Thanks KFB, I appreciate a lot your comment.
After all you are one the few "coin flip" scholars (a real expert, I mean) capable to know when things could get us an advantage after the inevitable HE we are destined to face.

An important quote I particularly like is one from Alrelax that more or less sounds as:

"Do not bet what you desire will happen".

Paraphrasing it, that means that long term statistical data surely help us but at baccarat we shouldn't forget that we have to deal with many "short term" distributions easily deviating from the "norm".
Thus in those natural situations we have to set up a kind of perfect compliance over the actual results (AR) in relationship of the most expected results (ER).

Sometimes AR seem to be so deviated that any devised ER plan need a lot of hands (or shoes) to get its full power.
On the other end, the ER strenght acts constant and almost always toward the house.   

More later

as.

"Do not wager for what you desire.  Wager with what is happening"

Because, what you desire probably the highest majority of the time will not occur on your wagered hand(s).  It very well might occur while you are standing, seated, waiting, etc., and not wagering on anything, but that is how most all will get sucked in. 

Short term bursts of 'whatever is being presented' is by far the most lucrative wagers to place, IMO.

The entire point of playing is to win money.  Not to prove therory, statistics or pattern/trend triggering or anything of the like. 

Thanks Al, here is the correct quote:

"Do not wager for what you desire.  Wager with what is happening"

Of course that line of reasoning is the basic approach every bac player will adopt and without the proper experience we know the results...

So a more careful strategy would be to put within ranges what is more likely to happen with what actually happens.

In our opinion cutting off many hands from the betting opportunities is the first key point to transform a sure loser into a possible advantaged player.

The second important tool is to control the betting amount: We'll never know precisely (say in the vast majority of the times, at least) when our probability to win is valuably shifted toward our favor, even after having assessed that a same situation running infinite times will get us an edge.

Third, constantly assessing the average sd values even by a general point of view (expected results) and by what is showing up at the shoe we're playing at (actual results).

Fourth, when we have reasons to think that the production isn't so random (basically the vast majority of the actual scenarios), dilute at most our betting unless a univocal line seems to weirdly predominate what we're more likely to expect on average.

Now let's consider those points from the casino's point of view.

1- More hands a player will bet, higher will be our (casino) return as the HE constantly works at our favor.
At any rate, we (casinos) can safely rule out that a selected strategy will get a player any edge: itlr every player will lose, maybe less but keep losing.

2- Variations of standard betting amount, imo, should be limited within percentages of it (say 10% or 20%).
At a general negative edge game, it's always better to consider losing situations than winning situations.
Doubling or even tripling up (or more) a standard bet needs a very long experience and of course strong data to support such move.
We think that it's way better to win less and lose less in a particular spot than losing a lot or winning a lot in the same spot.
Of course there are some (very rare) players capable to get the best of such progressive betting, mainly because they approximate at best the spots to get a kind of advantage.

Anyway any player raising his bet is loved by casinos.

3- Sd (standard deviation) values are the watchdog of randomness but most part of the actual shoes we have to face aren't random by any means.
I've already stated that we have strong scientifical reasons to think that RNG instructed productions are not random and I've provided some examples.
Therefore, consider a shuffled machine shoe as a strong "actual" production weirdly shifted toward one side or another of the AS or S world by strong clumps or deviations.
Basically and paradoxically such productions will elicit an abnormal number of S patterns intertwined by few AS situations (despite of the 0.75 AS probability), but good news is that some shoes will produce interminable sequences of AS streaks up to the point that the entire shoe is asymmetrically shaped (a thing that we haven't encounterd once at real shuffled shoes).

Seen from the casino part, S patterns could get the players a distorting perspective that symmetry tends to overcome the rest up to the point that some shoes will totally deny it by producing long asymmetrical successions that are not so easy to detect (unless one had read my pages).     

