IMO, more parameters we're inserting in the algorithms and greater will be the probability to get a poor prediction as the baccarat variables are so many that we risk to sink in the undetectable ocean where casinos take their huge profits.

For practical reasons, we've chosen to set up both algorithms in the simplest way and obviously 'training data' take care of the old 'average card distribution' that cannot be disregarded for long.
As already sayed, when an actual card distribution tends to deviate from the 'average', algorithms stop their action even if they have collected a temporary loss.

Good news is that when both algorithms are in action, in the vast majority of the times the positive clustering effect of one al. will overwhelm the other performing bad and not by a kind of 'opposite' way of considering things.

After all predictions are made upon a very restricted field of operation where the main goal is to  get all winnings along the shoe dealt.
Remaining situations, albeit producing more wins than losses at the end of the shoe, are very welcome but considered by the algorithms just as 'incidents'.

More later

as.

BTW, next week we'll talk about the 'overfitting' problem. 

as.

The algorithms are built after having tested thousands and thousands of real shoes coming from various sources:

- Manually shuffled shoes

- Shuffle Master Machine shoes

- Preshuffled shoes (typical of HS rooms)

- Different mix of the above procedures

Results and profitability are the same, there are no significant statistical values favoring one or another method of shuffling.

Therefore we have thought that the partial unrandomness or the nearly perfect randomness of card shuffles won't make a role in determining the excellent alg's prediction.
That tends to collide with our past hypothesis that only a defect of randomness could make a game beatable.

Since we are not so naive to think that we were up on something without a reason, we thought that the old key cards asymmetrical distribution makes an important role about the results, moreover reinforced by what we name as "average card distribution" where alg A put the most emphasis on.
At the relatively less likely occurences where the 'average' seems to fail privileging the outliers and despite of a huge betting selection, the alg A stops its action conceding room to the alg B, but this is just a back-up less important move as more shoes we'll play more consistent will be the probability to exploit our EV+.

For sure there will be better algorithms to beat the game, yet we have tested an unbelievable number of different possibilities and so far those are the best.

The decisive factor, in our opinion, was to get a possible long term 'optimal play' based by comparing expected math and statistical features with the actual results, so even density and rhythm of presentation are both valuable tools to instruct the algorithm to make a proper job.

As already sayed, an ironic aspect to consider is that a same pattern (commonly considered as a single or a double, etc) could be a losing or a winning spot in relationship of when the alg decides to act.

To make a simple counter intuitive example, alg A may easily suggest to bet P after PP or PPP, or to bet P after B as those are the slight more likely occurrences related to the actual distribution.   

Obviously and to make the already complicated concepts simpler and to implement the authors ideas, the algorithms consider BP results as 50/50 spread, thus having the expected same probability to appear.
So in the process and while betting P side, we're conceding a general (but volatile) -0.18% math edge to the house.
Actually the algorithms are capable to catch situations where P probability is way over than 50% whereas B side gets more difficulty to spot situations surpassing easily and safely the 51.3% profitability cutoff point.

But this is a well known baccarat problem while betting B side: Either we're rarely astoundingly right (8.6%) or absurdly consistently wrong (91.4%).

See you next week

as.

Derived roads vs algorithms

Everybody knows that a given plan performing bad at a given derived road very often will form opposite situations at one or both other roads, so enticing us to change the succession to be followed, with the hope that things keep staying in the 'good' territory.

Sometimes the 'trick' works and other times does not and of course most of the times the probability to succeed is 50%. So worthless.
 
The derived roads invention was a brilliant accomplishment made by some Macau colleagues in the 70s (there are some statistical features to exploit by playing them) but somewhat flawed from the start.

The main problem of the DRs is that they are geometrically produced like bricks forming walls of different height, so 'too much' affected by the actual card distribution without giving a proper  role to the decisive math features.

In fact, whereas natural difficult situations arise at both DRs and algorithms and for different reasons, DRs do not give us the luxury to rely upon a 'well calibrated and controllable' scenario, the paramount condition to set up serenely a profitable plan itlr.

In truth, each DR is capable to provide longer positive situations than our algorithms but with the fatal downside to make more probable long negative sequences to happen.
A thing that we must avoid at all costs.

Obviously the same problem applies to the Big Road but at least here we possibly get additional factors to rely upon (see 'codes' plan for example).

But the most interesting thing we've found is that DRs are providing 'symmetrical' events, in the sense that every road will whimsically present good or bad situations in relationship of the actual distribution without any link between the three lines, whereas alg A when seems to fail makes alg B to get a more normal 'course of action'.

Despite of being both algorithms built with the same math and actual distribution issues, the alg A always takes the lead over the alg B as this one is considered just a back-up (still very profitable) plan. 

