Seventh sample taken randomly (10-shoes data considered as first and second pattern):

WW, WW, WW, WL, WW, WL, WW, WW, WW, LL.

Eight sample:

WW, LW, LW, WW, WW, LW, WW, LW, WW, LW.

Ninth sample:

WW, WW, WW, WW, LW, LW, LW, WW, LW.

Tenth sample:

WW, LW, WW, WW, WW, WW, WW, WW, WW, WW

11th sample

WW, LW, WW, WW, LW, WW, LL, LW, WW, WW.

12th sample:

WW, LW, WW, WW, LW, WW, LW, WW, WW, WL.

13th sample:

WW, WW, LL, WW, LW, LW, LW, LW, WW, WW.

14th sample:

WW, WW, LW, WW, WW, LW, LW, WW, WW, LW.

15th sample:

WL, LW, WW, WL, WW, WW, LL, LW, WW, LW.

16th sample:

WW, LW, WL, LL, WW, WW, WW, WW, WW, WL.

17th sample:

WW, LW, WW, LW, WW, WW, WW, WL, WW, WL.

18th sample:

WW, WL, LL, WW, LW, WW, WW, WW, WL, WW

19th sample:

WL, WL, WW, WW, WW, LL, WW, LW, WW, WW.

20th sample:

WW, WW, WL, WW, WW, LW, WW, LW, LW, WW.

I can run my datasets forever or any other reliable bac source and things won't change.
Maybe some harsh variance could come out along the way but itlr everything must be placed accordingly to that statistical appearance.

In this 140 shoes sample we got:
- 47 W clusters;
- 11 W singles;
-  7 L clusters;
- 41 L singles.

But that's just a start, there are more powerful tools to take advantage of.

as.

b]Gaps[/b]

It's obvious that the game could be beatable by qualities and quantity qualities and not by mere quantities.
A quality (I consider the term in a wide sense) needs more linked factors to show up, so the field of randomness should be somewhat restricted especially when it's not a real randomness.
 
Let's make an example.

Assume we're using a sky's the limit progression wagering that the very first pattern of any shoe dealt won't be a 3+ streak (a streak longer than 2).
So hoping that the first pattern will be a single or a double.

I take randomly 10 shoes from my datasets and see what happens.

L, W, L, W, W, W, L, W, W, L.

If we use a 1-2 progression for any step progressively raising the bet after a two-step loss (1-2, 4-8, 16-32, 64-128, etc), this sample would be a winning one featuring just single losses between wins. Despite of the total W/L ratio by flat betting accounts to -6 units (vig ignored for simplicity) as W=6 (x1) and L=4 (x3).

Arrange this WL succession into all possible permutations and it could happen, albeit quite unlikely, that the sequence will look as L,L,L,L,W,W,W,W,W,W.
Now our progression would be as 1-2 (L), 4-8 (L), 16-32 (L) and 64-128 (L), that is 255 units spent to win just one miserable unit.
We needed a 256-512 progressive unit plan to recover all the previous losses (but vig could lower that ratio).

So far we may infer that playing singles/doubles vs 3+ streaks at a preordered point could be a very dangerous BS plan, actually it is even at any other point of the shoe considered.

Another observation we can make at the original succession is that after a L every next outcome will be a W.
On the other hand betting W after a W provides just one loss being followed by two wins.

Finally, when the number of L is inferior than the number of W at a 10-shoe sample, at least one WW clustered event will mathematically happen.

Anyway by wagering this exact first pattern situation, we're playing a quantity.

Go on and see at the same original succession what's the second pattern coming out after the first one whatever it is, always in terms of W=single/double and L=3+ streak.

L, W, W, W, W, W, W, L, W, W.

Now our progressive plan no matter the possible permutations involved remains good as no back to back L came out.
Moreover the L singled outcome trigger remains good and just the first and second pattern produced two L in a row.

By flat betting (1-2) this second pattern we got a +2 units, so we are still behind 4 units (after vig).

Let's take another 10-shoes sample and see what happens.
First pattern:

W, W, L, W, W, L, W, W, W, L.

By FB (1-2) it's a -2 units loss.
The single L trigger stands and the same about WW clusters.

Notice that overall we got -6, +2, -2 so we're still behind 6 units.
Obviously by adopting the sky's the limit approach so far we did't get any bust.

