Baccarat Bac Stats

I thought I would start a thread sharing some Bac Stats and so here it is. Can it help? Who knows! Maybe beauty is in the eye of the beholder.

For each 100 shoe / 8 deck sample, I will display 3 things.

1) Session Analysis.
2) Best Bets.
3) Losing Bets.







The first two are pretty self-explanatory. For the third one (losing bets) The first line shows Bank appeared 3,659 times. The longest losing streak for a Bank not to appear after a Bank was 13.
The longest losing streak for a Player not to appear after a Bank was 16.

I have not included any analysis regarding ties in any of the above.

 

Batch 2)

 

Batch 3)

 

Batch 4)

 

Batch 5)







Running the stats are food for thought if nothing else!
One thing that stands out in my mind is how the groups of 4 with at least three Bank in them (so BBBB, BBBP, BBPB, BPBB) roughly perform around 6.50% and the groups of 4 with at least three Player in them (so PPPP, PPPB, PPBP, PBPP) roughly perform at around 5.50%. Now if you were looking for clusters of these groups which pay out at 15/1, could you trap some of the better performing B clusters? Like I say, it's stuff to think about!

 

Looks like this could make sense for template play. Setting up columns for matching pattern betting?

 

Hello Jimske, I was thinking about what you were saying above.

BBBB=1
BBBP=2
BBPB=3
BBPP=4
BPBB=5
BPBP=6
BPPB=7
BPPP=8
PBBB=9
PBBP=10
PBPB=11
PBPP=12
PPBB=13
PPBP=14
PPPB=15
PPPP=16

The groups with three B's are BBBB (1) BBBP(2) BBPB(3) BPBB(5) PBBB(9)

Here is a randomly selected shoe...



Here is a template grouped into 4...



So from the 1,2,3,5,9 groups....3,9,9,9,9 appeared.

More thinking required!

 

Ha...so many different thoughts/ideas come to mind when looking at stats....Looking at the above and thinking about groups of 4 in a kind of template style had me thinking about Lugi's old Balanced vs Unbalanced ideas using a group of 6. Quite a lot of the time, you will get a 4 over 2 towards the bank side to total the 6 hands. You could wait for a 3 vs 2 in B favour and then bet B or wait for a 4 vs 1 in B favour and then bet P.

 

I like the little idea above looking for the 4 vs 2.
Three players up first means no bet.
Not too many bets in a shoe and the variance could be manageable one way or another.

 

eugene how about see true bias where two events strike several times in a row.

For example singles and series of two-strike sometimes three times or five times or ten times or twenty times in a row.
Or you can see singles and series of three or higher strike for five, ten, twenty times in a row.
Or you can see a series of two and a series of three or higher strike five, ten, twenty times in a row.

All shoes show small, medium and strong bias using Sputnik's March and the principal of 1/3.
This way you can catch strikes several times in a row.
My personal solution.

You also have Marigny De Grilleau who show you the math and probability and statistics.
The law of series and measuring them is completely different than your stats.

If you want to read variance and fluctuation and understand where, when and how things happen you need a playing model.
For example singles versus larger series where you want to know the likelihood when the next larger series will appear.

Then you need imbalance and you can measure imbalance or another word variance.

Singles has the value of 1
Series of two has the value of 0
Series of three has the value of 1
Series of four has the value of 2
Series of five has the value of 3
And so it continues...

So seven singles with an unknown amount of series of two are around 1.5 STDV
Pretty common imbalance or variance.

14 singles and 2 series of three are 3.0 STDV
Rare but it happens with regularity

After a certain benchmark of imbalance, you have the expectation that a larger series will appear.
Regression towards the mean.

They can come as tiny, small, medium or large events.
You measuring overrepresented events and underrepresented events using match and statistics.

The French word for STDV is ECART

Someone might want to ride the wave or variance grow stronger the opposite way versus regression towards the mean.
So you can measure the strength and the weakness.

Here is a dummy chart and the formula to calculate the statistical correct value for variance strength and weakness.


First, you have to get the Absolute Ecart when you calculate.
So let's assume you have a sequence with 14 series alternating with two singles present.

Then you take 14 - 2 = 12

Now we want to get the statistical Ecart so we continue with...

14 + 2 = 16

Now we take the sqrt of 16 = 4

And finally, we divide the absolute Ecart whit the sqrt

12 sqrt 4 = 3,00

The Statistical Ecart 3,00

This is just the scratch on the surface how to use statistics based upon real stats, math and probability calculations.

 

In the old days i did many hundred thousands simulations and will end this topic about algorithms or march with money management.
Here i will go deep into how you can measuring the random flow or distribution.
Its based upon taking advantage of the law of series in all existing aspects.

This is the values and existing playing models based upon pure math and probability.
This is how you find you windows of bias or overrepresented events.

Series contra Singles.
Series has the value of 1 and Singles has the value of 1.
There is as many singles as existing series no matter length.

Singles contra series.
Singles has the value of 1 and Series has the value of 1.
there's is as many singles as existing series no matter length.

Singles contra larger series.
Singles has the value of 1
Series of two has the value of 0
Series of three has the value of 1
Series of four has the value of 2
Series of five has the value of 3
Series of six has the value of 4
And so it continues

Series of two contra larger series.
Singles has the value of 0 (you just skip them as none existing)
Series of two has the value of 1
Series of three has the value of 0
Series of four has the value of 1
Series of five has the value of 2
Series of six has the value of 3
Series of seven has the value of 4
And so it continues ...

