Why bac could be beatable itlr

[AsymBacGuy]

@whatwhats

Basically only a large number of complex approximate algorithms working together will get the best EV+ situations, where some of them ascertain the relative unrandomness (or real randomness) of consecutive shoes and the other part will take care of the "more likely" deviations every shoe is entitled to produce.
Mostly common bac successions we're destined to face are 'biased' in the sense that they seem to get a bit greater  number of univocal deviations than expected, yet the problem remains to understand if such deviations will come out from "natural" fluctuations (sd values) or artificially endorsed by the bias.

Obviously when in doubt betting towards the deviations will be a minor mistake than wagering to have that deviation to stop.
Anyway a steady betting plan directed to get deviations or moderate/strong deviations around any corner is destined to fail unless the asym/sym factor is implemented in the approach.

So any strict mechanical plan (unless suggesting over selected situations) will surely lose because we have no means to know if the shoe is randomly or unrandomly distributed.
I mean that even the 2nd bet could endure long consecutive losing situations, so waiting for a moderate/long fictional 2nd bet losing succession to show up before real betting won't make the job. Actually it should tell us that that shoe is either following a natural deviation or that it wasn't properly shuffled. So no hints.

What you call as "reverse" strategy is an interesting point, providing you'll put in a proper balance what is theorically more likely to happen with what is really happening and that is often best determined by the asym/sym patterns shape and lenght considered by each relative step.

For example, we've tested several thousands of real shoes dealt by a perfect "random" shuffle and we got no one complete asymmetrical pattern succession (that is up to 21 patterns had featured at least one symmetrical pattern per shoe) but in the real world the almost same sample got two shoes without any symmetrical pattern.
Conversely, the longest symmetrical consecutive sequence in our random sample was 6, but in the real world we've accounted a 7 and a 10 long sym succession, supporting the idea that actual real shoes are not properly shuffled.

Conclusion is that nowadays at most (say the entirety) of shoes dealt, the asym/sym feature considered by each step will be less likely to provide specular (so symmetrical) patterns than the opposite situation.

as.

[KungFuBac]

Hi Asym.

"...Obviously when in doubt betting towards the deviations will be a minor mistake than wagering to have that deviation to stop..."

    I think this is the optimum approach for most events. 


Q: Approximately how many events(i.e., Betting Spots) do you consider in most shoes?


Q:What is your typical deviation-from-expectation requirement for betting into that spot? For example do you look for events that lets say occur four times per shoe. Then after say 60% penetration (with -0- occurrence) in the shoe you start wagering for that event to occur  after the first stages of said event have shown?
    OR 
Are you more likely to only wager on events that lets say only occur every 3.5 shoes?


Thx in advance.

[AsymBacGuy]

Hi KFB!

Q: Approximately how many events(i.e., Betting Spots) do you consider in most shoes?

This depends a lot about the actual texture of the shoe, sometimes we need a lot of hands before approximating at best the prediction.
So if the shoe is getting too many weird situations (mainly from an 'hand results' point of view) we prefer to stay put or wagering very few spots.
We think that it's slight more likely to cross a WW situation by diluting the betting than getting the same WW by a consecutive betting approach.
More or less the same about a LL sequence,  anyway those considerations are strongly linked to our specific approach.
Recently we have implemented a kind of additional (very diluted) strategy based just on this: so betting the very next hand toward a L after a single W and betting the very next hand toward a L after a single L.

Once WW and LL patterns had formed we take care of the actual and expected deviations basically by running two different lines:

1) W and L patterns (so "events") seem to get a 1-2 distribution (1 or 2 gaps);

2) W and L patterns seem to provide 3/3+ streaks and few 1 or 2 gaps.

Notice that I'm talking about W/L sub sequences coming out from a selected plan and not necessarily about B/P hands.
If we implement the asym/sym factor on such sub successions, more often than not we are not going to face 'many' symmetrical situations, meaning that WWW/LLL or WW/LL, etc won't be common findings.

It's now that "expected" values will help us to define whether the 1 or 2 line will be predominant at which level of apparition and the idea that per every shoe dealt a perfect balancement between two opposite situation patterns widely intended is out of order.