4- Shuffle machines dislike the "average" word (finite and dependent random models do like averages) and when "average" doesn't act at the shoes we're playing at it's better not to risk any money or to ride the situations when the S patterns seem to be "too much" silent.
In any instance I'd suggest to bet toward symmetry, let such symmetrical weird long patterns go without risking a dime from our part, instead waiting for the relatively rare shoes where AS strongly overcome the S counterpart.

as.

So it's the rhythm of distribution that counts, you'll climb the ladder by one step each time with the constant risk of falling down two steps (or three steps if you adopt a 1-2 mini progression).

Of course once we are up several steps, we should focus our attention about HOW we got so high and the "how" means how many first or second winning situations we've encountered so far. I mean first or second CLUSTERED situations.

To cut a long story short, most part of our edge and providing a decent bet selection (with the guidelines already traced) come from a careful evaluation about first winning attempts distribution on the shoe we're playing at.
Therefore when there were few or none, stop to bet toward winning the first step but even about winning the second step as the asymmetry works slightly better on the patterns giving us first wins.

Being this the case, that enable us to set up a difficult but quite profitable "counter strategy" where instead of chasing separately the second W attempt (or the first W attempt, but it's more risky), we might bet not to get the general asymmetry we're looking for. Just for one winning step at both cases, of course. 

This move is particularly worthwhile when shoes are RNG distributed.

Anyway baccarat results, whatever distributed, tend to take a "clustered" line than an overalternating line.

It's up to us to approximate at best how long a cluster lasts or when a cluster of any kind is more likely to show up after a sort of overalternating succession happened.

More later

as.

Suppose the core of a shoe (it's a real shoe) went as:

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

A=14, S=5

Assuming A=+1 and S=-3, we didn't find any A favourable gap (14/15).

Nonetheless such distribution will get some precious hints to make "educated guesses".
And in reality such shoe is a perfect shoe to deal with, providing to take care of the quality factor.

For example A clusters= 5, A isolated= 0.

Then S isolated= 4 and S clusters= 0.

So a "perfect" exploitation of this shoe would be to bet toward A clusters and S isolated, meaning will get 9  wins and 0 losses (before vig).

Notice that we have wagered 9 times out of 73 resolved hands.

Anyway the scope of playing baccarat is not to prove complicated distribution features but to win money, so we could easily secured a profit well before the end of the shoe.

Accepting a decent profit without chasing more wins must be correlated with a specular action made at those (rarer) shoes presenting many opposite situations, that is A isolated events and S clusters.

In that the former A and S shapes will help us to define our future course of action as we've seen that most part of symmetrical events are coincidental (and/or coming out from an unrandom shuffle).

Practically, whenever we see that the S/A ratio is too high than expected we ought to think that the production is affected by too math volatility and/or unrandom features.
That means that here we have to be very careful to chase the asymmetry for a couple of reasons:

1- math advantaged situations that went "wrong" in spots we consider as "more likely" are consumed without getting us an edge (actually producing a damage);

2- unrandom shufflings markedly polarize the A or S sides by values being quite distant from a real random production.
In reality it's true that generally speaking S clusters will make more room to subsequent A clusters, yet we prefer to place bets when S patterns come out by distributions fitting REAL random requisites and not from RNG softwares that totally negate a perfect randomness by definition.

Those considerations helped to build the concept of "playable and unplayable shoes" with many categories in between (mainly playable, moderately playable, mildly playable).

There's no such a thing a player can get the best of it by wagering ALL shoes dealt, even by betting few hands per shoe.
The real purpose of playing baccarat is to be ahead of something after EVERY shoe PLAYED.
And since not all shoes are "beatable" (that is offering us a kind of advantage), we'd better prepare to stay put and wait (or ask) for another shoe.

Betting good money after "bad" (that is unexpected and/or weird events) tend to increase house profits so reducing at 0 our possibilities to win itlr.

That's why is so important to join tables having some players already seated or to jump at other tables open.
No action and the casino in question do not like to deal free hands? Good, do not bet a cent.

Consider the actual bac A/S ratio shifted towards the right in the same way black jack counters take a negative count.
With the important difference that at baccarat you are not forced to bet anything and changing tables is way easier. (Not speaking about the "no mid shoe entry" policy).

as.
 