See the second pattern of this second 10-shoes sample:

W, W, L, W, L, W, W, W, W, W.

Again +2 units by FB, single L trigger remains solid, W clusters quite good.
No matter the permutations.

Maybe someone could see that the second patterns are more likely to produce a W succession after a L one (and perhaps vice versa) but that's not the point.

Let's see about a third 10-shoe sample taken randomly.

W, W, W, W, W, W, L, W, W, W.

Easy game, huh?  Not really. 
The total count by flat betting is +6, so erasing the previous deficit (again before vig).
Notice we are considering 30 shoes.
But a single spot we didn't have to put in action our pogressive plan.   

Second pattern

W, W, W, L, W, W, W, L, W, W.

+2 by FB, just two spots needed a first-step progressive plan.

Fourth 10-shoe sample (again taken randomly).

L, L, W, W, L, W, L, W, W, W.

Same considerations about the first 10-shoe sample, but here we got a LL sequence.
By FB our total account for this sample is -6.

Second pattern

W, W, L, W, L, W, W, W, W, W.

'Randomness' is so capricious, again a W/L 8/2 ratio (+2 by FB).

Fifth sample.

W, W, W, W, W, W, W, W, W, W.

Wow, no 3+ streaks at the very start of any shoe.

Second pattern

W, W, L, L, W, W, L, L, L, W.

It seems we can't stay ahead by FB for long, -9 units by FB and the L singled trigger seems to not working.
Fortunately W clusters keep winning but it's a coincidence as 5 L and 5 W could easily distribute to get multiple singled W situations.

Sixth sample

W, L, L, W, W, L, W, L, W, W.

FB = -6 units
Now we have two W clusters and two singled W situations.

Second pattern:

L, L, W, W, W, L, L, W, W, W.

FB= -6 units.
No L singled situations and two W clusters.

If we'd think that betting towards singles/doubles vs 3+ streaks could get a kind of direct advantage, well it seems it's not the case.
So far our progressive plan got the best of it as only one time we had to utilize the fourth level of the progression, but we know that some different permutations would make us to lose our entire bankroll (either for bankroll finitess and for the maximum limits).

Are there other tools coming at our help to mitigate a negative variance or to raise our probability of success?

Let's consider now the back to back result (first pattern and second pattern) per each shoe dealt.
Again W= single or double and L= 3+ streak

Now the picture looks as

1) LL, WW, LW, WW, WW, WW, LW, WL, WW, LW

2) WW, WW, LL, WW, WL, LW, WW, WW, WW, LW

3) WW, WW, WW, WL, WW, WW, LW, WL, WW, WW

4) LW, LW, WL, WW, LL, WW, LW, WW, WW, WW

5) WW, WW, WL, WL, WW, WW, WL, WL, WL, WW

6) WL, LL, LW, WW, WW, LL, WL, LW, WW, WW.

Now the L clusters are well more defined in their distribution, more likely roaming around a 0 point.
On the other end and despite a slight than average apparition, W clusters distribute more clustered than isolated (of course after having considered the 3:1 probability ratio).

as.

Thanks for your detailed answer, as always KFB!

I think that almost no one bac player will take care of the important 'streak rate'/bet amount ratio you've written, choosing a more 'confused' approach dictated by the actual results.

I mean that whenever we choose to bet a higher amount than our standard wager, we ought to consider that such bet should involve a greater than expected probability to win and not hoping to overcome a natural temporary negative variance.

So a flat betting player must solely rely upon a profitable bet selection working for him, maybe a progressive player of any kind could implement to his strategy additional tools as the streak rate compared to his bets, etc.

Unfortunately too many players think that raising the amount alone (at either winning or losing side) will do the job, forgetting that their bankroll is minuscule related to the house's one and just in case there are maximum limits to stop their action.

But the worst thing to do while losing is starting to wager the side bets that besides the Dragon Bonus at P side are heavily burdened by a 7% HE or more.

Thanks if you have time to further elaborate your thoughts, KFB!!

as.

Hi KFB!

- horrible MM
- starting to desperately trying to get the Money back by wagering side bets
- Betting side bets no matter what
- Betting huge after losing and conservatively while winning

What's your opinion?

as.

