Hi KFB!

The BP model is biased at the start for two different reasons.

1- B is more likely than P just on rare occasions, I mean it's not steadily more probable than P as the overall B/P probability is assessed on the very long term, many times mixing up different sources of data and/or considering pc simulated shoes.

2- any live bac shoe dealt in the universe is affected by a kind of non randomness.

#1: I've provided here and for the first time publicly (for free) the math values why and by which degree B>P.

#2: math values work only at real random propositions, yet we can assume that such randomness doesn't affect most part of live shoes dealt.
But we need more advanced tools to ascertain a possible non randomness, namely getting rid of simple B/P successions.

Several years ago Ed Thorp devised a card counting scheme assigning a value to Banker's positive cards and Player's positive cards.
After a cutoff point was reached (needing many many hands to show up on average), Thorp demonstrated that a side was more likely than the other one and, surprise, only the Player's side got a small positive edge over the house (0.33% or so). Banker side remained negative no matter what.

More recent studies (always apllied to pc simulations) have shown the in very rare occasions some shoe distributions could get the Banker the 51.3% cutoff probability capable to erase and invert the negative HE. Unfortunately being so much restricted that it's considered as worthless.

Everybody here knows that a card counting technique cannot be a viable option to overcome the negative BP edge.
But at least we may conclude that baccarat is not made by independent successions and that even though B>P, one study showed that the only profitable situation to be ahead of the game is by waiting a strong deficit of Player positive cards.

Another eminent gambling expert, James Grosjean, stated that "the game is symmetric so that are no cards that massively favor one bet or the other".

Actually and knowing the average asymmetrical key cards impact acting at every live shoe dealt we should transform the above statement into "the game is so symmetrical that asymmetrical spots will tend to get more probable cutoff values than expected".
And of course B/P values tend to be way less predictable than A/B models.

B/P classifications must consider a BP distribution acting at every hand dealt without any previous consideration, yet A/B models could start the registration after precise situations had arised (or not).

I mean that the more hands we must classify to get a A/B result, higher will be the probability to get A after B or B after A.

Consider this shoe sequence:

PPPPP
BB
P
BBBBB
P
B
BB
PPP
BBB
PPPP
...
as a

A
B
A
B
AA
BBB
A
B
A
BBB
A
BB
A
BB
A
B
A
B
A
B
succession.

We transformed the BP sequence into a AB succession by a  simple mechanical way.

Are we more favored to detect the AB sequence than the BP original sequence?

What about very long term situations?

as. 

Quote from: alrelax on June 28, 2021, 01:55:08 AM

Please do not take it the wrong way, I'm not criticizing your writing and your comments, I'm just making an observation to stress the drawdown, the patience required, the frame of mind and everything else that is mandatory in trying to catch all of this, which I've learned over the years.

I do not, Al and I never did.
I take your observations very seriously, after all what you say it's very effective at the tables.

I just want to reiterate the idea that baccarat can get the player a sure indeniable edge by a strict scientifical point of view.

To beat baccarat nobody is totally wrong or totally right, thus we better put different ideas together (imo).

as.

Quote from: KungFuBac on June 23, 2021, 06:09:07 AM
Thx AsymBacGuy--good thread/ post.

Your sentence:
"... An average card distribution will more likely produce clustering win situations. The more we are considering winning clusters by strict parameters, higher will be the probability to win.  ..."

Can you clarify the phrase in bold? or give an example. thx

Continued Success,

Hi KFB!

Say our method is tested on 70.000 live bac shoes.

The probability to be ahead by randomly selecting just one or more BP betting spots per shoe is very close to zero, even if we are acting after knowing the exact BP distributions happening at this 70k shoes data.
Ok, maybe 12+ P streaks are going to get a B hand more often than not (vig considered) but the ROI still remains negative. Not mentioning that the hands observed/EV ratio will be constantly shifted toward the left side. No matter how deep we'll start the betting.

I mean no general rare BP triggers coming up along the way are valuable itlr as the actual shoe card distribution will make a decisive role about it.
Shoe per shoe.

Variance

When we transform a 'random' BP succession into multiple AB unrandom sequences, the variance will be way more restricted than expected. For good peace of mathematicians.

It happens that BP successions are way more affected by volatility than AB sequences, as B or P are going to get a place just by the simple nature ot the actual outcome, whereas an AB model must take into account a back to back probability made on multiple situations before getting a place.