More later

as. 

Think that our algorithms give a bighorn.sh.it about what math experts keep claiming, that's why they actually work.

We're clowns but knowing very well what Richard Von Mises, M. v. Smoluchowski, Konold, Nickerson, P. Revesz, Marigny de Grilleau and many others have written about randomness and statistical or gambling topics.

Oh well, math experts know better than them...

Thanks, thanks, thanks!!!!

as.

Consider every possible shoe's scenario and its probability to happen and you'll have the answer about how we have set up the algorithms.

Remember that we're dealing with a finite slight dependent proposition being asymmetrical at the vast majority of the times.

So "long" symmetrical scenarios are the exception and not the rule.

Moreover we have added a decisive factor in our r.w.'s by implementing strict math and statistical features that cannot disregard an average card distribution as any hand is not completely coming out 'out of blue' as a roulette spin.

In a word, nearly 90% of the times we'll possibly lose just for a less likely permutations issue and at the 10% remaining part we are forced to bear a quite unlikely card distribution that could be extremely good around 6% of the times or extremely bad 4% of the times.

Interestingly, when one algorithm seems to fall into the negative permutations issue or being prey of a strong negative deviated field, and we're not willing to wait for a more natural course of action (whether the shoe is still considered 'playable'), the other one performs so good that it's a child's joke to select the most profitable situations.

After all, both algorithms move around the same math and stats concepts, that is the relative unlikelyhood to get symmetrical results for long.

See you in a couple of days.

as. 

There's no way to make fictional bets appearing as genuine wagers at casinoscores for many reasons.
So we better talk about general thoughts.

 

Hi KFB, thanks for your interest!!


Since we have verified that the alg. provides a slight number of more Ws than Ls (after vig), we consider any single session just as a minuscule part of the whole picture, so putting the maximum effort toward not losing severely than winning a lot.
Moreover, when a shoe is not featuring our 'average' expectations, we'll raise the requisites of future bets up to the point where we won't bet a dime anymore.

Oppositely, when a shoe seems to fit 'too much' to the average distribution, we may start to bet more spots but the betting amount more or less tends to stay at the same level.

So it's impossible to make us to lose a buy-in at a single shoe as the 'buy-in' concept simply doesn't exist. Let's assume we can rely upon a virtual'infinite' buy-in.   
 
So far the alg is not so sophisticated to cover all the situations when to stop (or prolong) our betting in relationship of the previous wins/losses register as it simply tries to spot each time what should be more likely to happen after a series of pattern results we've inserted on it.

Obviously the RTM effect makes an important role partially balanced by the rare most deviated card distributions denying it.
When in doubt and unless the rare strong deviations are coming at our favor (because are 'due'), stopping the bet is the best move to take, IMO.

Despite of having ascertained that the algorithm works wonderfully, we have learnt at our expenses one more time that most part of negpro plans are just making the casinos' fortune, that is that variance cannot be controlled by varying the betting amounts as even a verified edge could need a lot of time to show up.

See you later and again I'll try to make fictional bets at casinoscores site.

as.

First thanks for your interest.
I've tried again to forecast results at casinoscores site but it's a very difficult task to achieve, mainly as sometimes results are not properly registrated and then for a lack of proper betting time.
All this despite of my algorithms are perfectly working at my pc (so differently to a live world where we'll have to make the job manually) but even then they need some time to insert the actual results, not mentioning the problem to post in the site my forecasting ASAP.
(An additional addendum is that we don't like playing online for several reasons). 

Anyway you can be so sure about the edge I'm talking about that you can safely put at stake whatever you have on your name as no negative variance in the world could break our advantage.

BTW, it's very likely that among the best baccarat players in the world most of them are here.
So you're not wasting your precious time reading these site's pages.

as.

Good choice, KFB! 


Are we sure to play with an edge?

In every asymmetrical proposition (itlr a side is advantaged over the other one, say our plan) it's natural to expect a W>L ratio, but there are more important tools to consider, that is the winning streaks shape and their distribution.

Obviously to exploit an edge, itlr W clusters must be superior than W isolated events meaning that very often it's not the actual W streaks lenght to shift things in our favor.
That's because the variance will put a strong obstacle to expect homogeneous situations featuring all of the time a greater amount of W clusters than W isolated situations.

In fact, W isolated events may easily come out clustered, so if you have a plan that went through a 4 or 5 or even 6 W isolated series without reaching superior values you can safely assume you were just lucky.

Even worse are the L sequences that cannot be 'controlled' by their lenght as under normal circumstances the edge remains small.