BTW, when you are losing do not think you made wrong moves or got unlucky.

Luck doesn't intervene at baccarat but for a 1% impact and wrong moves are restricted into this list:

a- Betting too many hands;

b- Keep staying (and betting) at an unfavourable shoe;

c- Raising the wagers while in the negative territory and mostly negative situations;

d- Start betting side bets in the hope to quickly recover the deficit.

My comments

a- It's the biggest mistake every bac player makes.
Profitable opportunities come out quite rarely: long positive sequences are rare to happen, but long negative sequences can't end up (lol).
Obviously and besides the joke I'm referring to the constant HE to be overcome.

b- Once we didn't get a profit after the mid part of the shoe, it's better to consider it as unplayable.

c- If your plan itlr doesn't win by flat betting it won't by a kind of progressive wagering. If you'd think that some progressive betting spots are more likely to win, so capable to invert the HE, well bet only those spots.

d- Side bets wagering is a delicate issue, curiously the only field to get a math edge at baccarat despite of their huge negative edge.
But if you're not card counting or getting a very huge experience in dealing them, catastrophic results are around the corner.
Anyway it would be silly to break even (EZ bac tables) or getting half of the winning bet (no commission Tiger tables) when a key bet is placed at B side.

as.

Later I'll present some ideas about the importance of getting more convergent parameters than we can to orient our betting (or no betting) decision.

as.

Thanks Asym

I'll present some ideas about the importance of getting more convergent parameters than we can to orient our betting (or no betting) decision.

I look forward to the next post.

Thanks KFB for your interest!

Every math/stats/gambling expert will teach us that no matter how we dissect baccarat outcomes every hand will be dealt by a probability close to 50/50.
That's ok, but at baccarat probabilites are dynamically supposed to deal most hands by hugely shifted situations, think about the first card being an 8 or a 9 (or a 7) or the probability that an asymmetrical hand will take place (B>P) and so on.
Of course we can't have the means to perfectly spot the situations where one side is more likely to show up, so we have just to approximate at best our betting options.
In our opinion (and according to our results), some parameters tend to help us in defining such a matter.

General parameters

a) Any tie happening at a shoe acts as a sort of "resetting" hand, meaning that what was supposed to be more likely all of a sudden is neutralized by this tie.
More ties are showing up at the shoe we're playing at and less precise will be our wagering plan, unless the previous pattern was so robust to keep going that we can "risk" the same bet for the "after tie" next hand.
Therefore most of the times a tie means that we'll have to wait the next hand result before betting again.

Notice that RNG productions are slightly more likely to form rich ties shoes, obviously in any way capable to erase and invert the strong HE.

b) The number of naturals

We know that naturals account for slightly more than 1/3 of the total outcomes (ties included).
Whenever the naturals/every other hand ratio seems to be too strongly shifted toward the left (too many naturals) or the right (too few naturals), we have reasons to think that in the vast majority of the times the production doesn't fit the random requisites as natural variance considered at real random models is limited within more probable deviated ranges per each shoe dealt.

c) The number of 6 cards utilized to form a final hand

When many hands are resolved by 6 cards we are playing a "gambling game" where theorically P side gets more reasons to be bet than average.
Obviously 6-card hands remains the prototype of gambling no matter which side we'll wager, think that 6-card hands coming out consecutively are not so probable at real random models, yet RNG productions (for example) will form several back to back 6-card hand patterns.

d) Third card impact (5-card hands)

Standing points vs drawing points are hugely favorite to win or tie (4.33:1 or 6.5:1), curiously some productions are "too much" privileging the right side, meaning that miracle cards happen with too much frequency than expected for many shoes.
Naturally and to get long term expected values, some shoes are not going to give any help once for the drawing hands, all in complete "random" fashion that in reality is a strong form of unrandomness.