I mean that itlr it's way more likely to get multiple detectable AB situations than simple BP successions as the former category cannot give a damn about the asymmetrical outcomes nature acting just one time over an average number of 11.62 attempts. Sd values will help us.

Summarazing.

- all shoes dealt are affected by a sure asymmetrical card distribution acting by some values;

- the fact that B>P should be considered as an irrelevant factor as AB streaks are way more detectable than BP streaks.

- the probability to get a given AB pattern is at least 1.5 times higher than the BP probability to get the same pattern, providing to get rid of the unplayable shoes.

as.

Besides other topics, Alrelax is deadly right about the importance of ties affecting the BP distribution or, generally speaking, our plan.

Recently I've crossed a really weird live shoe (a manually shuffled fresh shoe):

first card is a 5, thus five cards were burnt. First hand is Player, second hand is Banker

1)  3-6  K-9  TIE

2)  3-T (3) Q-6  TIE

3) 8-T   K-8   TIE

4) 5-2   4-J (3)  TIE

5) A-9 (9)  K-Q (9)  TIE

6) 4-5   Q-9   TIE

7) 2-2 (6)   4-K (6) TIE  That's amazing....a complicated way to get the seventh tie in a row when a simple third zero value card could have made the job.

2-A (2)  8-7   TIE

9) T-2 (  4-3  Banker wins, end of the ties streak.

In summary this shoe presented 8 ties in a row at the very start of it; just two over eight tie hands employed six cards (the more likely situation to get a tie); at hand #7 people at the table went crazy.

Ties are one of the worst betting opportunity at baccarat, yet jackpots happen.
And in reality few people have made serious money at this shoe.

In my tests I've run hundreds of thousands of pc simulated shoes and this thing never happened at the very start of the shoe, just a couple of times 8+ ties came out but in different positions. And of course my live shoe data consider a significant smaller percentage of shoes.

I mean that if sh.it happens, well, jackpots must happen, especially at a game where we have reasons to think that each 'spin' cannot be perfectly independent from the previous one.

Math goes right down the toilet whenever each distribution isn't perfect randomly shaped.
That is nearly 100% of the times.

as.   

Hi KFB!

I think that at gambling games more imperfect informations a player will get higher will be the probability to lose.
At baccarat everything seems to be so "volatile" that players' efforts to look for predictable results are worthless. Imo, this is not the case.

Baccarat results move around two distinct fields of probability:

- the math probability to get B advantaged over the P side;

- the average card distribution probability eliciting patterns of some lenght. That's the main factor we should be interested to assess.

Itlr, the vast majority of patterns could be restricted into precise lenght situations up to the point that we can consider B=P.
After all, an 8-deck shoe will present, on average, just one more B hand than P hand. Thus enlarging a possible B streak at one spot or shifting the P sequences at one spot.

Obviously CSMs deny a sequential probability of some kind and even though we can assess the BP distribution by multiple derived roads, the lack of dependence factor will invalidate the power of the average card distribution issue.

For that matter, many high end casinos know very well the baccarat vulnerability, they'll simply hope that players like to get huge winnings within too short intervals or liking to wager the insourmountable negative edge apllied to side bets.

[b]Simple back to back outcomes and complex back to back outcomes

It's the key to win itlr.

An average card distribution will more likely produce clustering win situations. The more we are considering winning clusters by strict parameters, higher will be the probability to win.

The clustering effect will form situations of different lenght, anyway we are interested about back to back W or L spots.
We know there's a general probability to get singles and doubles, the probability to get losses in such sequences is specularly placed as we shouldn't consider as B and P as opposite results.
Anyway, all streaks surpassing the 3 cutoff point are going to our favor as they'll produce opposite situations from a A/B point of view.

as.

Q: Do you know of houses that do this on a regular(daily) basis  ?

Thx in advance,
[/quote]

Best example is the Salon Privés at Monte Carlo casino in the Principality of Monaco.
Needless to say, it's the most prestigious historical gambling premise in the planet to bet the money at.
In the summer season Salon Privés games (single zero roulette, bj and baccarat) are offered at a outdoor terrace directly overlooking the Mediterranean Sea.
Few players can get the admission to play at this room even though baccarat tables limits are as low as €100-€30.000.

Notice that in Monte Carlo and in most european casinos, no free hands are dealt. 

In Vegas, few casinos are worried about dealing the vast majority of the shoe and very often players  instruct the dealer to stop the actual shoe in order to get a new one.

