Quote from: alrelax on July 13, 2022, 01:17:28 AM
As you said, "mainly as many casinos won't allow bets higher than $20.000".

For serious play, it is worth going to, Rivers Casino in Des Plaines, Illinois has between 24-30 baccarat tables and the majority are at $10.00 minimum with a max at $100,000.00.

Good to know, those are very good limits!!!!
Thanks Al!

as. 

Hi KFB!

Thanks for your words and thanks for your inputs!

Yesterday norwegian Espen Jorstad won the WSOP Main Event collecting $10 millions.
(BTW, before canceling it, someone could remember I've made a post by 'guessing' that the eventual Main event WSOP winner or runner-up would be a EU player, I've guessed right and winning some money on it).
The prestigious title from Germany went to Norway.

In an hypothetical game where a given amount of starting money can be bet by an 'all-in' and limitless procedure and by facing a kind of infinite coin flip propositions, a player would transform a $10.000 initial investment into a 1000x $10.000.000 (and more) prize by winning 10 hands in a row (1:1024 odds).

Naturally it would be a utter nonsense to compare a poker tournament having 8663 entrants with this hypothetical game, nevertheless at both scenarios the final winner would be labeled as 'particularly lucky' (and definitely skilled).
Not mentioning that maybe 70% or more percentage of WSOP Main Event entrants have no one single probability to win it. The famous 'dead money' poker pros would rely upon in order to increase their EV.   

What I wish to emphasize is the role of 'time' on both games, as WSOP Main Event winner will be awarded after 8 days of play    (each day lasting about 10 hours or so), whereas at the game prospected a super lucky person could win $10m right at the start or within a very limited amount of time. Actually we know this can't be done as this game doesn't exist in reality, mostly because no one casino in the world would accept wagers superior than maybe $500.000.

Another difference is that at poker tournaments after around 80%/90% players pool has been eliminated, every surviving player will be ITM (in the money); so freerolling for the entire enchilada by getting a small, moderate or huge profit for their initial investment.

Now, let's consider a real game where we could set up a kind of 'poker tournament philosophy', so trying to get the most value of our buy-in (tiny fraction of our bankroll). It's baccarat.

1) We could set up our goal (that is our 'first prize') by hoping that our buy-in will be transformed into 5x, 10x, 15x, 20x or more factor, always considering the possibility to quit the tourney whenever we want.
But we ought to remember that poker players making real money aim to get the final three prizes, so finishing ITM doesn't make the job.

2) It's impossible to win a poker tourney in the early-intermediate stages of it and it's very unlikely to win at baccarat the above best prizes within too short played sessions.

3) Poker tournaments make the 'surviving mood' one of the primary  tools to aim for.. At baccarat this thing is way more important as profitable patterns need some time to show up. But, differently than  poker, they'll come out.

4) At baccarat we don't have to post blinds and ante, we can 'fold' whenever we want without losing a dime, that is not playing a hand until we wish.
Moreover there are no progressive blind limits, forcing us to play or push some hands unnecessarily.

5) At baccarat we could set up many 'back-up' plays impossible to make at poker tournaments (allowed at some cash games).
I mean that after having lost a hand, we could make the same bet maybe enlarged at the very next hand or next trigger we've decided to take as 'good'.

6) To eventually win a lot of money, we must be willing to put a large part or the entire part of our stake at risk at the possible favorite opportunties, maybe by betting a two-step 1/3 then 2/3 wager.
It's a sound math move as a 0.75% probability cannot be silent for long and of course casinos are more worried about an i.di.ot betting huge few times then by facing 'wise' players betting light-moderate sums for long.

7) At baccarat we can't bluff to win, yet baccarat patterns must take a more likely line sooner or later.
It's just a matter of time (time!) that 0.75% probability events will produce univocal steady lines.

Baccarat is the second best game to play in a casino, having the least unfavourable negative edge. Featuring the luxury to bet the side we wish anytime we wish.

Next how to play a given bankroll as a poker tournament buy-in.

as.