Unfortunately it's very difficult to follow casinoscores as many results are missing from the displays and there's no interruption between shoes.
The only way to collect outcomes is to write down them from the streaming, but even this sometimes doesn't work.

So end of the 'experiment'.

as.

Sorry, it's very harsh to follow REAL evolution results and writing down my forecasting bets.

I have to adopt my strong bet selections to do that, inthe meanwhile writing down manually the outcomes.

new fresh shoe

PT so far

PTB

BET B 1.5

B WON

NB

P 1.8
TIE

NB

P 1.5

P WON

BET B 2.6

B WON BY 9 OVER 6

NB

P9

NB

P WON

BET P 2.1

LOST

AGAIN BET P 2.1

LOST

BET P 3.0

WON
NB

BET P 2.6

WON

BET P 2.8

WON BY A N9

NB

BET P2.0

WON BY A N9

BET B 2.4

WON BY A 4 VS 0

NB

P 1.8

WON P4 VS B2

B 2.4

LOST 9 OVER 3

NB

P 1.9

WON 8 OVER 7

NB

NB

B 2.2

LOST 1 VS 8

B 2.5

LOST

NB
NB

BET B 2.7

WON 1 VS 0

BET B 2.5

LOST

NB

BET B 2.8

WON 8 VS 0

GAIN SAME BET B 2.8

WON 5 VS 1

NB

N9 AT B
NB

B 2.2

LOST 9 OVER 3

SAME BET B 2.2

WON 7 VS 6

NOW B 2.8

WON B7 VS P3

NOW P 1.2

WON 8 OVER 1

NOW P AGAIN 1.5

LOST 2 VS 9

I CALL IT A NIGHT

SEE YOU TOMORROW AND SORRY ABOUT THE ISSUES COMING OUT, AFTER ALL IT'S JUST A FIRST ATTEMPT TO DO THAT

TOMORROW I'LL START A NEW SESSION AT 8.30 GMT

SEE YOU AND THANKS FOR YOUR PATIENCE!!!

as.


 

It's 3.08 GMT

#1. Last results are BPTPPBBBBP

From now Ties won't displayed and the before tie bet still stand otherwise indicated

#2. last decisions are PPPP

#3. PPPPP

#4. NB of course

#5. bet B1.5

#6. Won 1.5 was B

#7.
After a tie a B should have won

#8. Ok, np, you have to manually write down the results
Scores are inaccurate

#9. Keep track of the streaming results and not of displayed outcomes!!!!

#10. B 2.0

#11. W B 2

#12. NB for long, I have to track manually the results
Anyway up of 3.5 units before tax

#13. B 2.5

#14. Won

#15. so far 5.5 units won before vig

#16. tie NB

#17. banker wins by a 5 point

18. player wins by 3-5 N

#19. banker wins by a 5-4 draw NB

#20. player wins NB

#21. b wins by A-7
Now bet 2 at B

#22. Tite 7-7 NB

#23. NB

#24. Bet B 1.8

#25. YESS!
Won 4-5

#26. NB

#27. banker won by a 8
Bet B 1.0

#28. easy fkng game won

#29. NB

#30. player won by a 6
NB

#31. P5 B4 player won
NB

#32. P won by a 5-4 point
NB

#33. Bet Player 2.5

#34. LOst

#35. P wins by a 6

#36. NB

#37. easy P bet 2.8

Stop talking about theory, let's put in practice all the bigh.orn.s.h.it I've stressed you so long here.

Taking an idea from another forum's member, I'll make fictional bets in real time at live outcomes.

Sessions will be made following casinoscores.com/lightning-baccarat site.

We pretend to get a $1.000.000 bankroll by wagering $10.000 unit bets, maximum bet will be $30.000 . Reason to consider a 3x standard bet is because I do not want to make an endless series of NB (no bets).
Bets will be written under the 1 or 1. form, so for example a $15.000 bet at Banker side will be a  B1.5

Games are assumed as normal commission games (5% vig).
vig will be acconuted at the end of each session .

Bets will be placed as:

B = Banker bet
P = Player bet
NB = No bet

I'll make my best efforts to spot the next bet ASAP and to write down the actual W or L result.

Despite that, I've recently experienced a very bad connection on that site, so whenever this thing happens I'll simply report this.

For simplicity time considered will be GMT.