Of course whenever acute players had considered a shoe as a unplayable one, this stopping procedure will go to their benefit. But in the remaining and more likely occurences, acute players' interest is to get the shoe dealt up to the end.

To clarify things more and for one time taking the Jacobson's book direction to consider the two opposite parts (players side and casinos side) I would suggest:

Players side

- bet only at shoes were most part of the shoe are supposed to be dealt. Preferably when the first card is an ace, deuce, three or four, thus cutting off from the initial part of the shoe just one, two, three or four cards.

- avoid shoes where the red card is too 'light' placed, meaning that too many cards are cut off from the play.

- play at tables where you're not compelled to place a bet for whatever reason, that is tables where more than one person likes to bet every hand.

- play at manually shuffled shoes or Shuffle Master Machine same shoes dealt in alternating pace. 

- for practical reasons serious money can't go unnoticed at Bac Theaters, so true HS rollers must bet at live tables. At live tables nobody gives a fk whether you'll place an occasional yellow or multiple yellow chip denomination, but at Bac Theaters you need to introduce several $100 bills to get a proper bankroll capable to bet the same amount or to endure the invariable losing situations. Not mentioning that the maximum betting limit at BTs is, in the most fortunate case, set up at $5000.
Moreover chips are money in distinct denominations, tickets cannot be splitted.

Casinos side

- only a pc software always starting from a perfect 'neutral' point where everything is equally probable to show up could provide real random results.
In the remaining cases, your card distributions will be affected by a kind of bias. 
SMs acting at the same deck won't fit the random parameters, let alone manually shufflled shoes starting from precise card sequences.

- more hands are dealt and lesser are the decks utilized to form a shoe and higher will be the probability to face players capable to get hints from the actual card distribution.
It's not a coincidence that at Monte Carlo casino (where players can regularly bet 30.000 euros, that is $35.000 at this time, almost the double max limit allowed at Vegas casinos) shifted a 6-deck shoe offer to a 8-shoe offer cutting off from the play at least two decks.

- besides the above considerations, the only sure way to neglect a possibile (sure) player's advantage is by dealing bac hands by a CSM, that is by totally denying a possible back to back influence over the outcomes.

We'll see more deeply this issue in a couple of days.

as.

Think that our claims consider all possible outcomes' successions and strictly measured, classified upon the same shuffling procedures made at the same deck.

Whenever a new deck is offered (HS rooms) we have accounted a general probability compared to the actual probability, so unless cards are precisely dealt by a software (and even if we suspect this fact, the post manual shuffling happening at every HS shoe dealt must neglect such possibility), the probability to get equal or opposite results at back to back shoes remains quite asymmetrically placed.

Btw, casinos have no interest to shuffle cards in a certain way to promote players to lose.
That's a total nonsense, it suffices to study the 4 derived roads directly displayed on the screen.
There's no one BP succession  in the world to get all losing sequences on all four derived roads and even though they know our precise preferred personal random walk we like to use, they can't arrange cards to get multiple losing sequences at back to back shoes.

Casinos get their huge profits at bac tables about players' ignorance or fake statements and not only about the math edge.

Most players like to bet upon asymmetrical situations lasting for long, a kind of trending based action, unfortunately asymmetrical situations will proportionally mix with symmmetrical situations and unless those spots are math studied and properly classified and measured, the EV will be negative.
By a 1 trillion % degree.

The same about the Banker math propensity.
Banker bets are better than Player bets by getting a lesser than 0.18% ROI disadvantage.
Anyone who is used to play at HS rooms knows very well the commission weight, I mean that commissions add up at the end of the shoe, very often producing a total loss.

After all, our bets must erase or invert a more than 1% math edge, thus no help comes from lowering it by a st.upid 0.18% long term value.

It's more likely to get a better than 50% win rate at P bets than getting a 51.3% cutoff value to get B bets to be worthwhile as the asym strenght favoring B bets come out one time out of 11.62 hands on average.

But it's whenever we consider the BP sequence as A or B result successions that we can get a better idea that no one side is particularly shifted toward one side as the actual card distribution will make a huge role on that.

As long as A or B are different than B or P, well we're playing a winning game.

as.

Quote from: KungFuBac on June 14, 2021, 04:11:52 AM
Good, Better, and Best.

Its difficult at times for us to pass on the Good/Better spots and wait for the Best. However, the latter is certainly more lucrative/yields a better ROI. 

Perfect!

And we can bet everything we have on our name that long term winners wager only at the Best spots.