Q:What is your typical deviation-from-expectation requirement for betting into that spot? For example do you look for events that lets say occur four times per shoe. Then after say 60% penetration (with -0- occurrence) in the shoe you start wagering for that event to occur  after the first stages of said event have shown?
    OR
Are you more likely to only wager on events that lets say only occur every 3.5 shoes?

I'll answer this later.

as.

[AsymBacGuy]

Regarding your second question, KFB:

It depends.
For example a shoe per shoe registration will make plenty of opportunities to exploit an expectation/actual deviation ratio especially at the very first pattern happening at each shoe as being complete randomly determined.

Suppose we're constantly betting that the very first pattern will be an asymmetrical pattern (so not followed by a same quality second pattern and according to the guidelines decribed in my pages).
Obviously we'll expect a fair amount of AS first patterns or, at least, that S counterparts will be somewhat restricted in their back-to-back appearance. The AS/S pattern ratio (utilizing a 0.75 p) is 3:1 but even though it could be slight lesser than that (2,92:1 or so), itlr such ratio will approach the expected value, especially after having assessed the consecutiveness of the results.

But more importantly and besides the real numbers, it's the quality of such first patterns as single S or double S-S will be easily followed by an AS pattern and of course ranges of AS clusters will be particularly probable.
Obviously this first-pattern distribution translates into a permutation issue more insensitive of a possible symmetrical distribution bias of the entire shoe.

To get a better idea of that, let's try to adopt the reverse strategy, that is wagering toward first S patterns and everyone will see very soon that it's impractical to say the least.

Once we want to bet into an entire shoe, things will change a lot because the boundary between expectation and actual distribution becomes more subtle (yet more profitable with some experience).
I'm sorry but by now I have no time, see you next time.

as.

[AsymBacGuy]

Any symmetrical (S) pattern needs two hands in a row to be equally distributed as the previous pattern.
If cards are really randomly distributed, it's a simple task to ascertain how many clustered S patterns will show up as isolated or (more unlikely) as clustered.

No one serious bac player can miss the profitable spots a real random distribution (random.org, for example) will provide up to the point that a multilayered betting plan will be able to destroy every possible distribution in the world.
This fact counterfeits the idea that every hand is totally independent from the previous ones, thus if our plan is based upon S isolated events, we'll be in very good shape to get more wins than losses, especially if we wait for some fictional losses to happen.

Actually let the house to hope that S clustered events will happen for long but they can't as whatever the cards are arranged a kind of asymmetry will take the lead over the counterpart.

Since just one hand will break a more likely asymmetrical distribution (so producing a less likely S pattern), we need to restrict our field of intervention so waiting for a S pattern to stop independently of its consecutiveness.

Therefore once a S-S pattern shows up at the shoe we're playing at and knowing that more often than not long successions of S isolated events are more probable to come out, we might infer that S clustered patterns will be slight more likely followed by another S cluster. Especially when shoes are unrandomly distributed (machine shufflers, for example).

On the other end, S clusters will slight make more probable A clusters so in the end the only successions we should fear are A-SS...-A-SS... sequences.
And such situations aren't going to come out so often and whenever they'll show up they'll constitute an astounding trigger to get our future bets affected by a huge EV+.

Suppose we have four distinct a-b-c-d fictional players betting for us:

a) player will bet toward A-A just one time;

b) player will bet toward A one time after any single S;

c) player will bet toward A-A after any S clustered event;

d) player will bet to get a A-A-A (or longer) situation.

Our long term data told us that in the vast majority of the times isolated A (so negating an A-A sequence) aren't coming out by a level suprassing the 3-level.
Therefore way more often than not negating a fourth A isolated appearance.

Isolated S patterns are affected by a very low volatility, meaning that isolated S events are more likely to show up clustered than followed by a S cluster.

Once a S clustered event happens (S-S or S-S-S and so on) it'll be slight more likely to face an A cluster.

Clusters of A getting the exact two value (S-A-A-S) aren't going to get many back to back sequences without getting a more natural superior A succession.

Overall we won't face many situations getting ALL four players to lose for long.
Actually it's very likely that at least one or two (or more) players will get the fair amount of positive situations they're entitled to get.
It's just a matter of time and actula deviations, way better to be resolved by a strong diluted bet selection.

as.

[AsymBacGuy]

Selecting same spots patterns at a multiple shoes succession

Suppose we are registering A/S patterns by assigning a progressive number per every shoe played and arranging them into columns.

