 | 120 | 96 | 82.4054% |
101 | 1 unit on 17:1 | 118 | 100 | 83.2534% |
102 | 1 unit on 17:1 | 119 | 101 | 84.1057% |
103 | 1 unit on 17:1 | 120 | 102 | 84.9649% |
104 | 2 unit(s) on 8:1 | 120 | 102 | 85.8240% |
105 | 3 unit(s) on 5:1 | 120 | 102 | 86.6832% |
106 | 1 unit on 11:1 | 117 | 105 | 87.5342% |
107 | 1 unit on 11:1 | 118 | 106 | 88.3904% |
108 | 6 unit(s) on 2:1 | 120 | 102 | 89.2606% |
109 | 1 unit on 11:1 | 120 | 108 | 90.1314% |
110 | 2 unit(s) on 5:1 | 120 | 108 | 91.0021% |
111 | 1 unit on 8:1 | 119 | 110 | 91.8691% |
112 | 1 unit on 8:1 | 120 | 111 | 92.7481% |
113 | 1 unit on 5:1 | 118 | 112 | 93.6150% |
114 | 3 unit(s) on 2:1 | 120 | 111 | 94.5061% |
115 | 1 unit on 5:1 | 120 | 114 | 95.3970% |
116 | 2 unit(s) on 2:1 | 120 | 114 | 96.2879% |
117 | 3 unit(s) on 1:1 | 120 | 114 | 97.1788% |
118 | 1 unit on 2:1 | 120 | 117 | 98.0938% |
119 | 1 unit on 1:1 | 120 | 118 | 99.0211% |
120 | no bet | - | - | 100.0000% |
From my tests I would say, that it is better to use the table as a positive progression, and start with lower bankroll and make sessions from that, there can be some nice upswings, but also long times of no win, of course. If you start with 105 units, it tanks too often in my opinion, wins are small and cannot make up for losing ones. I have seen several loses killing it and only maybe two or three wins in between. On the other way, several times it survived 10000 spins and in the end returned to positive. Other times the 10000 spins window was just a losing graph. Be warned, it can fluctuate very much. But probability of ONE win is maximised, price is that you either win 15 units or lose all 105, nothing in between.
Thanks
You are probably right for my progression.
My experience is that also positive progressions are a slow way to the agony and death.
It's a matter of time!
Guetting ,Pluscuop,Maxxpro,D'Alencon tested as positive progressions hundreds times.
Small losses very frequent,big wins every now and then that don't offset previous losses.
But I agree negative progression are kamikaze behaviours.
IÃÆ'Ã,¬ll try your table,that is dangerous too,I'm afraid.
I don't Know a more solid progression of a Marty eleven terms(the maximum possible in Montecarlo,1023 units!):Risk of ruin 1/1524 attacks,but exposed to house edge on high terms!!
Headache......
QuoteYes,the problem is that the 995 losing players on straight numbers,lose a huge amount of money,much more than each EC players.
The nearly optimal strategy you mention is what I play in Montecarlo Casino.
My win goal is very low(no greed),15% at the most of my bankroll.
Progression:1,2,4,8,10,12,14,16,18,20:105 units(Marty 4 terms and D'Alembert 6 terms).
Risk of ruin/bankruptcy 1,2% or 1 every 83 attacks
(you are expert in Computers,can you check these % as risk of ruin?Thanks)
Risk of ruin of your progression is computed there:
link:://rouletteforum.cc/math-reference/how-to-compute-probability-of-a-progression-winning-or-losing/ (link:://rouletteforum.cc/math-reference/how-to-compute-probability-of-a-progression-winning-or-losing/)
Enjoy :thumbsup:
Hi and many thanks.
It's very kind of you to send me the computed risk of ruin of my progression.
1)I'm a bit surprised by the high probability of losing.
My calculation was close to 1,20%.
Unluckily i'm not as good as you in software and computer ,even I have already download Octave on my MAC.
I have calculated 1,20% of losing all bkr, with a software of a friend of mine, based on Krigman formula.
Is in your simulation the partage option taken into account?
I have also checked with the informations of the following book: Roulette odds and profits by Catalin Barboianu (available on the web) and my result is what I said above.
But I'm afraid that your calculation is much more accurate,so I tend to trust your figures.
In both cases,my or your risk of ruin,is a progression that I must give up asap.
2)I have given a look to your OSBD 128MM table.
Very interesting,useful,but dangerous.
All my conclusions about roulette,that I have studied for 25 years are a bit disappointing.It's normal.
If you play even chance with partage option,you must play the lowest possible number of spin.