Specific parameters

Even assigning the most power to the natural variance and coincidental variance to bac productions, symmetry remains underdog vs asymmetry yesterday, now and in the future.
We've found a strict relationship between the general parameters and the specific parameters, meaning that more general factors are acting at the actual shoe and stronger is the symmetry strenght over asymmetry.

On the other end, even the most unrandom shoe must provide room to the more likely asymmetrical patterns up to the point that an entire shoe will be asymmetrically placed along its entire course (a thing that happens very very rarely at random shoes).
More importantly is assessing what happens at the vast majority of unrandom shoes dealt: that is when symmetry concedes space to the asymmetry for whatever reasons (and vice versa).
Now we have to restrict the field of operations as the "clumping cards" factor doesn't intervene here, so the problem is more focused about the symmetrical different patterns succession.

In some way unrandom sequences are less detectable as a whole but way more detectable whenever some previous patterns happened at the shoe we're playing at. Thus we'll get a lesser number of profitable situations but involving a larger edge.

See you in a couple of days.

as.
 

You wrote ......"happen with too much frequency than expected for many shoes."

Yes, understood.  But that is why a player can not rely on bac stats for mechanical wager placements. 

Those stats, as I said before are usually derived from 1,000,000 shoes/80,000,000 hands.  Those stats have all kinds of deviations within those results.  Long and short presentments of everything imaginable. 

Random fashion, coincidental and Unrandomness are all part of the game.  One or two or all three, throughout most shoes.

The bottom line is, you cannot condense the stats down to 3 or 4 or 7 shoes that comprise your session. Which a lot of people think you can but unfortunately you cannot. 

You wrote ......"happen with too much frequency than expected for many shoes."

Yes, understood.  But that is why a player can not rely on bac stats for mechanical wager placements.


Exactly.
Once a mechanical plan should work (and of course it can't without implementing actual shoe results with expected average results), cards could be shuffled to get inimaginable volatile results we can't do anything about but guessing.

IMO to win at baccarat we need to approximate at best the average pattern formation and lenght and not guessing or hoping for.
Once the normal deviations seem to be surpassed on the "wrong side" we have two different approaches to  follow:

a- trying to follow such deviations (up to a point), that is in a sense playing to "lose"

b- quit betting for that shoe and wait for the next one.

More later

as.

IMO at baccarat it's not about how many "good things" happen, but how to limit the "bad things" impact; only players capable to correctly estimate the bad things impact could think to make a living (or prospering) at this game.

Flat betting is the paramount source to understand whether we're really utilizing an EV+ bet selection or not, most of the times progressive plans only dilute the problem and never resolving it.

We've seen that "quantities" need very large samples to find a kind of propensity, so they are affected by huge variance being positive or negative.

Pros hate variance, pros like numbers, pros like more likely situations.
 
The fact that at baccarat there's no math edge to be exploited doesn't mean nothing.
Relying on a statistical advantage needs a careful "quality factor" assessment as quality patterns are a lot more detectable than quantity patterns (that obviously aren't detectable at all).

Whenever the patterns quality tend to strongly deviate from the more likely distributions, we better stay put or, maybe, to ride such deviation for fewer spots than when more probable successions show up.

The best way to separate money from our pockets is hoping to get the most probable ranges when situations dictate otherwise and even worse is keep wagering a shoe where things seem to go too much "unnatural".

Do not fall into the trap of thinking that every production is randomly distributed and belonging to the same category.
RNG productions aren't random by any means.
So do not play at them or act accordingly, that is by hyper selecting the playable shoes that obviously must be classified by an actual propensity to deal "more probable" or "hugely probable" patterns that happened by coincidence and not by natural variance.

Generally speaking, unrandom shoes will make a slight greater number of S-S patterns than random shoes, meaning that if you'll have the patience to wait for a S-S sequence the next probable outcome will neglect the 1:3 expected ratio so making proportionally more probable an A pattern than another S pattern.
(S-S-A > S-S-S)
And curiously this feature tend to reproduce itself at many derived roads.
It's like that a symmetrical RESULTS distribution propensity springing from an unrandom source tend to privilege the symmetry at most derived roads whatever considered.