Here a brief example of 20 shoes:

A-A-A-A-S-A-A-A-A-A-S-S-A-A-A-A-A
A-A-S-A-A-A-A-A-A-A-A-S-S-A-A-S-A-S-S
A-A-A-A-A-A-A-S-A-A-A-A-A-A-S-A
A-A-A-A-S-A-A-A-A-S-A-A-A-A
S-A-A-A-A-A-A-S-S-A-S-A-A-S-A-A
A-S-A-A-A-A-S-S-S-S
S-A-A-A-S-S-A-A-S-S-A-S-A-A-A
A-S-S-A-S-S-A-A-A-S-A
S-S-A-A-A-S-A-S-S-A-A-A-A-S-A-A-A-A
A-A-A-A-S-A-A-A-A-A-S-A-A-A-A-A-A-S-A-S
A-S-A-A-A-A-A-A-S-A-A-A-A-S-A-A
A-S-A-S-A-S-A-A-S-A
A-A-A-A-A-S-A-A-S-A-A-S-S-A-A-A
A-S-A-S-A-A-A-A-A-A-S-S-S
A-A-A-A-A-A-S-A-A-A-A-A-A-A-A-S-A-A-A-A
S-A-A-S-A-A-A-S-A-A-A-A-A-A-S-S-A-A
A-A-A-A-A-A-A-A-A-A-S-A-S-A-A-A-A
A-S-A-S-S-A-A-A-A-A-A-S-A-A-A-S-A-A-A-A-A
A-S-A-A-A-A-S-A-S-S-A-A-A-A-S-A-A-A
S-S-A-S-A-A-A-S-A-S-S-S-A-A-A-A-A-A-S

A= 233 and S=87 (x3=261)

Despite of being voluntarily taken by a kind of S innatural predominance (A:S gap=-28), we see that the above guidelines still stand even by a vertical registration.

For example column #1 (first pattern of every shoe) provides a A-A-A-A-S-A-S-A-S-A-A-A-A-A-A-S-A-A-A-S succession.

Column #2 a less appealing sequence as A-A-A-A-A-S-A-S-S-A-S-S-A-S-A-A-A-S-S-S

Column #3 A-S-A-A-A-A-A-S-A-A-A-A-A-A-A-A-A-A-A

Column #4 A-A-A-A-A-A-A-A-A-A-A-S-A-S-A-S-A

Column #5 S-A-A-S-A-A-S-S-A-S-A-A-A-A-A-A-A-S-A-A

And so on...

This simplified scheme (again...voluntarily taken from a moderate/strong negative standpoint) should get us some hints about approximating at best our action when we want to consider same spots at back-to-back shoes. Especially by assessing that bighornsh.i.t could happen for quite long (see the column #3 providing a cumulative -16 units loss before vig if we'd bet every pattern).
On the other end, A streaks longer than 3 must happen and of course they should be "chased" by selectively wagering and waiting that A patterns reach the 3 consecutive value (AAA).

Finally pretend to embody each column as a distinct player's destiny. There are no many columns getting "easy" A/S positive final returns, whereas more than one column experienced harsh times to endure.

Fortunately things will work way better than this example as in the real world the A/S ratio will be very close to the 3:1 expected ratio.
But being prepared to face negative variance is one of the best recipe to try to get the best of the game.

as.

[AsymBacGuy]

No matter the strategy employed, at baccarat there are two shapes of pattern presentation.

1- Clusters

2- Alternating movement

We've seen that for "cluster" we should consider any same event happening more than once. The simplest (worthless) cluster is B-B or P-P. The problem of those simple clusters is that they are coming out from a consecutive succession (ties ignored).

More interesting, at least theorically, are those patterns considered by a NOT consecutive succession.
For example how many BBB... or PPP... events (3/3+) are coming out in a row when intertwined by a given number of respectively P or B patterns.
Therefore BBB(P...)BBBB is a cluster of two, BBBB(P...)BBB(P...)BBB is a cluster of three and so on.

Since we want to restrict at most the variance impact, we should put a "limit" at those binomial "fights" between a given level of clustering and the superior one. (For example 3/3+ B clusters of two as opposed to 3/3+ B clusters greater than two).

