Sim Smart

I have completed developing several programs to help me model baccarat methods.  My first level of programs analyzed existing baccarat shoes from the Zumma and Wizard of Odds collections.  My next level of programs enabled me to generate a practically unlimited number of realistic baccarat shoes with statistical characteristics and expectancies agreeing with theory.  (Reference: My Baccarat Shoe Factory.)  Because I now thoroughly understand my data set, I am able to confidently proceed to the next level of modeling baccarat methods with the assurance that the simulations realistically reflect real-world playing.

I have begun testing some simple methods as I continue building the programs to greater levels.  Ultimately, I will be able to simulate relatively complex baccarat methods such as Ellis C. Davis’ System 40, Mark Teruya’s Maverick/Maverick Ultimate, and any other baccarat method with mechanical rules.

Before the fun begins, I felt the need to first establish why I am doing what I am doing.  I’ve received some criticism regarding my efforts, which I’m sure were offered with the best of intentions.  I thought it would be wise for me to address them first.


What are you trying to prove?

I’m trying to prove something to myself first and foremost.  I believe it is important to have confidence that the methods I use have objectively demonstrable advantages.  Since no method can offer a mathematically derived player’s advantage, the only other way to determine whether a method is robust is to perform simulations of them over a substantial, realistic data set.

I am not trying to prove which methods are bad or good, or which ones are better or best.  I am simply trying to ascertain for myself whether the risks of using certain methods to play baccarat is justified by the expected rewards.

If a baccarat method is viable in the long run, simulations can objectively demonstrate it.  If it is not, simulations will reveal why not, and hopefully lead to new insights which will improve them.


Playing baccarat is more art than science, and over-analyzing the game won’t help you win.  (There’s no such thing as a consistently winning mechanical system.)

True, there is such a thing as analysis paralysis.

But without exhaustively understanding every nuance of the game, how can one hope to win in the long term?

The baccarat artist may think he is relying on his intuition to make a decision, but he is deluding himself if he thinks he’s not just guessing.  If he is lucky, he might win for a while, but the odds are against him, and he will lose long term.

The only reason the game is being offered at all by casinos is because its  designers have completely analyzed the game and made mathematically certain that the edge belongs to the House.  No art was invoked to create the game.  Likewise, no amount of artistry will ever help a player consistently win.  If there is anything certain about the game, it is this.

All gamblers who rely on “art” eventually lose.  The  only way to consistently win is to take advantage of exploitable, persistent deficiencies and/or statistical anomalies in the game, and the only way to discover those hidden faults is by exhaustive, critical analysis.

With respect to the statement, “There’s no such thing as a consistently winning mechanical system,” if a method is not objectively systematic and mechanical, then by definition, it is subjectively artsy, and thus no better than simply guessing, doomed to fail in the long run due to the House’s mathematical edge.  If there truly are no winning mechanical systems, then trying to beat the House is hopeless, because we cannot win in the long run.  If there truly are consistently winning methods, they must by definition be objectively systematic and mechanical.

In the end, if the verdict is that there is no consistent way to gain a definable edge, then the scientist’s analysis paralysis will thankfully save him the many dollars that the artist blindingly loses to the House.


Computer simulations will never adequately model the real-world, so they can’t be trusted.

It depends of what you’re trying to simulate.  If it is something very complex with many unknown variables, for instance modeling the weather in climatology, or the evolution of the universe in cosmology, or the movement of prices in economics, then, yes, simulations are inherently incomplete.  Even in very complex studies such as these where many assumptions must be made to fill in the gaps, a great deal of knowledge and a certain degree of accuracy can still come from good simulations.

Unlike the climate, the universe, or the economy, a game like baccarat is relatively simple to model.  The rules are definite and complete.  It is a closed system with a finite number of possibilities.  Baccarat is many orders of magnitude less complex than chess, and chess has been modeled successfully, so successfully, in fact, that chess computers routinely beat human grand masters at the game.  So, using a computer to model a baccarat method can be a very effective tool to study its long term profit potential.  If a method has definite rules which can be programmed, it can be computationally modeled.