The trick to raise the probability of success is just a ploy to more likely catch the W clusters, well knowing that it can be valid only when fitting an 'average card distribution'.

Algorithm action


The algorithm is set up by two levels:

a) a mechanical classification about the probability of getting this or that by math features applied to a coin flip proposition;

b) an evaluation of the above results by statistical standards more or less deviating from an 'average' card distribution so enticing or not the betting.
I've utilized the word 'evaluation' (a topic already touched in a previous post), as the actual action must be calibrated upon the goal the players aim for: there are people who wants to play a quite number of hands (I hope mainly for comp reasons) and there are players who want to be right at very selected situations by wagering huge sums.

At any rate, the algorithm gets the best of it no matter what as it was instructed to take care of average card distributions, frequently stopping its action when things tend to not conform to those distributions.

Another important and counterintuitive issue is that the a) classification provide many shoes featuring an overall unit loss, so enhancing the concept that very often it's not important what we bet but 'when' we bet.
Obviously the (over) 'balanced' part is made of shoes providing ALL wins, it's just a matter of (few) time we'll exploit our edge.

The backup algorithm

Say that for some reasons the average card distribution is disregarded for long or for the entire shoe(s), so tossing into the trash our algorithm.

Besides the fact that such distributions will likely make the fortune of recreational or gambling players (so the almost entire baccarat community), we still have the tool to get our profits.

It's sufficient to postpone the algorithm A by a 1 factor, so building an algorithm B getting a different scheduled a) pace but a same b) rhythm.

Unlike the derived roads where the same Big Road will contemporarily produce quite diverse patterns at the three lines, algorithm A and algorithm B will produce the almost same number of expected spots but just by different permutations.

Since itlr the edge is mathematically insensitive of the permutations issue (as long as the card distributions fall into the average field) but relatively susceptible of short-intermediate variance, we may find reasons to put in action the algorithm B when the algorithm A stalls for long without suggesting any bet.

Now and in no way we're playing a kind of 'opposite' plan, we're just playing the probabilities under a 50% different pace plan.

So the only real problem to face is not about the probability to win but to properly set up manually the algorithms as it's quite easy to make mistakes, even if we just use the main algorithm A.

See you next week

as.

Another brilliant comment made by one who had shown us here an endless piles of Benjamins.

The challenge with the house is asymmetrical from the start, but don't make such battle as "too much asymmetrical", meaning that we don't want to risk a lot in order to win a little.

Let casinos fear we're risking a finite X bankroll to win a virtually infinite amount of money and not luring us to stop the action after having collected a miserable profit.

Defending the bankroll is of utmost importance but "to win the tournament" sooner or later we must put in jeopardy a fair or a large part of the money won whether the proper conditions are met.
I say 'fair or large' as most of the times we're moving around tiny profits. 

Let Steve Wynn keep thinking that "the only way to win at a casino is to own one": he doesn't know and obviously he couldn't care less about how much money we've extracted from his Wynn and Encore premises.

BTW I suggest the reading of Bill Walters freshly released book: "Gambler: Secrets from a life at risk".
You won't be disappointed.

as.

Average algorithm's betting frequency

Since one or more same specific patterns might be good or bad in relationship of 'when' we're classifying them, our algorithm makes the best efforts to 'align' them around the more probable average card distribution giving a lesser bighorn.sh.it about single hands, thus considering the process as a 'whole'.

Obviously such process will produce asymmetrical results, meaning that itlr (but even in the intermediate terms and in the vast majority of short terms) a slight greater amount of positive scenarios will overcome the negative counterpart so getting us an edge.

Hold on.

Mathematicians and gambling experts will say to you that every single hand is burdened by the negative HE no matter how's sophisticated our strategy.
But they've made the fatal error of considering baccarat productions as a kind of an independent undetectable world comparable to a coin flip succession, giving an obvious less negative role to the B side math propensity.

The average situations where B side is really advantaged and by how much were only presented by a keen baccarat expert and not by a mathematician considering them worthless.

The algorithm was and is proven to be wiser than those math pundits as it does consider an optimal play extracted after thousands and thousands of live shoes dealt, not after rattlesnake.s.h.it simulations made with some softwares.

At the end the algorithm considers B=P, because it's more important to be right at a series of probable hands belonging to a stereotypical world than to catch the improbable 'astoundingly' right situations (with a fair degree of error) at very few occasions.

Betting frequency per shoe

1) The worst risk/reward ratio (EV+, but getting huge volatility) considers an average amount of 15 bets per shoe. It's not the mere outcomes issue that matters, just the permutations issue.
With this ratio the algorithm tries to get all wins.