The alternating movement acts in the same way, that is no clusters at level 1 (one alternate pattern then a cluster arrives), no clusters at level 2 (two alternate patterns then a cluster arrives) and so on.

Both different 1 (clusters) or 2 (alternating) pattern shapes will mix up in any shoe dealt by making relatively difficult to get homogeneous situations lasting for long.
On the other end, slight more likely low levels of clustering or alternating shapes in turn constitute a form of clustering and it's here that casinos will get plenty of opportunities to catch players' money as players aim is almost always directed to get univocal lines lasting for long.

More later

as. 

[AsymBacGuy]

Suppose our trigger is the very first B 3/3+ streak happening at every shoe dealt and we want to register what pattern happens at the second B hand dealt. We'll keep registering such 3/3+ streaks (so the third, fourth B pattern, etc) until a B single/double will show up.
For simplicity we name any B 3/3+ streak as a "3" and everything different than that (so any B single or B double) as "X".
Any line is corresponding to any shoe dealt at the same shuffling conditions.

3-3-X

3-X

3-X

3-X

3-3-3-3-X

3-X

3-X

3-X

3-X

3-3-X

3-X

3-X

3-X

3-X

3-3-3-X

3-X

3-X

3-X

3-X

3-X

3-X

3-X

3-3-X

3-X

3-X

3-X

3-3-X

3-3-X

3-X

3-X

3-X

3-3-X

3-X

3-3-X

3-X

3-X

3-X

3-X

3-X

3-3-X

3-X

3-X

3-X

3-X

3-X

3-3-X

3-X

3-3-X

3-X

3-X

3-3-X

3-3-3-X

3-X

3-X

3-X

3-3-X

3-X

3-3-X

3-3-X

3-X

3-X

3-X

3-3-X

3-X

3-3-X

3-3-X

3-X

3-X

3-3-3-3-3-X

3-X

3-X

3-3-X

3-3-X

3-3-X

3-X

3-3-X

3-3-X

3-X

3-X

3-X

3-3-X

3-X

3-X

3-3-3-X

3-X

3-3-X

3-X

3-3-X

3-X

3-X

3-X

3-X

3-X

3-X

3-3-X

3-3-X

3-X

3-X

3-X

3-X

3-X

3-3-X

3-X

3-X

3-X

3-X

3-X

3-3-X

3-X

3-X

3-3-X

3-X

3-X

3-3-X

3-X

3-X

3-X

3-X

3-X

3-X

3-X

3-3-X

3-X

3-3-X

3-X

3-3-X

3-3-X

3-X

3-X

3-X

3-3-3-X

3-3-X

3-X

3-3-3-X

3-3-X

3-X

3-X

3-X

3-X

3-3-3-3-X

3-3-X

3-X

3-3-X

3-X

3-X

3-X

3-3-X

3-3-3-3-X

3-3-3-X

3-3-X

3-X

3-X

3-X

3-X  (1261)

Overall we got:

3-X= 103 times;

3-3-X = 41 times;

3-3-3 (or longer successions) = 10 times.

Since X=+1 and any back to back 3 is -3 (and no vig is acting here as ALL bets are placed at P side) it seems that best bets should be oriented to bet B after any SECOND B hand dealt following a previous B 3/3+ streak (no need to  "chase" another B 3/3+ streak, just any kind of streak---any double---will be good), then after a 3/3+ back to back streak came out, the best bet should be a two-step P bet negating a third (or longer) 3/3+ B streaks succession.

Actually a P double-step wagering after any back to back 3/3+ B streak cannot lose any money itlr, or at least we're dealing with a way better B/P proposition the game mathematically provides in B/P winning probability terms.

Now let's consider the P side under the same shoe conditions.

3-X

3-X

3-X

3-X

3-3-X

3-3-3-X

3-3-X

3-X

3-X

3-X

3-3-X

3-X

3-3-X

3-X

3-X

3-X

3-X

3-X

3-X

3-X

3-X

3-X

3-X

3-X

3-X

3-X

3-X

3-3-X

3-X

3-X

3-3-X

3-X

3-3-X

3-X

3-X

3-X

3-3-X

3-X

3-3-X

3-X

3-3-X

3-X

3-X

3-3-X

3-3-X

3-X

3-X

3-3-X

3-3-X

3-X

3-X

3-X

3-X

3-X

3-X

3-X

3-X

3-3-X

3-3-X

3-X

3-X

3-X

3-X

3-X

3-X

3-3-X

3-X

3-X

3-X

3-X

3-X

3-X

3-X

3-X

3-X

3-X

3-X

3-3-X

3-X

3-X

3-X

3-X

3-X

3-X

3-X

3-3-3-X

3-3-3-X

3-X

3-X

3-3-3-3-3-X

3-X

3-X

3-3-...