If you are selectively betting with a 0.75% general probability, you put casinos to hope to get many clustered 0.25% situations to occur, so you don't necessarily need to get long winning clusters, just to avoid losing unlikely clustered situations.
And the best way to get rid of those situations is to put an alarm bell whenever a losing spot of such kind will come out as sh.it happens clustered.

Greater is the room consumed within a shoe to get those unlikely scenarios and lower will be the subsequent probability to win.
And higher is the hope to get such winning streaks being quite or very long and greater will be the probability to throw into the toilet a previous profit.

With some caveats.

Think to pretend to play baccarat the like a major poker tournament works.
Nobody has ever won a major poker tourney by setting up multiple bluffs (that can't double the stack in any way), actually one had won it by winning many key coin flip situations, along with many underdog spots or wonderful unlikely card distributions.

Make your $5.000 bankroll to face many 0.75 probability spots by betting whatever you wish, knowing that at baccarat some long winning streaks must happen and at the same time not fearing that your QQ or KK will be opposed to a KK or AA hand.
For most poker players, the objective is to finish ITM so after getting, say, a $7500 bankroll you'll be quite happy no matter what. Think that almost 90% of the players will abandon the tables by losing their $5k buy-in.

Of course at poker after having reached the ITM boundary, everything will be a kind of freeroll with a fair shot to get the best prizes whereas at baccarat you can risk to lose the ITM goal.
Yet at baccarat we can quit the game whenever we want not waiting to be eliminated.

On the other end and assuming a first prize of $350.000, a second prize of $210.000 and a third prize of $130.000, a bac player could choose the spots to risk his/her money by distinct fractions, mainly as many casinos won't allow bets higher than $20.000.
I mean no key 'in/out' hands work out.

Think that a poker tournament of such kind will be spread along at least a 32-35 hours of play.

Moreover your $5000 bankroll will play by a 1.06%/1.24% disadvantage but poker tournaments must overcome a well higher negative percentage at the start.

Comments

a) play baccarat only when you hope to transform your bankroll (before finishing it) into at least a 10x or 20x or more accomplishment, providing a proper amount of time employed at the tables.

b) defend your fkng bankroll at all costs, avoiding the most part of st.u.pid coin flip confrontations

c) do not be shy to bet all your money (when allowed) in some selected appropriate situations. It's the same situation poker players will do everytime but unless they hold AA (220:1 odds), such players do not know how favorite they are. Most of the times they are flipping a coin for their tournament life.

d) consider that to multiply your bankroll you have to get the best of it at some math unfavorite situations more than once (for example winning a P wager having a 4 initial point vs a 7 B point and catching as third card a 4)

e) at baccarat no need to employ ICM poker strategies as blinds/bankroll cannot affect your play in any way.

f) be patient the same way poker players are. You can't hope to win with your 2-8 the same way valuable bac patterns are not coming out so frequently, especially after having lost two or more hands in a row.

g) let casinos to lose serious money and not playing by risking a huge bankroll to win a measly amount.

h) finally it's true that at baccarat some long winning situations could happen at the start of the session making a player a huge winner (a far difference with poker tourneys),  yet those are very rare occurences and being prepared to survive a lot of hands is, imo, the viable way to get the best of baccarat.

Next week I'll make some real examples about that.

as.

Hi KFB!
Good thoughts and thanks for your examples made at real situations.

According to the scientifical paper I've provided some posts above, the only way to win at this game is to catch a kind of pattern more likely to occur at the actual shoe we're playing at.

Best but much volatile and 'brainless' approach is to hope that a given streak or series of streaks will prolong by a 'sky's the limit' feature; since at baccarat long streaks are uncommon we should discard this plan from our opportunities, after all 99% of players will mostly rely upon such rare occurence, sometimes not getting the guts to bet until the streak/s will end up.

Then there's the wide panorama of more likely occurences forming the vast majority of situations every bac player should face to cross a possible valuable pattern.