Finally we know that not every RNG software acts in the same way; some productions will elicit a larger unusual number of S-S-S patterns than expected by a random distribution, obviously at the detriment of lower S-S classes.

Anyway, when consecutive S patterns seem to show up too often regardless of their lenght, avoid to play that shoe in terms of asymmetry.
Or, wait for their relative appearance (S-S or S-S-S) and start to bet toward A, a wise move to play with an edge.   

as.

It's a sin to let a "very good" shoe go without exploiting it by making several wagers, but it's a real disaster stubbornly trying to "catch" a very good (or just good) shoe around any corner, so totally disregarding the probabilities.

Taking care of quantities, "Bad" not only is specular (notice, not symmetrical) to "good" but getting unexpected deviated situations even at the more controllable quality features.
That's why strict mechanical systems cannot have a single chance to win for long unless the plan is carefully adapted to what we're facing in relationship of what we're supposed to get.
This translates into a kind of "human work" at the shoe we're playing at as we cannot use algorithms to define the issue.

W/L random walk

Suppose we want to utilize a diluted betting plan (few bets wagered per shoe) and then registering our W and L outcomes in a line.
Does this "personal new road" help us to approximate at best our prediction on future W/L patterns?

Impossible task at an independent model and according to gambling experts it's impossible either at baccarat.

In reality we disagree with gambling experts as baccarat provides many nice spots (that is where the raw "actual" corresponds to the "expected range") by levels capable to reach or better surpass the famous 51.3% winning probability at B bets and the 50.1% winning probability at P bets.
Overall that means to play with an advantage.

Suppose our betting plan provided such W/L line on this shoe: (Not always W and L are real wagers, just outcomes based on the AS/S approach):

L-W-W-W-L-W-L-W-L-L-L-W-L-W-W-L-L-L-L-W-W-W-L-L

W=11 and L=13, so W predominance cannot help us in this shoe.

Are there possible REAL betting plans capable to offer us a possible valid bet selection (that is quitting the shoe with a profit?)

More later

as.

Start with a general consideration that any WL sequence will get an average number of shifts (W changes to L and L changes to W).
In our example (where the BP real betting options were "concentrated" into the W/L form) we got 12 shifts within a 24 WL outcomes succession.
Moreover out of 12 shifts, 8 were included within the 1 or 2 step ranges (that is 2/3 of the times a univocal W or L sequence will stop after one or two W or L).

This simple consideration (of course completely worthless at independent models) will help us to define the expected average number of consecutive wins or losses: once such a homogeneous W or L pattern had surpassed the level 2 (two consecutive wins or two consecutive losses) we are relatively not interested to chase the W or L side anymore, so waiting for a new shift to happen. 

It's particularly important to understand that the number of shifts roam around average values but consecutive W or L successions could take strong deviated lines, so denying the aforementioned average shifts number.
That means that sometimes the "expected" will make more room to the "actual", so we've chosen to either let the shoe without betting or, maybe, to ride it by a kind of "reversal" fashion.

Since we have theorized that symmetry (at least under our empirical guidelines) tends to succumb to asymmetry as more often than not is a by product of "coincidental" events, once a symmetrical pattern had come out clustered more than twice (that is producing a LLLL or longer sequence), we'll have reasons to consider such shoe as a kind of "abnormal" shoe. That is not fitting the average requisites, so considering it as unplayable or worth to be chased by an opposite line of thought as long as no wins towards the asymmetry show up.

So in some sense, most part of asymmetrical patterns will show up either heavily clustered or not happening at all (or quite dispersed in limited ranges of apparition), so conceding a greater than average probability that asymmetry will be denied right at the first level of consideration.

See you in a couple of days.

as.

The number of shifts is another way of considering the CFS, but there's another factor to take care of, that is the RFS (rows filling speed) that starts to act after two same sides happen as the row #1 is 100% filled by definition.

More later

as.