+ Register

[Jimske]

This thread quickly turned into a discussion of progressions.  Why?  Most believe they will lose more hands then win I guess. 

Nice replies.

Imo, FB is just the start to set up a possible long term winning method.
Variance will be very high and, imo, a FB method expecially whether not mathematically based (being an impossible task, we know) will cross many unfortunate situations no matter how we'll be patient and disciplined.

Totally agree.

If a FB method will have the best of it, it means that our winning sequences will be proportionally longer than the losing ones; so our W sequences in a row will be higher than our W isolated sequences. The same for 3W in a row will be higher than 2W in a row. And so on.
Same about our losing patterns taken in a reverse thought.

This then is the meat of it, isn't it?  Reducing LIAR through bet selection will prove winner.

It should be a good idea to register such situations.

Monitoring the data as you go is the next step.  What to look for?

To reduce the variance's impact, I'd prefer to bet after a given L sequence and I don't care if I have to wait many hands to bet.

Yes.  Sequence of losses not necessarily only LIAR but L patterns and the abundance within a subset.

as. [/quote]Anything else?

[Jimske]

With LIAR limited to 10 or less your best option would be to play a mini series. Similar to GR8's progression.

A series of 9 or 11.

11111111111   If you lose 6 or more double your unit size to 2 and
22222222222   monitor your draw down, always play for a profit before you reset to 1u bets.
33333333333   Be confident in your bet selection, you can't afford to lose 2u and win 1u
44444444444   You will need a bank roll larger than 20u.
55555555555   You will have to decide what limits you put on this MM.


Flat betting with 52% hit rate or greater doesn't ever Bust. Once you enter into a progression, you will have to set its parameters. Theoretically, you should be able to continue in this progression or one a bit more aggressive and never blow through your bankroll with liar limited to 10 and a win rate above 50%.

I would also monitor your recent strike rate. Past 50 bets or so. I know you recently went +17u flat betting inside of 40 bets. I would expect that not to continue and start the progression with 1/2u bets. Or be prepared for a deep draw down and have the confidence to bet through it to profit.

5, 7, 9, 11  this, IMO, is just another stretched progression.  It is by no means infallible and I think you'll fnd some serious drawdowns as escalation increases.

[Jimske]

How often you expecting to be hit by 9LIAR? 

With what you suggest, you need to be "more right than wrong, as well as requiring early wins" at every level.  I personally would use STAR or E-STAR.  There is  a valid reason why the teacher doesn't use it any-more ;-)

ooops - just saw you said this already.  Quite true.  REQUIRE EARLY WINS!

[HunchBacShrimp]

Hello i don't understand why you would use does strings - does it mean i only have to win back half my bets when i am at secound level or do i have to win with exact 50% or more.
My TRNG march have won over 50.000 placed bets flat betting and i still get serios fluctation and variance that destroy that kind of progression.

For example this session - was i only unlucky or how ofthen do you bust even if you have strike ratio around 53% and above.
And i don't understand what is the Point haveing a string of 9 when you move up next level after 5 or less.

I use this strings...

1 1 1 1 1 1 1
2 2 2 2 2 2
4 4 4 4 4
6 6 6 6

Sputnik,

I'm in agreement with you. Even with a 52% hit rate variance can still take deep plunge against you. What you provided is an excellent example. 15 wins and 33 losses isn't anywhere near a 52% win rate. But it is a very realistic expectation after Jimske's latest session of +17u flat betting inside of 40 bets. Which must be very close to 28 wins and 11 losses.

The purpose of a mini series of 11 bets was an attempt to be very conservative, hopefully capturing his most extreme LIAR of 10 inside of just 1 series. Winning 6 and losing 5 is a profit of 1u. No need to increase unit size here. Losing 6 or more in any series of 11 produces a loss. The progression to 2u bets is an attempt to milk larger profits with a win rate of 52%. Having an expectation of soon winning more than losing. Once you start betting larger units in the mini series, it isn't completely necessary to play each series out to 11 bets. Draw down needs to be monitored and if a WWWWWL betting sequence appears at the beginning of any series it should have performed a recoup with profit. Or perhaps reduced the draw down significantly enough to forgo completing the series out to 11 bets and start a new series at a smaller unit value.

However controlling LIAR isn't exactly controlling variance. So it is a risk as you have demonstrated. Progressions are for those of us with a win rate below 50%. An attempt to deploy a clever MM scheme to take advantage of the W/L string configuration to generate profit that overcomes all the losses.


Asym states that an increase win rate above 50% produces more consecutive win sequences than WLWL sequences. Or even consecutive loss sequences. This could point to the use of a MM with a parlay characteristic such as Mongoose. However, imo, variance is still a problem.

I still think it would be better to adhere more to a flat bet method. And after some more thought on the matter think the following to be an even more conservative approach.