To better clarify such important issue, say we consider patterns as 1-step or 2-step W or L situations, we've seen that raising the probability from 0.5 to greater values will get a sensible impact over our long term results.

It's like we're trying to get the best of it by comparing actual statistical happenings considered in various forms as opposed as the constant math probability working for casinos.

Column key spots and row key spots


Along any shoe dealt there will be times where a kind of predominance must take a role in prolonging or stopping a column or a row. 
Math can't do anything about that, so if we're hoping that some additional rows should come out and after having examined the average lenght of previous streaks, we'll bet that back-to-back spots slowing down the natural row flow are less likely to come out.

At the same time, in the vast majority of the times, too rapid rows apparition will make more probable the 'stopping' streaks mood, always in terms of clustered or short-gapped spots.

In the vast majority of the times, those predominances move around more likely row clustered spots (codes) and same level or almost same level column situations taken at precise 'clustered' parameters.

I mean that considering both the row and the column issues by 'clustered forms' and taking into account probabilities higher than 0.5, maybe waiting for one or two fictional losing events to show up (so no need to employ a progression), the advantage a player will get over the house will be astoundingly high.

See you in a couple of days.

as. 

We're talking the same language, then.

I prefer to set up things by more specific issues, you by a 'sections/turning points' feature but the overall product is the same.

A) sections

Patterns belonging to any class or multiple classes belonging to the same category.
They could be singles, streaks in various shapes of appearance: isolated, clustered, isolated by a 2 factor, clustered by a 2 factor, isolated/clustered by factors >2, predominances, unb plan #1 or #2, etc.

B) turning points

Everything that stops a given section considered under specific parameters.

Itlr A=B. So theorically no points of intervention could be spotted to erase/invert the HE.

Since bac shoes are affected by a huge asymmetry of some kind, we may infer that A and B occurences are more asymmetrically placed along any shoe. Thus in the vast majority of the times, a given 'room' space is going to show up, obviously restricted by very long streaks that tend to consume 'space'.

If we're restricting the field of operations, so assigning the same class to those long streaks, we'll get a better picture of how many sections are going to happen, at the same time neglecting a possible 'unlikely' strong predominance factor that might be a 'section' by itself.

IMO, the problem is that many players want to get too many 'sections' to happen, many times considering B successions (turning points) as A sections.

According to my statistical findings and knowing what the few pro bac players do, A sections must be chased very few times per shoe, and the same is true about B spots.
So such players are not interested about HOW LONG a section will happen but about WHEN a section will show up, so trying to restrict at most the B turning points feature.

Naturally there are no strict SURE guidelines to know when a B succession will shift to A, everything relies upon the statistical probability.
At the same time and considering a higher than 0.5 probability to show up, average values per shoe won't be so easily disappointed for long.

The simplest measurable way to consider A sections and B turning points is to take into account 3+ streaks (any 3 or greater lenght streak).
Now A sections are made of singles and doubles and B turning points are made of 3+s.
Even waiting that a first single or double will show up before trying to get a A section (so getting rid of rare back to back long 3+s consecutive streaks), in the vast majority of the times A sections won't happen clustered after a 3+ streak, so B turning points are going to be somewhat clustered (in relationship of their probability to happen).
Good thing is that in the vast majority of the times, such 1-gapped 3+s spots are rarely surpassing the 2 value. So making a fair room to get more likely A sections.

The same is true about doubles. No doubles successions are A sections and 1-gapped doubles sequences are B turning points.
The difference is that we will expect a higher amount of 1-gapped doubles than 3+s, nonetheless the variance values are quite different.
Meaning that we'll get a well greater probability of success to bet that A sections made of singles and doubles will happen (as the 3+s value is more likely roaming around averages) than to cross through the same doubles counterpart.

As you correctly state, we can't know what will happen at a given shoe.

So let's set up this strategy, having two different fictional players bettng for us.

First player gets his/her enemy at 1-gapped 3+s, so playing toward A sections made of singles and doubles greater than 1.
We know that on average he will get at least one losing spot per shoe.