It's true that in some shoes Good and Better could last for long thus enticing us to bet a lot of hands, yet only the Best part yields the advantage we're looking for.

Regardless of how whimsical the card distribution seems to be, it will produce a succession whose properties remain the same.
It's just a matter of 'finite space' that the properties we're looking for will present or not in the actual shoe.

Curiously, but no so much, bad shuffled shoes are going to consume less room than good shuffled shoes as in the latter category the symmetry tend to reach 'perfect' thus unbeatable values.

It's a fact that the vast majority of each bac hand will yield a probability quite different than 50/50 or 50.68/49.32 as it strongly depends about the actual card distribution.
In a sense, when a player places his bet he should expect to be quite wrong or quite right, and not equally wrong or equally right.

The above math and commonly accepted values come off from fake 'collectives', that is large samples made on pc simulations not fitting decent conditions happening when we bet real money at real live tables.
And of course considering each step as perfect independently placed from the previous one/s, assuming that the probability to get this or that comes from the same perfect random source.

More technically speaking, that every single card distribution could come out at specific points to break a given strategic plan.

This is a total fkng rattlesnakesh.it.

First, we need a perfect random source to get so called "unbeatable" expected values and of course the vast majority of live shoes do not belong to this category.

Second, baccarat is not black jack where some card clumpings favor or not the player or the house, at the same time knowing the bj player must bet something at every hand dealt.

Third, a baccarat deck is almost entirely dealt, thus endorsing at various degrees the probability to get (or not) an expected situation.

Fourth, at baccarat we have many tools to estimate how much a given card distribution tends to surpass the 'average' card distribution, a parameter that can't disrespect for long certain values, unless very rare situations consume a lot of space.

The 'space' concept was so seriously taken by certain high end casinos that even though the only side bet offered at their tables are ties, 8-deck shoes are played up to 50-56 hands. Then they shuffle again.

Probably those casinos' customers (btw wagering maximum or close to max limits) seemed to be smarter than average, it's quite probable that sooner or later all premises offering baccarat tables will adhere to the same procedure.

Is baccarat a kind of bj game where some features will get the players an edge?

Ooh it can't. Math geniuses state otherwise.
Fortunately for us.

as.

One more shoe.

P
B
PPP
B
P
B
P
BB
PP
BBBBBBB
P
B
PP
BBBBB
P
B
PPPP
BBBBBBBB
P
B
PP
B
PPPP
BB
PP
B
PPP

Total B= 32
Total P= 28

New sequence built on the same features seen above will be:

AAA
B
AAAAAA
BB
AA
B
A
B
A
B
AAAA
BB
A
B
A
BBBB
A
B
AA
BBBB
A
B
AA
B
A
B
A
B
A
BBBB
AAA
B
A
BB
AA
BBBBB

Total A= 34
Total B= 34

Here a slight BP asimmetricity shifted toward B side produced a perfect balanced final A/B ratio. Getting more valuable spots to bet at.

Think that to beat infinite so called 'random' finite and slight dependent successions, we need to transform them into unrandom sequences getting limiting values of relative frequency not fitting the general probability numbers.

Btw, it's funny to see that some math experts like to label baccarat scholars as complete i.d.i.ots.

Really?

Collect your fkng money and face our bets, after all you'll have the math edge on your side.

As long as we can bet whenever we want or not along any shoe dealt, you can put your math edge right on your behind.
In a way or another, some baccarat players know better than you, you must accept this.

Are you going to rewrite statistical laws acting at a finite and card dependent live shufflle deck?
I guess you can't.

as.

Hi KFB!

Let's suppose to face this shoe:

PP
B
PP
B
PPP
BBBBB
P
B
PPP
B
PP
B
P
BB
PP
BBB
PPP
BB
PPP
BB
P
B
PPPP
BBB
PP
BB
PP
B
PPPPPP
B
P

Total B=27
Total P=38

By just considering the mere B/P hands gap we could think as this shoe as being quite asymmetrical, actually it's one of the best example of strong symmetrical hands distribution.

It suffice to utilize a simple "hand converter" to see that most situations are distributed quite balanced along the way. Of course knowing what to look for.

So now our shoe becomes as

A
BBB
AAA
B
AA
B
A
BB
AAA
B
AAA
BBBB
AA
BB
A
BB
A
BB
AA
B
AA
BBBB
A
B
AA
B
AA
B
AA
BB
AA
BBB
A
B
A
B
AAA

Total A=33
Total B=32

In our new sequence, strictly math derived from the original BP succession, some properties have changed and here it's easier to see what to look for before betting.