3-X

3-3-3-3-X

3-X

3-X

3-X

3-3-X

3-X

3-3-X

3-X

3-X

3-X

3-X

3-X

3-X

3-X

3-X

3-X

3-X

3-X

3-X

3-X

3-X

3-3-X

3-X

3-X

3-X

3-3-X

3-3-X

3-3-X

3-X

3-X

3-X

3-X

3-X

3-3-3-X

3-X

3-X

3-3-3-X

3-X

3-X

3-X

3-X

3-X

3-X

3-X

3-X

3-3-X

3-X

3-3-X

3-3-X

3-X

3-X

3-3-X

3-3-3-X

3-X

3-3-X

3-3-3-X

3-X

3-X

3-3-X

3-3-X

3-X

3-X

3-X

3-3-3-3-X

3-X

3-X

3-3-3-X

Surprisingly now we don't get "shifted" situations proportionally favoring the more likely math advantaged B side for long for the simple reason that asymmetry will reign supreme over the outcomes.

as

[Whatswhats]

Since X=+1 and any back to back 3 is -3 (and no vig is acting here as ALL bets are placed at P side) it seems that best bets should be oriented to bet B after any SECOND B hand dealt following a previous B 3/3+ streak (no need to  "chase" another B 3/3+ streak, just any kind of streak---any double---will be good), then after a 3/3+ back to back streak came out, the best bet should be a two-step P bet negating a third (or longer) 3/3+ B streaks succession.

Dear this is a fallacy, because theoretically and statistically is better the 3 (3) X then bet for the 3 (X)

between () where bet, but if you bet just ONE time after a second hand dealt B (after a 3+ streak) so

BBBB
PP
BB(NOW BET B)

we wil win in your case 41 times but lose 103 times,

because
3 X WILL HAPPEN 103 TIMES
3 3 X WILLC HAPPEN 41 TIMES

BUT WE BET WHEN THE SECOND "3" ISN'T COMPLETE SO IF WE LOSE THE BET EXAMPLE

BBB
PP
BB(WE BET AND GET A P)

WE WILL HAVE A

3 X PRATICALLY BECAUSE

BBB
PP
BB
P

So if you just bet one time after second banker hand it lose more then it win, then yes that we didn't calculate that 3 X can be also

BBB
PP
B
P

so in this case we didn't bet, but anyway you didn't calculate that in this post.

so have you said something wrong or I made some calculations mistake?

[Whatswhats]

So you tell that is better to bet one time after the second banker of

3 3 X

And not 3 X, but is false, explain if I miss something but seems not.

[alrelax]

I'm extremely sorry to interject here, but I have to.  If Asym wants this post deleted, delete it it's okay. 

Reference last few posts on both sides.  I have found in over 40 years of actual brick and mortar bac play, that anything and everything has an equal chance to occur.

Test and stat all one enjoys and until you are confident with numbers, triggers, patterns over the long run.

However, and a huge giant HOWEVER; those triggers that occurred and came about in all one's tests, cannot be transposed into a live bac game of a few shoes and may never come about in that session made up of 1-2-3-4 shoes. 

But, maybe I am off course?

[AsymBacGuy]

@Alrelax: lol, I haven't deleted a single reply in years let's figure out if I'd delete one of yours..

The purpose of presenting this trigger was to give the idea that running the same situation for many times could present some non linear probability values as cards are finite, rules are fixed and a kind of "average" distribution will work providing the outcomes are really randomly produced.

@whatswhats:

Overall we got:

3-X= 103 times;

3-3-X = 41 times;

3-3-3 (or longer successions) = 10 times.