Second player gets his/her enemy at 1-gapped doubles, so playing toward A sections made of singles and 3+s greater than 1.
We know that on average he will get a whimsical amount of losing spots (going from 0 to 4-5).

The only probability to get BOTH attacks losing for consecutive spots is whenever a precise hands succession is like as:

2-3-2-3... or 3-2-3-2...

In those scenarios we'll lose 4 bets and for that matter such successions are quite rare to happen, meaning that 1-gapped doubles and 3+s paced sequences are not so likely to happen.
Even if they happen, we know that such state is going to transform very soon into more likely A patterns, after all in a 4-pattern succession probability not to get at least one single is 1/16.

Now compare such 0.25 x 0.25 spot probability to lose with the overwhelming probability to get a more comfortable 0.75 x 0.75 probability and make such attacks to run on very long samples.
Maybe by utilizing a multilayered progression scheme either at positive side and/or at negative side.

Of course always taking into account that the more probable step to look for is 1 and clustered 1s, so we can't give a lesser damn about those very unlikely 2-3-2-3 or 3-2-3-2 occurences coming out clustered and surpassing the 1 cutoff point.

as.

Ok, I know what you mean, but at baccarat the 'everything can happen' must be quite different than 'everything can happen' at roulette, for example.
Otherwise we are talking about the 'reading randomness' myth and we know the probability to win at an independent and taxed game is absolutely 0.

It's good to look for 'sky's the limit' situations only as we have ascertained that bac rules and features in certain circumstances (say, sections) will make more probable such things than a 50/50 perfect game. Otherwise we're simply gambling.

Am I wrong?

as.

LOL, Al, so true what you've written! 

From one part there are 'magical system sellers' that instead of betting thousands and thousands not giving a fk about selling few bucks systems or books swear they're right without providing any distant possible reason why they should, on the other end there are stubborn math geniuses that do not know a fk nothing about baccarat, yet considering any bet or series of bets a EV- proposition, no matter what.

Does a person want to get a carefully detailed scientifical approach to beat baccarat itlr?
Good, it's on sale at $5 millions.
Way better to extract some ideas presented in this site, doing some work and getting it for free.

as.

KFB, you are so good in considering things that casinos do not want to get us playing together, let alone by a 3-way team with Al.

3+= any triple streak, so 3, 4 or superior streaks lenght belong to this category.

Mathematicians consider bac outcomes as a endless succession of EV- spots, so assuming that every bet will be EV- no matter what. 
This is a strong fkng unbelievable sh.it as EV+ or EV- spots move around dynamical probabilities coming out from the actual card distribution getting well determined sd limits per shoe.
If such math experts would have studied RVM and M.v. Smoluchowski works, and of course they didn't, they would reach at different conclusions and hopefully they haven't.

Our ideas supported by tests made on very large live shoe samples rely upon strong scientifical findings, so you can be assured that itlr and after some variables had happened, a X event vs a Y event will come out by a larger probability than expected and, more importantly, capable to erase and invert the HE. At the very least capable to get very low sd values where a simple multilayered progression would get the best of it. Itlr.

We are so sure about that that we can present our findings at MIT or whatever the fk College of Math Experts one wish to call.
As we're talking about gambling bighornsh.it, we must put at stake some serious money. So let's see how many 'geniuses' are willing to prove that we're telling rattlesnakesh.it by risking a lot of money to prove they are right by getting math on their side.

The challenge is open to everyone so sure about math, providing:

- the sample must come out by a verified live shoes sample;

- we can choose the spots to wager at;

- we can choose the amount to bet;

-at least five or more bets must made per every shoe dealt (including ties);

-we can bet any side bet anytime we want;

- final results are registered only when at least a 1000 shoe LIVE sample was dealt.
(A factor privilenging math geniuses as the probability to be coincidentally ahead at a EV- game by betting at least 5000 spots is very very close to 0).