Yet, by adopting this average 15 bets per shoe ratio, we'll deadly sure to encounter very soon an all winning sequence at the same shoe, obviously by backing-up the first losing step (when it happens, that is almost a slight less than half of the times occurrence).
Say that winning 8-9 consecutive two-layered bets per shoe is the ideal world we should aim for.

2) Permutations issue.
Positive situations are more likely to come out clustered and losing sequences isolated, yet long positive situations might be intertwined by clustered (albeit short) losing successions and so on.
This is one of the main situation to look for clustered positive spots, regardless of their lenght.

3) Clustered positive sequences

They are more likely to show up than the isolated counterpart, yet any isolated winning spot will be more likely to be followed by a clustered positive succession of any lenght up that waiting one or a couple of consecutive isolated winning spots constitutes the best trigger to aim for.
More often than not, 'long' isolated winning sequences are interwined by isolated losing spots.

4) Isolated positive spots followed by clustered losing situations

It's the only very bad situation to take into account unless we properly consider the #3 point.
On average they happen less than one time out of 5 shoes dealt, in the meanwhile the more likely occurences that gave us a profit totally or partially cover such unfortunate but inevitable occurences.

The beauty of the algorithm action is that it's virtually impossible to get a back-to-back sequence of such kind happening at the same shoe.
A thing that it's relatively frequent at BR, BYB, SR and CR by applying a corresponding bet selection.

5) Isolated losing sequences

Completely unplayable unless intertwined by consecutive isolated positive spots.
It's the proof that the algorithm is not taking advantage of positive patterns caught by chance as  losing spots tend to be more clustered than following a natural 'probability of success' line, what really happens at the common roads.

See you in a couple of days.

as.

Winning spots are not extracted from 'creative' strategies just coming out of probabilities


Each shoe distribution is not affected by past shoes distributions as the probabilities we're looking for are determined by the constant 'average card distribution'.
The algorithm simply takes care of the most likely distribution lines every shoe must take along the way.

The interesting part of the algorithm action is that same patterns could be positive or negative in relationship of its actual rhythm and of course this rhythm is mechanically scheduled and based upon the average shoe.
In a word, this algorithm works toward a kind of 'results alignment' that must happen at various levels of probability and not toward precise patterns that are always considered as 'good' or as 'bad'.

Obviously the 'positive' attempts of alignment are slight greater than the 'negative' attempts and that's where the edge comes from.

The beauty of this plan, besides of its verified edge, is that the algorithm starts its classification action by the very beginning of the shoe where some cards are 'randomly' burnt before the results start flowing.
Thus even if the casinos know what the algorithm is really looking for, it's virtually impossible for them to deal results not belonging to the 'more likely' probability aspects for long.

Another important aspect involves the possibility to set up the algorithm among different risk/reward categories so dictating the bet selection frequency and, less important, the betting amounts utilized.

See you later

as.

Hi KFB, thanks for your reply!

I'm a firm believer that much of our war with the casino is won by what the bettor is doing between wager #1 and wager #2. For we shall see the same length of Ws(and Ls) streaks, regardless of the size of our bet.

I can't agree more on that! 

as.

Get rid of your gambling attitude before thinking to play baccarat successfully

Whereas at poker an educated gambling attitude may transform a good player into a great player, at baccarat gambling must be completely tossed through the toilet with no exceptions.
We've lost a lot of money before realizing that and best baccarat players we know did the same thing.

The edge exists but under normal circumstances will be quite small. Sometimes the edge merges with the normal positive distribution, so luring us to bet too many hands or to improperly increase a lot the betting amount.
Think that when the edge suffers the negative variance, the natural negative distribution might add up then making a stronger impact over our results.

For example, when we'd unwisely think to suddendly double our standard bet, we should understand that never ever our edge will be doubled.
Differently than black jack where our edge is mathematically ascertained (and proportionally related to the actual count), at baccarat we are just approximating our winning probability by statistical features needing some time to show up.

The algorithm cannot give a lesser damn about how 'we're feeling lucky', the way was devised just suggests optimal (imperfect) choices considered after having tested large LIVE samples.
It hopes for the best but expecting the worst and itlr (but even at short/intermediate terms) best < worst for the HE.

IMO best MM to exploit a small edge is to selectively wager huge amounts at very few spots (a kind of 'Bold strategy') or to slowly increase moderate amounts by tiny percentages at supposedly more likely positive occurences (KFB made brilliant examples about that).
And Alrelax pointed out that it's wiser to 'press' the bets at the earlier stages of a positive pattern. 

Algorithm takes care of what should be the most likely B/P occurence, the risk/award ratio is up to us.

See you next week, I'll present you how's the algorithm betting frequency per shoe.

as.