The best way of betting,from a mathematical point of view, is to bet the maximum allowed on a straight number.But you need a lot of money and the risk of beeing ruined is still there.
No roulette players enter a Casino and like the two examples above.
That's why I appreciate your yesterday table even if they have the goal to last more and to lose less,but not to win,it's impossible IN THE LONG TER:
Because the bet selection doesn't exist,now I'm am studying a way of playng that of course will lose but the risk of ruin is very far away,maybe behind my life of player or may tomorrow,but very very rare.
Of course the Murphy law is still there.
If you are interested I'll give you a snapshot of my studies.
Sad truth about roulette is, that with a math help you can get as near coin flipping as possible. You are coin flipping for your bankroll. If your target is to double your bankroll, then best probability you can get is a very near 1/2*36/37, but never 1/2 or higher. Your chance is always somewhat smaller than fair coin. With bad systems, your chance is singificantly lower. With good systems, your chance of winning should be very near (36/37)*(bankroll/(bankroll+target)).
As for bet selection, I am not absolutely sure nothing can be done, but it is most probably true. There are ways how to control chaos a little, but roulette is random and not chaotic. But random system should sometimes act as chaotic system because the chance is there. Then some advanced method could be used there.
I am interested in this idea - if you use singles vs. series on even chances, then if you observe them, then you moved to another dimension, because while probability of seeing 10 series in a row is same as probability of seeing ten reds, you will play more spins on average before this happens, because series last always longer than one spin. That's why I will look into bet selections even if there is no reason for them to work. There must be some way to do something...
1)Even if you succeed in reaching exactly 1/2 probability,you will lose in any case,because the money available of a Casino is ,for sure, higher than your bankroll.
I'm sure you know the formula that confirms that!
2)If sometimes roulette behaves cahotic and not random,because there is the chance,is the confirmation that to win is a matter of chance only!
3)Here in Europe all serious players have read Marigny de Grilleau book,600 pages,500$ on the web if available, the biggest book ever written on roulette.
The topic about singles vs series on even chances has been explored in all the ways in this book and the conclusion of Marigny is to wait for a "difference" of 3 between singles and series,difference calculated like that:4/5*square root of total number of spins.
When you met this situation,play FLAT BET at the most 5 spins for reducing this "difference"(ecart in French).
It seems that it doesn't work,computer tested.
The behaviour of singles and series is the same of Red and Black!
4)I'm testing with quite good result a bet selection based on "Law of periodicity" of streets and lines.
Please look at :.win-maxx.com (link:://:.win-maxx.com) for more informations
Mr ore
Very simple question:I play Red. European roulette,en prison/option.
My probability of losing is 51,35% or 50,675%?
Thanks
50,675%
Thank you very much.
I have appreciated all your previous post that I've read yesterday
The best,in my opinion, is the bold policy( 40 units ,wingoal 55 units),that I used today in a real casino with success.
I'm wondering if with the en prison option,your table must include some bets on even chances or the difference is negligeble?
3)Here in Europe all serious players have read Marigny de Grilleau book,600 pages,500$ on the web if available, the biggest book ever written on roulette.
The topic about singles vs series on even chances has been explored in all the ways in this book and the conclusion of Marigny is to wait for a "difference" of 3 between singles and series,difference calculated like that:4/5*square root of total number of spins.
When you met this situation,play FLAT BET at the most 5 spins for reducing this "difference"(ecart in French).
THIS IS WRONG. BAD INTERPRETATION. 1st is not the total number of spins and maybe that's why the computer tests failed... if they did...
Quote from: mr.ore on Mar 28, 05:50 AM 2011
Yes, I know what a bold policy is. Those are simple bold policy. Simple because there is always bet only on one location.