No kidding, we are really waiting for a such challenge.
Otherwise just shut the fk up, keeping to try to sell poker books.

as.

Hi KFB and thanks of your inputs, I appreciated them a lot!

Think, just my opinion, that a 'something' clustered more likely situation is going to get a way less variance than an isolated 'something' less likely  situation because 'rare events come out in clusters and then they disappear for long'.

Considering outcomes as singles, doubles and 3+s, itlr the most volatility happens at doubles just as they are the more likely occurrence itlr.
At the same time and taking into account that the vast majority of shoes are poorly shuffled, we'd bet toward NOT getting many 1-gap double spots, especially at the wonderful 6-deck shoes where things tend to be fairly polarized.

After all, itlr we'll get a more consistent 3s number rounding the average value than doubles number.
Moreover and considering baccarat as a game full of relatively short streaks, even shoe sections rich of doubles will make many winning points (anyway there are cutoff values , so in a way or another there's very little room to get our bets losing.

Next week I'll try to summarize such different strategies, anyway the keywords are 'gaps' or 'ranges'.

as.

Quote from: alrelax on June 23, 2022, 08:28:58 PM
If I understand the latest it would be the following. There has to be three or more repetitive winning hands and then you're looking at the cut. If it makes one or two on the cut, then the following would have to cut again and your trigger would be on the third spot to be a repetitive hand.

Example.

BBBB
P or PP
BB —- >>>>(this spot) <<<< Is the hand to wager on.



So Asym, on a score card laid horizontally or the main road, it would look like the picture below.  You are referring to the spot to wager where I put the X, CORRECT?

[attachimg=1]

Am I correct?

Yeah Al, you are perfectly correct.

Let math experts to say otherwise: 'it could happen that after a 3 streak and a 1/2 pattern, everytime a precise double instead of a 3s streak will show up clustered for long'.
My answer: 'Really, you fkng dumbas.ses?'

Second more polite answer:

-There's an average number of 3s happening at every shoe dealt in the fk universe. So if the number of 3s are going to get a  lower value than average, it means that more singles and doubles clusters must happen. But only half of the times we'll find reasons to bet the above trigger.
On the other end, if a greater than avg number of 3s streaks is going to happen, it means that 3s are more likely coming out consecutively clustered and we won't bet a dime on that.

- Any card distribution (even whether manually and voluntarily placed knowing our strategies) will get a hard time to form 3s streaks followed by a single/double than by another double for long not fitting a kind of 'clustered' streaks pattern (collateral strategy).
In the meanwhile, 'non acute' players (99.9% of bac players at least) will bet toward streaks no matter what or to get a steady single/double line of some kind.

- Quite likely a possible 'losing' distribution is whenever 3s streaks are going to happen by a 1:2 improbable long occurence (average value is 1:3), that is when after a 3 streak a couple of 1/2 events show up making the above trigger as loser.
Anyway just a double is going to make us losers. Not mentioning that a 'clustered single/double player will be winning at all those spots.

- The more probable occurence of such trigger is 1 and not 0. At most situations, no need to chase unlikely trigger situations higher than 1 unless you've registered that too many '1' won't' be balanced by higher values.
Notice that a fair amount of times a single or a double happening at the start of the shoe won't be 1/2 clustered, so followed by a 3 streak, and again a 1 gap pattern happened.

Practical guidelines

if you like to set up a single/double strategy getting 3s limit 'walls' (so stopping or starting to bet up to get a single or a double happening), you know you'll get an avg amount of winning streaks vs losing streaks per shoe.
Of course it's way more likely to get a 1/2 line prolonging whenever a 3 streak will be followed by a single than by a double.
Consecutive 3 streaks do not interest us as being affected by a huge variance.
And obviously more likely (p=0.75) patterns will come out clustered and not singled or, at worst, as singled losers than multiple losers.
Nevertheless, this strategy will face the probability that most shoes are going to get at least one losing spot along the way (that is our beloved 3/1-2/3 trigger).