Any type of instruction to do a certain thing when another thing happens can be programmed.  Even a “hunch” can be programmed, as long as you can identify what is the underlying reason for the hunch.  For example, when someone looks at the decisions of a particular shoe and “it just looks choppy,”  we can try to more objectively express this hunch in terms of a quantitative value.  One way might be to keep a running count of the instances of opposite decisions and repeat decisions, and then defining that “choppy” conditions exist when opposites outnumber repeats by a certain number.  Indeed, a computer can even be programmed to randomly guess, and its guesses will be mathematically more random than a human can guess, and possibly more effective!

Some people use computers to model the wrong way.  For example, some reverse engineer the data, a procedure also known as data fitting or curve fitting.  A would-be developer might extensively analyze the Zumma shoes to tailor fit methods which win the patterns present in those shoes.  This procedure may result in a spectacular profit graph, but it will only be valid for the specific data set from which the method was forged.  Try the method on a different group of shoes or in live play, and it will most likely fail miserably.  The same kind of data fitting is routinely performed by some who try to create trading methods for the financial markets.  They will take price data and optimize for the best, most impressive results.  Over-optimization creates pretty pictures but lousy methods. You can easily identify those who sell over-optimized methods from their stunningly positive winnings graphs.

If a method created independently from any specific data set is able to show consistent winnings when tested on any data set, it should be taken seriously.   A computer is able to very quickly, tirelessly, and perfectly execute a mechanical system over hundreds, thousands, and even millions of shoes, and those who consider the results gain a definite advantage over those who ignore them.


Real baccarat shoes in physical casinos are intentionally biased by shuffle control and have characteristics which significantly differ from what is theoretically expected in a random set.  So, you need to base your simulations on results from real casino shoes, not computer generated ones.

Possibly, but where is the objective proof?  Reason suggests the opposite is true (Shuffle Control: Why It’s Bad for the House and Beating Random), and I found no quantitative evidence of bias in the live shoes from Zumma (My Baccarat Shoe Factory).

Granted, I would also prefer to test using live shoes, but it would be extremely difficult to compile a database of 100,000 live shoes, let alone 10,000.  The best commercially available live shoe data is the 600 + 1000 from the Zumma books, and no statistical signature distinguishes them from realistically computer generated shoes using a virtual deck.  They share the same overall ratio of successive frequency of events, and the same average expectancies for the rate of Player, Banker, and Tie decisions.

For all intents and purposes, baccarat decisions whether computer generated from virtual shoes or dealt live from physical casinos appear statistically and practically identical.


My overall plan is to start by focusing on very simple methods to build my intuition and technology gradually.  The ultimate goal is to realistically simulate complex methods such as System 40 and Maverick/Maverick Ultimate.  By first studying and understanding the simpler parts on their own, I can gain a better appreciation for their roles and effects in the larger systems.

The purpose of all of this is to help me improve, build intuition, gain new insights, objectively analyze existing approaches, and hopefully develop more efficient and effective strategies.  I’m going be smart in properly performing and evaluating simulations to the best of my ability.  I will be wiser and hopefully richer in the end for my efforts.

So without further ado, let the smart sims begin!


13 Responses to “Sim Smart”

  1. Read your proposed testing & think it’s great! I often wondered why no one did a comprehensive test of “triggers,” or for example if you play TBL/OTBL & switch on 2 losses one way & 3 the other or vice versa, does it matter? Is one way better than the other?? Is there a best way? Now you can find out & thanks for sharing your journey.

    • My pleasure, L,

      Glad to hear the information is interesting to you.

      Yeah, you would think that those who sell systems would have done thorough tests such as these to first make sure their methods stand the test of time and analysis before ever trying to sell their ideas to the public. It’s kinda like if an car company sold a vehicle that they never fully tested and expected their customers to do all the safety and engineering tests to make sure the car is safe and reliable.

      I suppose there’s a more obvious possible reason, though: I have a hunch that the simulation studies will show what everyone already knows, but doesn’t want to believe: the House always wins.