Each session will be displayed by a fresh thread on my section (365FB and the number) 

Let's play!

as.

Nice post!

as.

Sorry I have some issues to display the data, hope to fix the problem very soon.

as.

Very true, yet imo there's a difference between 'trying to adhere at most at the actual shoe' and 'trying to adhere at most at the actual shoe kwowing the more likely patterns ranges'.

One of the best advices you gave us Al is that while crossing a winning situation, people tend to get confidence 'too late' in the process instead of 'pushing' sooner.
I've found this attitude to be a strong mistake as more often than not positive things become less and less probable.
Of course negative situations appear to come out endlessly, so there's no point to press anything just watching.

Later the data I was talking about yesterday.
 

as. 

At baccarat there are sure indeniable bet selections getting a slight edge over the house (after vig, of course) so bac results are affected by a kind of dependency neglecting a perfect unbeatable randomness.

The problem is that such 'slight' edge is quite dispersed within the various successions, needing some 'room' to show up.

Tomorrow I'll present many real examples of that, maybe it would be helpful to understand that a lot of times baccarat is more a silly than wonderful game even though we're playing with a verified edge.
 
as. 

Definitely it's a nice post.

Maybe the most important and difficult factor to grasp is the fourth one, needing a lot of experience and harsh losing sessions coming over the player's shoulders.

as.   

Postulating that bac shoes are not randomly shuffled doesn't mean that streaks of 'specific' something all of the time will take a univocal direction longer than expected, it would be too easy to exploit the game.
In fact we can't know what will be more likely to happen per every shoe dealt as things continuously change so not privileging one side or the other one of the operating world.

That's why the only tool we can rely upon is the watchdog of randomness: standard deviation values.

Beyond any doubt bac shoes are not perfect randomly shuffled but it's very likely they are offered quite close to that, so we should learn to distinguish when and how much the unrandom world will take a practically exploitable lead over the unbeatable random world.

Therefore we should think of our bets in terms of winning 'ranges' where most part of them will fall into the random EV- proposition but some of them do incorporate a greater than expected winning probability capable to erase and invert the HE working at all other bets.

It's the same math concept why Banker wagers are less worse than Player bets: most of the times they don't, all of a sudden they strongly are.

Then it's intuitive to think that the 'independence' factor cannot work at baccarat as unrandom shoes sooner or later will feature a kind of dependence more likely showing up at sensibile levels after the formation of certain 'complex' events that tend to restrict the power of randomness.

In a nutshell and differently to any other gambling game, at baccarat each shoe is a world apart where most outcomes are randomly offered but some events (due to the unrandom shuffling nature) are way more likely to happen than what a pure random world dictates (e.g. sensible lower sd values).

Our advantage comes right by selecting 'probability' ranges where one or more bets should involve a strong EV+ capable to proportionally erase and invert all the other wagers made on that betting range (where half are lost and half are won by chance), the same way why itlr B>P.

Fortunately for us, regarding baccarat mathematicians and gambling 'experts' have made two fatal mistakes:

a) Taking for grant that bac shoes are really randomly offered;

b) It's the corollary of the above point, that is considering baccarat as any other gambling game where the 'whole' findings (infinite shoes) matter instead of focusing about 'single shoe' dependent features and properties.

Baccarat works the same way as poker: it's better to appear stu.p.i.d than smart but with a substantial difference: itlr at poker some players are detected smarter than others, at baccarat we are all stu.pi.d.s with no exception.

as.

Lol, I like 'Mr Fu'...I guess Fu stands for 'luck' isn't it?

Yep, that succession was awesome and of course scientifically unsound (actually it's not strictly speaking), yet I guess most bac players would have collected many wins from that.

Is it rare to happen? Sure!
Are there other ways than following it in order to get multiple back-to-back winnings? Nope.

Independence

The notion of independence, which is, in a way, the heart of randomness, presents a major psychological obstacle.
This obstacle involves severe fallacies concerning random walks (Falk).

We have seen that 'normal' people tend to assign (improperly) a too much 'overalternating' strenght at random binary successions whereas bac players tend to do the opposite, that is hoping that random binary sequences will get homogeneous situations of many kind around any corner.

Obviously as long as the baccarat production is really random, both different ways of thinking 'probability' doesn't lead to nowhere as random=unbeatable.

More on that later

as.