Series of three contra larger series
Singles has the value of 0 (you just skip them as none existing)
Series of two has the value of 0 (you just skip them as none existing)
Series of three has the value of 1
Series of four has the value of 0
Series of five has the value of 1
Series of six has the value of 2
Series of seven has the value of 3
Series of eight has the value of 4
And so it continues.

Now to the underlying dimension.
You can divide singles only into singles of singles and series of singles.
And you can divide singles series versus a series of series.
With the same math and probability measuring the random flow.

 

thanks for the input.

So choose to bet variance grow stronger or regress to mean. 1) is this counter productive in that the former discounts the later and 2) how to make the calculation in real time? - how to convert to a working strategy?

 

Well without going into detail I just state that when you have a window of events like 14 singles and 8 series of two and 2 series of three.
Then the underrepresented event is a series of four and higher.
The overrepresented window of events has reached 3.0 STDV.

Now you know that series of four, five and higher will show in the future, but you don't know when.
But million simulation show that the worst is 5.49 STDV.
This means that you will not get another 14 singles and 8 series of two without a larger series will be present.

So you can say something about expectation or the likelihood for something to happen, not like having a crystal ball, but not far away.
You can see it that you wait for a vacuum pressure to reach is limits before cracking wide open.

In this situation, you build a march to attack the events to become larger series to win.
That is one example of regression.

The other way is no much different.
For example, let's say you want to be that Banker reaches higher values, imbalance.
Then you decide when to enter and when to exit.

Cheers

 

Sputnik, thanks for going to the trouble of writing up a detailed approach regarding the ECART. I know it's your baby so to speak. I also have to be honest and say it's one of the few things that I continuously don't seem to be able get my head around and it's not like I want a free lunch or anything because I have put in the time. The rub for me though with this approach is that if someone got themselves in a situation where for example the ECART was 3, 4, 5 or whatever, then if they were good enough, they would have already taken advantage of that. Does that make any sense? I certainly am not some maths guy and I have read all the stuff from the 'AP' crew saying that 'regression to mean' doesn't work but I take what they say generally speaking with a pinch of salt because of what I do know but again why would anyone even contemplate it because to my mind the return to mean factor is just the 'second part' doing something similar to what the first part did and if you can't win with that, well then, I don't really know.

 

After talking a bit about balanced vs unbalanced above, I did find an old thread from a now defunct forum with some stats on the matter.

https://web.archive.org/web/20121101003755/http://baccaratforums.com:80/t7427/

 

eugene there is a misunderstanding regression or correction is common and part of random bits behaviour.
It depends on what way you decide to observe the random bits to explore the phenomena.

For example, let's say you observe windows of imbalance where 12 events versus 2 events in any given combination.
Each time you do not find such a window of events you have variance and going up and down i waves depending on the way you observe the random bits.

Now take 12 singles versus 2 series.
Assume you look at singles versus series and want to know how often six singles and two series will hit six more singles in a row without regression.
Then each time you get more than three series among those singles you have regression.

It exists, is true, are math and probability and statistical laws.

Ap might debate that you can not take advantage out of the likelihood of something to happen and call it gamblers fallacy.
But saying that regression is not real is just being stupid.

Everything goes in waves, grow stronger and get weaker, that is the nature of random bits.
Regression is part of that.

What you don't grasp is that correction comes as tiny, small, medium, large events.
For example, each time you see six in a row and followed by three or more opposite events you have regression as they did not grow stronger.

You can use smaller samples and smaller windows of events to see how things grow stronger and weaker in waves.

For example, if I would like to bet on something to grow stronger you only do things the reverse way of things.
Take my example with six singles and unknown amounts of series of two, then three series of three would create a regression or a series of five.
If I wanted to ride that wave to grow stronger that would be my obstacles to avoid or my losing sequences.

The point is when you have a playing model you know what create regression and what does not.
Do you bet for the sequence to grow stronger or weaker?

Same as betting with or against Banke Player patterns with any combination.
The difference is that you use the law of series and a window of events as a selection and the benchmark is what you bet with or against.

eugene no one can say regression is fiction because is part of the random game at all times.
But to give one example you need to decide what to explore and measuring.

Cheers

 

So when ECD attempted to use weighted count based on averages he hoped to measure and predict what goes stronger and weaker. He was looking for regression.

McVince pointed out that his weighted count did not help future events so he decided to use an unweighted count using only 3 counts. No. of 1's, 2's and 3+. This allowed him to either bet for a chop (alternate) or a repeat. Never bet 2 goes 3. A good simplification IMO. I don't think they ever did a study deciding how well it worked. IMO this probably worked better when the count became more extreme. But maybe the flow or direction of the count is a better indicator.

 

There's a blast from the past. (Did Dr. Tom get his idea from this?)

Egalite played around a lot with this stuff. I've examined this and it seemed to represent the same lengths as B and P or converting to OTBL and TBL, etc. IOW, not changing the average runs. But looking at it from the perspective of template betting a different story.

 

Jimske can you post or email me about the weighted count, need some examples with Banker/Player results.
Then I might come with some working examples.

Cheers

 

You are not wrong! A lot of fluff to go through but one or two interesting threads in the archives.
I had a game online this morning for a few shoes playing for the 6 vs 4 unbalance in favour of Bank just betting when it's either 4/1 to Bank (bet Player) or 3/2 to Bank (bet Bank)

L WW LL WWW LLL WW LL WW L W LL WWW LL W L W L W L W

Using the divisor gained me 3 units with a highest bet of 2. I prefer the divisor to Holloway as it tends to keep the bets a little lower and delivers an extra unit or two profit. The bet selection was pretty resistant in quite a bit of testing I performed over the last several days (longest losing streak = 5) and so thought I would give it a go for small stakes online.