Simply bet 1u until you have accrued 10 losses. For each loss record a 1 on a piece of paper. If you won more than you lost, pocket the profit a start over flat betting one unit for another 10 losses.

If you have lost more than you won, then continue making 1u bets for a total of 10 wins, crossing off those 1's as you do. For each loss record a 2 on a piece of paper. You may have fifteen or even twenty 2's recorded before winning those ten 1u bets. Then proceed to bet 2u until all of those 2's are marked off and record a 3 for each lost bet. Rinse and repeat.

This is a long drawn out MM that takes advantage of a win rate greater than 50% and protects you from wide swings of variance more than any other method can aside from straight up flat betting 1u for life.

[HunchBacShrimp]

This thread quickly turned into a discussion of progressions.  Why?

Fork me, I dunno. I sure plunged headlong right into one though. Completely off topic. My fault. Feel free to delete all that nonsense I wrote.

Apologies for that. It was unintentional.

[Big EZ]

However controlling LIAR isn't exactly controlling variance. So it is a risk as you have demonstrated. Progressions are for those of us with a win rate below 50%. An attempt to deploy a clever MM scheme to take advantage of the W/L string configuration to generate profit that overcomes all the losses.

Controlling your LIAR is exactly what controlling variance means.

[HunchBacShrimp]

Controlling your LIAR is exactly what controlling variance means.

No. It is not. Above Sputnik has a series of bets posted where he has limited LIAR to no more than 7. And he still lost 33 bets out of 48. A win rate of only 31%. Illustrating that Variance is not under control.

It is true that a LIAR limited to the extreme of only 2 could at most produce WLLWLLWLL. Imposing a minimum win rate of 33% but it is still capable of acquiring the virtual limit of standard deviation. The control of LIAR is nothing more than a speed limit on variance. A strong influence, but not exactly control.

I also don't consider a limit of 10 to be much of a limit. Even 7 didn't do Sputnik any good. I can post 112 consecutive B&M decisions (one full shoe and a partial of the following shoe) and the bet selection I ran against them that triggered 67 bets and went 25 wins and 42 losses and LIAR never exceeded 4.

[Big EZ]

Im confused.

Are you looking at win % and not the variance. When you say it doesn't do any good with his LIAR at 7 you are correct, if you are talking about flat betting but not in regards to variance

This is the definition of variance through wikipedia :   

In probability theory and statistics, variance measures how far a set of numbers is spread out. A variance of zero indicates that all the values are identical. Variance is always non-negative: a small variance indicates that the data points tend to be very close to the mean (expected value) and hence to each other, while a high variance indicates that the data points are very spread out around the mean and from each other.


So when your saying LIAR is a "speed limit" to variance I do not see how you can make that point. For example, if you test 12million placed bets at you see you have no more then 4 losses in a row do you think you have your variance under control?

[AsymBacGuy]

Again this is one of the most interesting post.


Personally I don't like to base my betting after a cumulative series of fictional losses as any shoe is a distinct part of the whole baccarat world. In that sense I'm with gr8player and many others thought.

Adopting the same betting plan, any shoe will produce some WL patterns along the way.
For example we might target the WLL sequence as our enemy and with some selections there aren't many consecutive WLL patterns per any single shoe.
That means to wait a fictional W then betting two times hoping not to get two losses.
How many times are we expecting to get the WLL sequence per any shoe?
Imo, the difference between targeting a WLL sequence instead of a given LLL sequence is that a most likely outcome should come out in a couple of spots even in form of clustered W or in isolated single L.  Expecially if my search of W is built up onto a multilayered betting plan (to get that W I'm not looking for just one pattern).

Despite this, variance will pose a big threat on our plan expecially on some very deviated and unusual shoes (very rich P shoes, for example).

Again approaching this line, we'll expect to have more isolated WLL sequences than clustered WLL sequences, so we might wait the WLL occurence to better define our selection.

Unfortunately, some shoes will keep to mantain an unusual distribution from the start to the end and nothing can prevent to be in the unlucky situation to get consecutive unusual shoes.

Anyway, the vast majority of shoes will produce a given number of W/L patterns of some lenght and there exist certain spots where we could guess right more often than not (slightly).

Theorically, a supposedly system capable to get more W than L will distribute more single L than clustered L, more 2 L than >2 L, more 3 L than >3 L and so on.
But for some statistical and obviuos reasons, it's very difficult in the short-intermediate term to get every single ratio shifted toward the left. I mean that we might easily have more clustered L than single L and at the same time getting a large amount of double L than superior (2+) L sequences. And some other mix.



as.















 







   



 

[HunchBacShrimp]

Im confused.

Are you looking at win % and not the variance. When you say it doesn't do any good with his LIAR at 7 you are correct, if you are talking about flat betting but not in regards to variance

This is the definition of variance through wikipedia :   

In probability theory and statistics, variance measures how far a set of numbers is spread out. A variance of zero indicates that all the values are identical. Variance is always non-negative: a small variance indicates that the data points tend to be very close to the mean (expected value) and hence to each other, while a high variance indicates that the data points are very spread out around the mean and from each other.