Policy to win 11 units:
bankroll | bet | W | L | Pwin |
0 | no bet | - | - | 0.0000% |
1 | 1 unit on 8:1 | 9 | 0 | 8.7203% |
2 | 1 unit on 8:1 | 10 | 1 | 17.5148% |
3 | 1 unit on 8:1 | 11 | 2 | 26.4321% |
4 | 1 unit on 5:1 | 9 | 3 | 35.2262% |
5 | 3 unit(s) on 2:1 | 11 | 2 | 44.2668% |
6 | 1 unit on 5:1 | 11 | 5 | 53.3046% |
7 | 2 unit(s) on 2:1 | 11 | 5 | 62.3424% |
8 | 3 unit(s) on 1:1 | 11 | 5 | 71.3802% |
9 | 1 unit on 2:1 | 11 | 8 | 80.6623% |
10 | 1 unit on 1:1 | 11 | 9 | 90.0698% |
11 | no bet | - | - | 100.0000% |
For comparison, even chance only version - anti+Martingale:
bankroll | bet | W | L | Pwin |
0 | no bet | - | - | 0.0000% |
1 | 1 unit on 1:1 | 2 | 0 | 8.1882% |
2 | 2 unit(s) on 1:1 | 4 | 0 | 16.8313% |
3 | 3 unit(s) on 1:1 | 6 | 0 | 25.7125% |
4 | 4 unit(s) on 1:1 | 8 | 0 | 34.5976% |
5 | 5 unit(s) on 1:1 | 10 | 0 | 43.7549% |
6 | 5 unit(s) on 1:1 | 11 | 1 | 52.8534% |
7 | 4 unit(s) on 1:1 | 11 | 3 | 61.8523% |
8 | 3 unit(s) on 1:1 | 11 | 5 | 71.1174% |
9 | 2 unit(s) on 1:1 | 11 | 7 | 80.4107% |
10 | 1 unit on 1:1 | 11 | 9 | 89.9406% |
11 | no bet | - | - | 100.0000% |
You start on line with same number as your starting bankroll and according to W,L columns you continue. If you win, go on line in W column, if you lose, go to line with L column. At any point, you know your probability of winning. In this case, start on line 10 to win 11 units.
Number of spins is on average low enough, so house dos not eat so much, that's what I think. While my policy improvement algorithm is not always that good and needs some tuning, computed probabilities are correct, it is just a matrix multiplied many times. So if algorithm finds a policy, it might not be absolutely optimal, but if I see that computed probabilities are higher than Marty, then I know it is very near.
Also remember that those are just fancy versions of parachute...
i'm not american/English, so sorry for my idioma, however it's possible to increase the probability of this matrix
one chance can is split in dozzina (12 num) and 3 terzine (9 num)
so the probability of canche is 48,64
12 num + 9 num have 48.86
Quote from: mr.ore on Mar 27, 05:42 PM 2011
I am sure I am not mistaken. It works because on average you bet less than five units to reach your target, so they are not exposed to house edge. It is same idea as instead of betting two dozens, bet 1 unit on ec and if it lose, bet 1 unit on dozen. If you win, you are 1 unit up, if you lose, you are two units down. Probability of winning is 18/37 + 19/37*12/37 = 0.65303140978816654492.
If you bet on two dozens, each of them 1 unit, probability of winning is 24/37 = 0.64864864864864864865.
0.65303140978816654492-0.64864864864864864865=0.00438276113951789627
That is a difference of 0.4%, not much, but it is there.
We have two units, want to have 3 units. If we avoid betting one unit sometimes, probability of winning is increased. You do not bet on dozen, if you win on even chance. If you bet two dozens, you ALWAYS lose one unit while winning two.
Another example:
Let's have two methods to win one unit with five units bankroll, both are based on probability:
method 1: bet 1 unit on five lines, we cover 5*6 numbers.
Probability to win is p(method 1 wins) = 5*6/37=0.810810.
method 2: bet 1 unit on even chance, if it loses bet 1 unit on dozen, if that also loses bet last 3 units on even chance. Anytime it wins, profit will be 1 unit, while still risking 5 units bankroll like with method 1. Probability is different:
p(method 2 wins) = 18/37+19/37*12/37+19/37*25/37*18/37 = 0.821826
method 2 is better than method 1
Difference is 0.821826 - 0.810810 = 0.011016. That is 1.1% difference in probability. There are those little nasty tricks in gambling. Similar trick can be used in craps, for example. You will still lose in the end, and the difference is actually negligible. But you will play more spins this way for same price :)
Sometimes it happens that I will discover something very interesting in this forum, this is just the post.
Mr.Ore after the simulation, shows that the use of 3 step progresion gives a 1.1% advantage over the normal flat placement.
If his calculations are correct, it gives a very strong reduction of HE, using roulette with the la partage principle and HE only 1.35%, we are already close.
Now you need to determine all the methods that even theoretically cause a delicate relaxation of the HE by the bet selection.
Fast, I get two options, the first option that I tested a long time ago is to play EC and after two lost bets waiting for virtual wins, it gives about 0.50% edge in longrun, to confirm that I had 2 simulations after 100k spins and let two confirm it .
Unfortunately, we are playing the first EC bet the second bet is a dozen, so you will have to combine it a little differently.
The second option is to play, trends with dozens, using High / Low
I have long simulations, where I can see on the wheel without zero, that playing the last two dozen has an advantage. You can also combine to create EC bets and play them.
If someone has an idea and in a simulation on the wheel without a zero some strategy for the EC bets showed him better results than the random selection, add it on this topic, or maybe something will come up with an interesting