So we must choose to hope that something weird (albeit being 'natural') won't show up or that a relative unlikely scenario will come out sooner or later.

So let's go back to theory.

Differently than doubles, 3s are more consistent in their appearance (lower sd values): so itlr strong deviated values are more likely to affect doubles than 3s.
It's true that in rare cirucmstances such feature will be disappointed, yet if we'd make as a first condition to bet a 3 streak happening followed by a double (second condition), variance will be efficiently limited.

Now we get a random walk having its peaks of winning and losing hands distributed very differently than a possible 50/50 independent game and not necessarily taking into account the math edge favoring B side.

Instead of verifying such claim but following the scientifical method, let's try to falsify this hypothesis, for example setting up the same random walk now applied at doubles or whatever pattern you'd like to use.

as.

Quote from: klw on June 23, 2022, 06:50:15 AM
Surely a perceived edge is diluted the more decks a casino uses.

Exactly.

If a finite number of 'key' cards move around a more limited field, odds that 'no key cards' distribution segments will whimsically affect the real outcomes will be way restricted than when using a larger field (more decks).

as.

"impossible that any card distribution won't get at least one 'strong' shifted situation to happen very shortly."

    Do you mean in any one full shoe, or say within 3 spots,  5 spots? 8 spots? Other?
I do agree with your thesis above. Im just wondering what has been your experience with the upper end extreme u typically observe. In your opinion what are the merits for: Choosing to do a slow negpro or steep negpro til u catch a W, or no negpro=Flat, or Pospro, when pursuing this keyhole spot for the W?

HI KFB!!

With some experience and after having collected a large sample of live data, you'll see how much average 'concentrated' shifted situations will happen in relationship of the actual shoe.
At most situations, you do not need to spot them by quantity (long streaks of something) just by quality.
Pros I know do not bet that something must happen for long, instead that clusters of anything move and stop within 'more likely terms' surpassed whom they are not interested to bet 'em anymore.
For example, the 3/1-2/3 specific pattern obviously won't show up per every shoe dealt, even though the number of 3s is higher than average (as more 3s are showing up, more back-to-back 3s are going to happen, so enlarging the room for 1/2 clusters greater than 1).
Since the 'enemy' counterpart of 3/1-2/3 specific trigger is 3/1-2/2, we know some shoes will produce a 'long clusters of the latter pattern, making worthless or at least quite risky to set up a negpro on such shoes.
In some way, we should do a lot better to wait that a 'shorter' amount than average of 3s will be followed by the non bettable 3/1-2/1 pattern at some levels, then wagering when that final 1 becomes a 2 so enticing a 3 streak formation.
Notice that we could act in the same way by wagering doubles, but those patterns are more affected by volatility as they are more likely to happen (or not, when cards are so clumped to produce just singles or 3s).
The avg 3s rate per shoe is 8.75, so a possible negpro must take into account either this value and, more importantly, how many 'losing' patterns (3/1-2/2) had happened so far. (Along with other features I don't want to discuss here).

About 6-deck shoes vs 8-deck shoes

When facing 6-deck shoes, best variables to look for is whether shoes are manually shuffled and if very few cards are cut off from the play.
If those two parameters are fulfilled, it's virtually impossible to lose ITLR as patterns will be more consistent than at the 8-deck counterpart.
Meaning that many univocal patterns (and there are many 'clusters' to look for) will stand longer than at 8-deck shoes.
Now the 'quantity' takes a primary role about 'quality'.

It's now that the already mentioned 'code' strategy (if properly evaluated) will get the best of it by its various 'number' steps distribution.
Different code values move around the probability to be clustered or not, at the same time a perfect balanced code distribution is out of order. So some numbers must present clustered, especially if we merge two or more numbers together.

as.

Thanks KFB and thanks Al!

KFB:
my samples considered the 3/1-2/3 attack and each W spot is +1 (minus vig when applicable) and L= -1
1-2 trigger is any single or double happening after a 3+ streak.