So when your saying LIAR is a "speed limit" to variance I do not see how you can make that point. For example, if you test 12million placed bets at you see you have no more then 4 losses in a row do you think you have your variance under control?

With respect, you can lose me with too much advanced math and its semantics.

Variance is the distance wins and losses deviate from 50/50. Winning 60 bets and losing 140 bets is a win ratio of 30%. If I've done my math right the std is 5.55. Variance is 40 places from center. It is easily within reach of a limit of 4 LIAR. 4 LIAR demonstrated no visible control over Variance other than the speed of which it got there, 4 losses at a time.

As for 12 million decisions, the virtual std limit and maximum variance is still well within the reach of 4 LIAR.

For variance to exhibit any effects of control it would have to be constrained to no more than X places in front of or behind 0. Or expressed as a win/loss ratio no more than a 40/60 split for example. Limiting LIAR itself does not significantly affect its ability to accumulate large amounts of losses.

If you can add a limit of LIAR to a minimum WIAR, or a minimum win percentage greater than 50 inside x number of bets. Then you would be well on your way to controlling variance.

I hope I made my meaning more clear. I appreciate your feedback.

[Big EZ]

I think you are looking at it wrong or maybe in a different way..

I am not a math guru either but the spread sheet I use says that out of 200 bets if you only win 60 that you have a z-score of -10.03. To my understanding zscore and std are the same thing .

And you are talking about off center,that has to do with what "average" is on std. There is no "average" on variance. You also can not have variance below 0 as you state, that also has to do with std.

Also you keep talking about variance with how many losses you accumulate. I agree with you that limiting LIAR itself does not significantly affect its ability to accumulate large amounts of losses.  That does not matter.  You can have variance at no more then 4 LIAR and you can have a strike rate of 20% and be profitable (not flat betting of course)

ps...
I still accept I can be totally wrong, but this is how I have come to understand it

[HunchBacShrimp]

Big EZ,

Forgive me for taking so long to respond.

Let's skip past my standard deviation remark, I shouldn't have brought it up it is not part of out discussion. My fault.

When I've referred to variance being avg, I chose my words poorly to reflect the mean. To be more unclear, I meant the mean. 

Stating variance as a negative, or less than zero is again me poorly expressing my thoughts. In my mind I have the needle of variance pointing straight up at zero, the mean, the expected avg of wins and losses. Whenever more bets are won than lost, the needle moves to the right of center, positive variance for the bettor. Whenever more bets are lost than won the needle moves to the left of center, negative variance for the bettor. Center being the mean. Variance as you say is not a negative number, it is strictly the distance any given value is from the mean.

I've thought about this from several angles, and nearly came to be in agreement that limiting LIAR controls variance. However, I still adhere to my original statement that "...controlling LIAR doesn't exactly control variance"

And I'm not worming my way out by arguing the meaning of "exactly". But I meant it just as I wrote it. And your response was identical, that "...controlling LIAR was exactly what controlling variance meant"

My perspective may be off, but this is how I'm looking at it.

A limit of 10 LIAR ( which is not in contention ) is hardly any limit at all. Variance can be quite severe, more severe than what nearly any but the most robust MM could handle. Sputnik's example of 48 bets with 15 wins and 33 losses has a mean of 24 and a variance of 9 with a restriction of 7 LIAR. I will bring up the hit rate, because that is what concerns the bettor. A 31% win avg is still very difficult to overcome regardless of the control 7 LIAR had over variance. I know 48 bets is a small sample, and the following of 67 bets is not much better, but a limit of 4 LIAR produced 25 wins and 42 losses in 67 bets having a mean of 33.5 and a variance of 8.5. A 37% win rate. Which is a bit more manageable, though doesn't really exhibit any noticeable control over variance.

We nearly doubled the limit of LIAR from 7 to 4. And variance was only decreased by .5 in 19 additional bets. Producing a win rate only 6% better. This is what I mean, a limit of LIAR doesn't exactly control variance. Variance still has a great deal of freedom of movement. Still has the capability of working itself against the better quite severely.

From a different perspective. To control variance, one would have to say that variance will never be greater than 5. How do I achieve this? by putting a limit on LIAR? Not in my opinion, not to my understanding. At least not with a limit greater than 1. Even the most extreme and unrealistic limit of 2, has no restriction preventing it from WLLWLLWLLWLLWLLWLLWLLWLLWLLWLLWLLWLL. 12 wins and 24 losses with a mean of 18 and a variance of 6.

I'm not sure how to control variance. But I think it has to be confronted on at least 2 sides. Such as a LIAR limit and a WIAR minimum. A change in the expected win percentage would actually change the mean I think, I'm not sure what effect that would have on variance.

Now I could be standing over here in left field, facing the wrong way.