  2. Frankly speaking , I don’t believe that the HOUSE always wins.

    If , only if , by any chance we can bet 30 times using Martingale without any table limit , I believe it is such an infinite win , isn’t it ? ( Someone said in the forum :world record only have 28 streaks ).

    Now , for that reason , I have a system designed by myself until now always give me good results on every land based casino.
    I’ve tried some times ( including some forum members ) to program a simulation for that. It happens that no one could , due to the complexity of its rules and variables.

    When you are interest in it , please give me a shot.

    • @perks

      With a Martingale, infinitely deep pockets, and no table limits, yes, you can claim to technically “beat” the game. Some of my simulations show such winning outcomes, but the drawdowns and highest bets are impractically steep.

      I emailed you privately with regards to your system.

    • Hi folks, I wandered upon this site while looking for pre-generated shoes to continue my testing. my hobby for 30 years has been studying probability and odds, while attempting to beat the house. believe me I am aware of the math, I am no fool. over the last 18 months I developed a system also much to complex for any sim I have found and I have tried most. I I have exhausted all available shoes and hand dealt thousands more with my wife. Over ten thousand shoes so far and it works, within posted limits, within reasonable time frame. The bankroll requirement is moderately high, but I have accumulated 76 x my stop loss amount, and reached my stop loss amount only once. I wish to discuss further with anyone interested, mostly i am trying to find more pre-generated shoes. thanks,

  3. Hi, my Name is Gabriel Rusu and I am writing to you from Germany, Munich.

    I am curious to learn more about your System

    Thanks in advance
    Best regards

    • Hi Gabriel,

      Glad to see you are well. Enjoy!


      • Hi Dave,

        are you also a baccarat System developper ?

        Here is a description of my tool

        Here are 7 Baccarat systems that are tailored for Baccarat players. Input fields for player and banker and tie. The systems operate the ECs to the so at Green set, red stop. Integrated indicator is stop and go, as well as progression computer (overlay) per click and eCart.

        Each system in the ECs specifies a punctuation mark on the respective chance from a certain voltage or its own strategy with green. With two consecutive losses, it shows red when the ECs and becomes green again if the two red losses are eliminated and a plus.

        This function is also switched off.

        When starting the software, remove the checkmark at “Double loss wait”.

        The systems are built according to the third law and tendency.

        Indicators and figures running simultaneously (Alyettische characters) as well as with the corresponding rapporteurs.

        In the Alyettischen figures there is also symmetric and asymmetric, in the Rapporteur of each top right digital display. (See description at Chevere I)

        Also, the eCart at the ECs will be charged at the same time.

        These tools are very helpful in determining sentence

        Looking for cantact with baccarat professionals worldwide.

        Best regards from Bavaria ( Octoberfest !)


      • Thanks, Gabriel, Yes, I’m a developer, too.

        I’ve developed my bacc programs mostly for personal analysis use.

      • Hi,

        perhaps you can help me : I am looking for These 2 Systems in pdf. Format :

        24 Karat Baccarat
        3 Dmensional baccarat

        I am also looking for a ceap Software tool in order to protect a small baccarat Software I am selling online.

        Thanks in advance.


      • I’ll look for those systems in PDF, and if I find them, I’ll send them to you.

  4. Good Evening Dave,

    I think that the most effective way to do this (modeling the game) is to throw bet selection out of the window and simply line bet a method of your choice against one of the lines. By following this path, your programming will be simpler as you only need to apply the betting rules for each decision and you will get a very real sense of how a method of play stands up against the true volatility of the game.

    Bet selection only creates the illusion that the player is changing the underlying volatility of the game. For example, I can line bet the banker in a player dominant shoe and lose 14 bets in a row. Or I could pause, run from table to table waiting for certain events and be unluckly enough to get the same result.

    Finally, it really doesn’t make a difference if shoes are computer generated or not, because ultimately the game is a series of win/loss patterns where 1 you win and 0 you lose. The only thing we should care about is what happens to our money that is, how fast or slowly do we go broke when we subject our cash to this volatility.


    Lee Evans

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: