Why bac could be beatable itlr

[AsymBacGuy]

hi asym, I write better things then just " share a strategy .. "

I know you do

as.

[Whatswhats]

Thanks! If you want to reply to something I've written, go ahead.
If not, no worries obviously.
Anyway I'll be waiting for your next post.

[AsymBacGuy]

I'll be back to write very soon...

as.

[Whatswhats]

Ok asym, thanks!

[AsymBacGuy]

"There are no valuable patterns to exploit at baccarat"

That's the mathematicians mantra denying any possible edge and strictly speaking they're right.
Providing any card distribution to be really random and not simply "undetectable" from a results point of view.
And, frankly, there are more and more serious players who have found out that the word "random" linked to all "baccarat shoes" sounds at least as an hasty statement.

Baccarat results (then patterns) are the by product of randomness and bac rules, more specifically are a condensation of the probability that any hand starts with a given two-card point as opposed to another two-card point linked with the QUALITY of the two cards nature.
Then there's the impact of the third card(s) where in a tiny percentage of hands will be decided by bac rules mathematically favoring B side.

All those features move around more likely values (ranges) and it's now that we might insert the "more detectable" concept.

Obviously to reach this conclusion we had to compare thousands and thousands of shoes shuffled in different ways at different casinos and, alas, not every shoe production belongs to the same homogeneous category that for theoretical purposes we had to consider as "random".

There are many very interesting studies made by Prof. Persi Diaconis about the card shuffling randomness, I'd suggest to read them.

Anyway and knowing that a so called "perfect randomness" can only come out by atmospheric noise variations and radioactive decay processes, we preferred to take as a "random" control what the random.org site displays after having instructed it to shuffle 416 cards (8 decks), then arranging the distribution within bac results (and patterns).

At those random.org distributions you'll see that your standing 7 point won't be easily beat five or six times in a row by an opposite drawing hand (W/L odds are 6.5:1, 6.5:1, 6.5:1, etc) perhaps by the same "miracle" cards drawn. Things that regularly happen at actual live distributions.

as.

[AsymBacGuy]

Whereas it's possible to arrange the cards in order to reduce (or increase..adv effect) the probability of certain side bets, thinking that casinos do this voluntarily is ridiculous.
Even worse are the complaints of some players claiming that current shoes "do not contain consistent patterns" or "enough Bankers".

The problem is about math probabilities of resolving hands that in turn will affect the "average" actual patterns compared with a control model (random.org).

After all, the more a distribution is biased the better will be our predictions, providing to have collected a fair number of data.
Thus in some sense we should get rid of the "average shoe" definition belonging to a pure random model and so priviliging a strategy toward quite deviated (albeit not univocal) situations considered as a bit "abnormal".

An example is when "too many shoes" won't get streaks longer than three or four or, conversely, when many (too many) long streaks happen in the same shoe (doubles absence).
Another important feature to look for is when the B/P number is too balanced for many steps, meaning that the B/P ratio move around very small or small steps.

Then there are more intricate (and important) features regarding the asymmetry and symmetry getting sd values quite different than random models.

Obviously 'less is more', so betting very few hands and considering "complex" patterns having a relative diluted general probability to show up will amplify a lot our correct prediction.
When in doubt we're not compelled to bet anything as the game is set up to make us losers per every bet placed.

Probably a strategy working at pure random models and getting low deviations must be changed into a simple "ON/OFF" strategy featuring more univocal W or L sequences.

More later

as.

[AsymBacGuy]

Along the course of any shoe dealt, more deeply univocal patterns had shown up and less likely will be the probability to get a proper number of clustered counterbalancing situations, especially whether we take as triggers the complex patterns.
That's why is so important, in our opinion, to consider clustered "unwanted" patterns to flow in order to avoid the "clustered-clustered" event(s) intertwined by isolated searched situations.

Suppose we have at our disposal an infinite bankroll and progressively betting (even by a careful multilayered scheme) to get a side to win after a given number of losses whatever assessed.
We're not going anywhere as sooner or later a B/P or whatever random walk utilized will get an insormountable long losing streak capable to wipe out our bankroll (as there are betting limits).
Things change when we take care of the asymmetrical/symmetrical feature, maybe enlarged by how many A or S clustered/isolated events had come out.

Say we want to bet toward a symmetrical pattern to be isolated (that is not followed by another S pattern).
Even by adopting steadily this strategy (that is by hoping that any S event will show up as isolated) and assuming we are flat betting, we're not going to lose a lot of money.
Let's take a further step, that is waiting that a couple of S events came out in a row, so wagering toward S-S clustered events being followed by an asymmetrical pattern.
Odds are that we'll lose even less money itlr.
Symmetrical situations superior than that are not considered as they're very unlikely to happen and, more importantly, are not giving us an edge.

In a word, the theory applied to a baccarat random distribution (and verified beyond any doubt) tells us that S events must come out as isolated by low-moderate levels of deviation and the same about S-S events (so followed by an A pattern).
It's not that A>S at all costs, just that the random distribution will make very low or low deviation values about the S isolated or S-S isolated patterns.
Now a progressive multilayered plan will get the best of the model even if it cannot be riskless.

Yet we know that a possible EV+ plan must be the by product of a W/L ratio shifted towards the left and capable to get at least the famous winning probability of 51.3% at B bets and 50.1% at P bets.
Now we want to be more precise in our betting, so it's not sufficient to get any S event to start the betting.

So we're evaluating things by a further step, that is how many isolated S events (ASA) will be followed by another S isolated events(ASA...SA) or when they are followed by a S cluster (ASA...SS..); the same about a superior class of S events class that is ASSA...SA vs ASSA...SSSA...).

It's now that the ACTUAL distribution counts, as it's demonstrated that SS events coming out within the initial-intermediate portions of the shoe will make more probable a further S cluster and viceversa.
So it's slight more probable to get S isolated clusters when no clustered S happened so far and whenever S clusters happened the isolated S probability is slight diminished.

Next, again, I'll discuss the importance of properly assessing the doubles appearance consecutiveness, a kind of no brainer tool to get the best of the game.

as.

[AsymBacGuy]

Needless to say our bankrolls are nothing against the enormous casinos capitals, so we either play a "quit when you are ahead" losers long term game or we have to understand that things are arranged by detectable forms only whether considered within a large picture.
Professional black jack players or pro poker players know very well the variance's impact; the advantage of playing baccarat is that we can bet whenever we wish.

More later

as. 

[AsymBacGuy]

Most of the times patterns move around low or very low levels of deviation, more appropriately by ranges of univocal situations happening for short frames.
"Trend following" players (probably 99% of all bac players) will win by hoping (or confiding) that homogeneous (or at least controllable) lines will be superior than heterogeneous lines, so wagering that winning ranges could take a longer than expected distribution.

The remaining 1% part will adopt a kind of system/approach/method/strategy based on B>P propensity, B/P gaps, B/P pattern deviations, RTM effects or other features. Obviously their enemy is strong negative deviations (ranges) not properly balanced by the searched (expected) results.

Definitely both different approaches are completely worthless itlr as bac tables continue to collect huge profits either from trend followers and from "system" players.

Sayed that, baccarat is a game where searched "ranges" of either fashion shouldn't be "chased" while not happening for long along the actual shoe, no matter how strong are the statistical/math laws or data suggesting that.
Such a statement shouldn't be considered as an endorsement to adopt a trend following strategy, simply that when in doubt (say 70-90% of the times or so) we must stay put so no wagering a dime.
For the same reasons, "system" players must know that any kind of advantage should come out by hoping that more likely ranges could be easily silent for quite long, in a word that we can't toss good money after bad.

A possible edge, in our opinion and according to our data, come out when we'll set up the delicate process of comparing more likely ranges with the actual production.
The fact that at baccarat "overalternating" sequences are the least distributions to face works as a sort of decisive "bonus" in order to get and edge.

Do not think that BP "chopping" long successions are the prototype of an overalternating mood coming out "so often".
Run the common 4 derived roads (bead plate, byb, sr and cr) and you'll see that a chopping mood is not so frequent and presenting so consistently.

Moreover, the A/S feature of patterns (at least in the way I suggested in this thread) will be invariably shifted toward getting low negative deviations of S isolated patterns being well balanced with the S clusters counterpart.

If we use a 0.75% probability and the A/S registration, a perfect overalternating mood will be a succession as AAASAAASAAAS..., a thing way more difficult to happen than a sequence as AAABAAABAAAB...

as.