From this example we could think that the 'symmetry' or 'asymmetricity' concept would be totally relative, depending about what we really want to classify and register.

Of course the average 8.6% probability to get math favorite B hands to win stands, but this probability is hugely influenced by the actual card distribution.

Notice that the probability that an entire shoe will get ALL Banker winning hands at every asymmetrical B math favorite spot is very very slim.
In some sense we should know that when betting P side rarely (and with some reason), the expected disadvantage could be easily more restricted than what the math general values dictate. After all, just one hand out of 11.62 hands dealt (or wagered) will be B favorite.
Being wrong at sym spots after wagering Player side is a far inferior mistake than winning the same sym spots at Banker side.
And of course we need to win very very few spots per shoe along the way. 

Actually casinos like to face multiple bets by people preferring Banker side no matter what as they know that such B aficionados could more easily fall into the card distribution variance.

From the most part, a Player bet fears just two exact card situations:

- Player draws and Banker shows a 4 point (unless Player side gets a 5);

- Player draws and Banker shows a 5 point (the most B math advantaged spot).

At a way lesser degree of probability comes the Player drawing when Banker shows a 3 and third card is an 8.

There are no other card situations strongly favoring Banker side to be payed 0.95:1, thus we can easily assume that baccarat is a kind of coin flip game hugely depending upon the actual card distribution.

as.

Finally back home.

Baccarat vulnerability

People making a living at this game know very well that baccarat could be beaten only at very few spots arising along most part of shoes but not along every shoe.
It's the same concept why bj is beatable, albeit taken from different perspectives. Math issues at bj, card distribution issues at baccarat. 

A baccarat deck cannot refrain to produce more likely outcomes along the way, it's up to us to select what and when certain more likely outcomes should come out or not (and how long).
It's of paramount importance to understand that the vast majority of BP successions could be interpreted as different random walks getting diverse features of certain lenght.

Again the key word is symmetry, widely intended.

At baccarat cards cannot be distributed proportionally along every shoe, even though it could happen that unlikely whimsical results tend to produce "fake" symmetrical spots. Notice that the counter probability to get a hoped result by opposite issues will be specular itlr. So itlr weird unusual card distributions may be considered as neutral.

Along any shoe, the symmetry fights against the asymmetricity by various degrees and by various lenghts; to consider multiple sub successions will help us to better define their 'average' impact.

Since games must be accounted and measured by 'numbers', we should set up 'personal' limiting values of relative frequencies of both symmetry and asimmetricity acting at every shoe dealt.

For example (and assuming B=P for simplicity), a BBBBBBB sequence followed by a PPPPPP sequence is a sure asymmetrical situation, but we need 6 betting steps to find it out.

And a BBPPBBB sequence needs 4 betting steps to cross the same asym finding (this sequence forms two symmetrical spots and one asymmetrical spot).

Itlr numbers considered at various levels can't be wrong whenever the source (card distribution) is asymmetrical by definition.

See you next week

as.

Normally baccarat players consider shoes' outcomes as consecutive successions. Imo it's not the only tool to find out a possible bias and/or to take advantage of game's flaws.

Events (especially 'complex' events needing many hands to form) are distributed asymmetrically along any shoe, yet their pace varies continuously as long as new shoes are dealt.
Thus enforcing or not a general probability to appear whenever a positional study is in order.

The easy objection one could make is that itlr each event will be distributed proportionally at every position of each shoe dealt, but this objection only stands whether every card distribution is perfect randomly produced. And this is not the case, especially when same decks are shuffled back to back.
Anyway, our watchdog remains the sd.

Probably transforming shoes into a mere 8-15 digit number (of events) and comparing those shoe per shoe numbers by a positional study should be the best tool to know which spots are more likely (or not) to show up.
The 8-15 number is just an indicative value posted for practical reasons, actually the more we are restricing our field of registrations higher will be our probability of success.

Let's make an example.
Say our first shoe is read as 21513123413 (a real shoe, btw). 

Now we are facing the next shoe trying to get some hints before betting.
Since I have omitted the general probability why such numbers will form, we could think that an option might be to get the new shoe producing more different positional numbers than equal positional numbers (or vice versa if you knew the exact events general probability to happen).

Of course it's way more practical to bet that numbers will differ from simple values, for example numbers being equal or different than 1 or 2 at the same positions.