Almost impossible to get a multilayered progression to lose as it's virtually impossible that any card distribution won't get at least one 'strong' shifted situation to happen very shortly.

There are 96 'key cards' to fall here or there along with 128 zero-value cards. The remaining card combinations are just flowing without 'strong' math reasons to produce a winning hand at either side, itlr such part of outcomes will equal.
Imo we must spot such 'key cards imbalances' and not trying to win at the way more likely 'remaining' whimsical situations. 

And btw, the lesser the amount of decks utilized to form a shoe greater will be the probability to catch those 'imbalances' by a proper 'key card' propensity assessment.
That's why in Vegas and in Asian casinos 8 decks are utilized instead of the EU 6 decks.

In some way it's the same concept why casinos have cut off single deck black jack.
At bj, math is the king and at baccarat statistical distribution is the queen (but think at chess, queen has more moves to make than king).
So if at a single-deck bj game all aces are removed very early and all 5s are live, we could easily tell the casino 'go fk yourselve). But we have to bet anyway, so taking the worst of it.

However at baccarat a low level of predominance may or not may stand for long, yet we know which more likely ranges such low levels will take and of course a relative counterpart predominance must happen sooner or later. Always taking into account how many hands were dealt and how many hands are left to be dealt.
With the luxury to take the direction we wish and whenever we want, differently to bj.

Al is perfectly right on that: if a given line seems to stand for long let's ride it, otherwise stay away.
Maybe it's a seemingly more 'gambling ' approach than waiting for less likely favourable spots to show up but getting its value anyway.

We should remember that baccarat is a dependent card game not in terms of which side should be more entitled to win but about the more likely winning ranges dictated by the actual more or less unbalanced card distribution

Let's consider another wonderful approach to get the best of it, that is the already mentioned 'clustered streaks' factor.
Now taken from an 'opposite' way of thought, that is that predominant sitautions must stop when they reach a given deviated value.

Theory: streaks of a certain lenght are more likely made by predominant one-sided events that reach detectable ranges. In a word, when key cards are naturally consumed, a more natural hopping situation will take place, so making long whimsical streaks less likely to happen.
Always by a clustered fashion considered at given levels as shoes are unlikely to provide events getting huge 'non clustered' situations.
This theory was demonstrated to be so powerful in practice that it's probable you'll lose interest in playing this game anymore for a lack of 'suspence'.

Set up your 'random walk' on streaks of certain two-level lenght (2-3, 3-4, 4-5) and make them to show up by a clustered fashion vs an isolated fashion.
Test a lot of live shoes and you'll see that in specific situations the general probability to win (75% as you'll bet two spots to get a cluster) will raise up to 85%, a wonderful 20% edge minus commission (when applicable).

Moreover a fair number of shoes are so polarized that the opposite unfortunate event won't happen for long, so making the 'breaking down the house' more a reality than a dream.

Remember that math can work just on perfect random and independent propositions, and we have strong reasons (verified by practice) that baccarat results are affected by a kind of unrandom factor accompanied by a natural asymmetrical card distribution enticing the formation of some patterns than others.
Itlr.
In the long fkng run.

as.

Another samples

LLW
LLL
WWWWL
LWW
WLL
WWWL
LWWLW
WLLLL
WWW
LL

W= 18 L=18

WLWWLL
WLWLW
WL
WWL
LWL
WLWW
WWWW
WLLW
LW
WLL

W=21 L= 15

LLW
WWL
WLL
WL
W
WLWLL
LLLL
WW
LL
WLLWW

W= 13  L= 17

Here's the attack made on real live shoes randomly taken:

LLWL
LW
LLWLWW
LWW
WLWL
WW
LW
WWWL
WWWW
L

WWWWL
WW
WLL
WLLWW
WLW
LL
WL
WW
WWW
WLLWL

WLWWL
WW
WLLL
WWLW
W
WW
LLLW
LLWW
WW
LLL

Total W= 55
Total L= 39

A fluke for sure, so test your shoes to disprove such findings.

as.