Since X is any pattern different than 3 (so 1 or 2) it's obvious that we have to assess an average probability to get or not to get another consecutive 3 and this needs a two-step betting.
The general odds of any 3 vs (1-2) are 1:3, since itlr B>P we shoudn't be surprised that 3-X will lose money against 3-3-...
In my example and assuming a 3:1 unit W/L ratio (before vig), 3-X won 103 times and lost 51 times (51 x 3= 153).
Clearly by wagering toward another 3 after a 3 happened (now by an inverse positive 1-2 progression) will get the best of it even though long streaks of 3-X can naturally show up along the course of the shoes dealt (that's why Alrelax pointed out his legitimate doubts about these findings).     

Then after a 3-3 succession came out things seem to change as betting toward 3-3-X got a kind of propensity to produce more 3-3-X patterns than 3-3-3-... patterns.
In fact 3-3-X patterns have shown up 41 times and 3-3-3-... patterns just 10 times (10x3=30).

Notice that I've presented the very first (or when applicable) the second or third B 3 streak happening per each shoe dealt, supposing those are more "randomly" placed results than others.

My conclusion is that a bac player should be interested about what happens most now AND itlr or, it's the same concept, that things could distribute by huge levels of volatility but always and invariably following more likely probability values.
And one of the best tools we should employ to get the best of the "actual" related to the "expected" is betting very few hands (for example think how's unlikely to wait for 3-3 to come out)

as.

[Whatswhats]

Online with 30 table, find pattern is faster

[AsymBacGuy]

Online with 30 table, find pattern is faster

Probably it is, but maybe exploiting the derived roads distribution will make a similar job without internet issues, and we know there are plenty of them to overcome.

as.

[AsymBacGuy]

Here additional real shoes coming out from the same source (3s are only considered at B side):

3-X

3-X

3-X

3-X

3-3-3-3-X

3-X

3-X

3-X

3-3-X

3-X

3-X

3-X

3-3-X

3-X

3-X

3-3-X

3-3-X

3-X

3-X

3-3-X

3-X

3-X

3-X

3-3-X

3-X

3-X

3-X

3-X

3-X

3-X

3-X

3-3-X

3-3-X

3-X

3-X

3-X

3-X

3-X

3-3-3-X

3-X

3-3-X

3-X

3-3-X

3-X

3-X

3-3-X

3-X

3-X

3-X

3-X

3-X

3-3-X

3-X

3-X

3-X

3-X

3-X

3-X

3-X

3-X

3-X

3-X

3-X

3-X

3-X

3-3-X

3-X

3-X

3-3-X

3-X

3-X

3-X

3-3-X

3-X

3-3-X

3-X

3-X

3-X

3-3-X

3-3-X

3-3-X

3-X

3-X

3-X

3-X

3-X

3-X

3-X

3-X

3-X

3-3-3-X

3-3-X

3-X

3-X

3-3-3-X

3-3-X

3-X

3-X

3-X

3-X

3-3-3-X

3-3-3-X

3-3-X

3-X

3-X

3-X

3-X

3-X

3-3-X

3-X

3-X

3-X

3-X

3-X

3-X

3-X

3-3-X

3-X

3-X

3-X

3-X

3-X

3-X

3-3-3-X

3-X

3-3-X

3-X

3-X

3-X

3-X

3-X

3-X

3-X

3-3-X

3-X

3-X

3-X

3-X

3-3-3-X

3-3-X

3-X

3-3-3-X

3-3-X

3-X

3-X

3-X

3-3-X

3-3-X

3-3-3-X

3-X

3-X

3-X

3-3-X

3-X

Total:

3-X = 114

3-3-X = 31

3-3-3-..= 10

We see that 3-X vs 3-3-.. sequences are still unfavorite to win (W=114, L= 41x3= 123) and 3-3-X vs 3-3-3.. events are almost equal (+1, 31 vs 30).

Nonetheless notice that 3-3-X sequences do not involve any vig as all bets are placed at P side.
Moreover, the 3-3-X/3-3-3-.. ratio is so balanced that we could even think of adopting a multilayered progressive plan without worrying about the vig.
Yes, the only substantial obstacle will be a permutation issue, so let's pretend to face an hypothetical unbelievable scenario to distribute all losing patterns consecutively or strongly clustered.
But when the "bad" is clustered and we know the proposition had demonstarted to be harshly balanced or shifted at one side, chances we'll cross a win or multiple wins are approaching the certainty.

Next week we'll see how to exploit at a maximum level the AS/S patterns feature.

as.