Anyway, the next same shoe shuffled by a CSM (the very next shoe was not considered as belonging to a diverse 8-deck) produced a 422162113241 sequence.

1) 2-1-5-1-3-1-2-3-4-1-3
2) 4-2-2-1-6-2-1-1-3-2-4-1

There are infinite ways to consider such back to back outcomes, anyway we just consider 1s and 2s, that is the six numbers produced at the first shoe (positions #1, #2, #4, #6, #7 and #10) compared to the next shoe same positions.

pos 1: different number
pos 2: different "
pos 4: equal  "
pos 6: different  "
pos 7: different  "
pos 10: different "

Naturally any 1 will need just one step to be different than another 1, whereas 2s need a two-step betting to get a different value than 2 (first step betting toward a 1, next step betting toward a 2+).

Another interesting effect to be aware of after having tested several live shoes shuffled in the same circumstances is that single shoe positions could endure homogeneous results for long, a kind of weird clustering effect of rare events. When such thing seems to happen, best way to take is to simply get rid of that position.

It's important to add that we aren't forced to bet each position by any means, a thing particularly valuable at HS rooms where each deck is a new one.

Finally, the 8-15 events per shoe range was just an example, we could select more deeply our bet selection at the price of waiting and waiting and waiting but in the meanwhile raising our probability of success.

as.

Hi KFB and thanks!

The paper is "Probability in Decline" by Dean M. Brooks.
I've found that some general ideas contained here could be helpful at specific same deck shuffling situations.

Later about your first question.

as.

Frequentist theory of probability

Frequentist probability or frequentism is an interpretation of probability; it defines an event's probability as the limit of its relative frequency in many trials. ...


Words already heard in my posts...

So, ties ignored, is B showing 50.68% of the times and P the remaining 49.32% itlr?
Who gives a sh.it?

Instead let's work on the definition of "event" and thus about its relative frequency.
After all an "event" could be interpreted as a single B or as a single P or as a 5 B or P streak or as a 10 chopping BP line or as any other red/blue succession showing up at derived roads.

I mean that whenever an event is considered by a multiple hand situation, the probability to get the same or opposite event will be slight affected according to the patterns (and quality features) already happened on that line.
In addition, multiple hand events considered shoe per shoe are not producing the same positional features as long as a restricted amount of hands constitute an "event".

Even though a "more likely" world sometimes can't be valuably assessed within a single shoe card distribution, consecutive shoes will enlarge this probability as it's impossible to think that a given card distribution will produce the same random walks for long. Especially whether our betting points (events distribution) are dynamically insensitive to a precise hands' number and position.

Every shoe as a number

Depending upon what events we'd like to register (the few the better, of course) every shoe dealt in the universe will get an "event number" transformed in digits.
Since we're classifying "events", the number 0 doesn't appear in our registrations, when a 10 or higher number will appear we'll sign it as a "X".

Even though the possible card combinations are almost infinite, an already slight propensity to get something at a single shoe must be endorsed by a back to back shoe assessment but not by considering outcomes as mere BP successions but as "events" having a dynamic probability to form.

We've already seen that a given BP succession is directly displayed within four additional forms (derived roads), of course the bead plate displaying ties should be discarded by any means.

Positionally speaking, if any event is already more likely than the counterpart, vertical 'hand insensitive' spots considered shoe per shoe will be even more likely or not, meaning that shoe single digit numbers will deviate from a supposedly perfect random world.

Acting this way we're discarding most of the strong unlikely deviated situations being the heaven for recreational players and the hell for serious long term winners. As they are constituting few spots along the shoe number formation.

Moreover an interesting study has found out that rare events tend to come out in clusters then declining in probability.
The authors of this study claimed that such findings wouldn't get an advantage over gambling games.
We disagree. 

as.

I think that in the complicated gambling world people raised their expertise in different fields, about managing worst drawdowns you seem to be very prepared.

And btw, anyone stating that random successions can be controlled in player's favor no matter what, should get more emphasis than those saying that a game is not so random thus potentially producing a player's edge.
So congratulations are for you.

It's a fact that baccarat scholars like to stay on their findings, without trying to get inputs from  other players to possibly improve their strategy. And this is a pity, imo.

I still consider baccarat as a finite unrandom and multiple factor asymmetrical game; but those features on average will be very slight placed, and not happening valuably at every shoe dealt.

If a MM might get the best of it by wagering every shoe dealt, well chapeau!, yet I prefer to win by a strict flat betting procedure. That's all.

as.