1. Welcome to the #1 Gambling Community with the best minds across the entire gambling spectrum. REGISTER NOW!
  2. Have a gambling question?

    Post it here and our gambling experts will answer it!
    Dismiss Notice
  3. Discussions in this section are assumed to be EV- as they are outside of the Advantage Play section. For EV+ discussions, please visit the Advantage Play sub-forum.
    Dismiss Notice

Baccarat Follow the Last anyone?

Discussion in 'Baccarat Forum' started by fathead, Sep 7, 2022.

  1. TwoUp

    TwoUp Well-Known Member

    Joined:
    Nov 30, 2020
    Likes:
    321
    Location:
    Australia
    His code would not be that useful. The tools available now make most things that used to be very specialised much more approachable.

    But it would be possible to setup a project on say Google Colab and anyone could use it.

    What would be most useful is his training dataset and labels, the type of neural network (CNN, RNN or ANN) and the topology. I'm guessing he is probably using either an RNN or a CNN.

    An approachable video series on how the simplest artificial neural networks or ANNs which are also called a perceptatrons, actually work:





    As you will see there is a lot of computation involved even with the simplest of examples.

    Thanks to modern GPU's and a thriving ecosystem we have access to powerful tools at a fraction of the price that was available just a few years ago.
     
    Last edited: Sep 13, 2022
    JacobBlaze, Punkcity and Jimske like this.
  2. TwoUp

    TwoUp Well-Known Member

    Joined:
    Nov 30, 2020
    Likes:
    321
    Location:
    Australia
    One man's export is another man's import.

    For automated betting online, yes that is possible, and there are certain tools I would use for that. But judging by RZ's experience you have to move around and not overstay your welcome and eventually the options dry up.

    Was the same with sport betting, they don't suffer winners, even the so called "fair" betting exchange resorts to front running, delaying order execution, limiting bet size, and running dark pools against their exchange participants and then take over 40% of profits that you do make with cunty premium charges. Imagine that, owning a business that has zero downside risk on a bet but taxes almost half the profits when you're a winner.
     
    Last edited: Sep 13, 2022
    JacobBlaze and Punkcity like this.
  3. Punkcity

    Punkcity Well-Known Member

    Joined:
    Mar 31, 2020
    Likes:
    1,037
    Occupation:
    Sanctamonious bullshit baccarat advisor
    Location:
    In someone’s head MATE. Skipptophia.
    Excellent post thanks
     
  4. Punkcity

    Punkcity Well-Known Member

    Joined:
    Mar 31, 2020
    Likes:
    1,037
    Occupation:
    Sanctamonious bullshit baccarat advisor
    Location:
    In someone’s head MATE. Skipptophia.
    Totally concur
     
  5. Rz10000

    Rz10000 Active Member

    Joined:
    Sep 8, 2022
    Likes:
    64
    Location:
    Pennsylvania USA
    Thank you for posting the info. It will give some members a better insight into the scale of what it take for a machine to do simple functions.

    You are correct in that I can port the software onto a public computer platform. But I do emphasize the word public which makes my particular development available to the world. Any cloud based system is not private. I would have to build safeguards into the development stage. I do not need any of that when the machine is in-house. Also I employ a math built in a multi-dimensional world, at this point in time I am not publishing that particular coding. I can encrypt the coding and let it reside as a black box within the software.
    The rest is standard vector math and old time stochastic ( what everyone used for stock market charting from the 90's) which is the learning stage. Nothing really special there (just number crunching). Also I have to look at how to control the application of the design. If everyone has it and some use it to just crank out max profits what will happen to the games in the online casinos.

    Will our choice only be what slot machine do you want to play today....or maybe KENO!

    I never had any intention of producing a product for sale. The successful sale of it would create its own demise.
    This started out as a presentation of a different approach and has gathered quite a lot of conversation, which is very good. That is how ideas are created.
     
    JacobBlaze, TwoUp and Nathan Detroit like this.
  6. Nathan Detroit

    Nathan Detroit Well-Known Member Founding Member

    Joined:
    Dec 25, 2014
    Likes:
    2,020
    The length of a session is being determined by the win goal or what ?
     
  7. Rz10000

    Rz10000 Active Member

    Joined:
    Sep 8, 2022
    Likes:
    64
    Location:
    Pennsylvania USA
    It is a strictly arbitrary as when to stop, other than each bet would have to end in a profit. Eg. if I am up $650 and I place a $25 next play and it looses then the system will run until the profit total is at least $675. That could take 2 or more hands. That is what sets the stopping points. I set a minimum per day at 8% but that does not mean in cannot make more.
     
    JacobBlaze likes this.

  8. soxfan

    soxfan Well-Known Member

    Joined:
    Dec 29, 2016
    Likes:
    711
    Location:
    FrozenTundra
    Well since yer baccarats patterns cracking AI can capture 8% compounded profits daily and never lose a session I think you should call it the john-O 1000, hey hey!!!!
     
    Nathan Detroit likes this.
  9. Nathan Detroit

    Nathan Detroit Well-Known Member Founding Member

    Joined:
    Dec 25, 2014
    Likes:
    2,020
    John-O is no consistent winner and AI is by no means the Holy Grail .
     
  10. Nathan Detroit

    Nathan Detroit Well-Known Member Founding Member

    Joined:
    Dec 25, 2014
    Likes:
    2,020
    soxfan is more in the range of accepted reality . NO gimmicks at all.
     
  11. TwoUp

    TwoUp Well-Known Member

    Joined:
    Nov 30, 2020
    Likes:
    321
    Location:
    Australia
    I'm not suggesting giving everyone a golden goose or a product, just using modern tools because hardware eventually fails and what you are currently doing is constrained by the methods and tools you had available at the time.

    Definitely check out Google Colab. Nothing is public unless you want it to be and you control who you provide access to.

    Also checkout things like PyTorch and the continuous learning modules like Avalanche. You can use them all on Jupyter notebooks on Google Colab, Multidimensional calculations are a one liner now and parallelized over as much resource as you need.

    Here are some great resources to explore if you're interested in the modern approach and techniques where some of the most recent research into continuous learning is taking place:
    https://github.com/ContinualAI/colab
     
    JacobBlaze likes this.
  12. TwoUp

    TwoUp Well-Known Member

    Joined:
    Nov 30, 2020
    Likes:
    321
    Location:
    Australia
    All except that it is, and you use AI every day and probably don't even realise it.

    Have you ever taken a photo on a modern camera or smart phone? AI makes the noise go away and optimises the focus, contrast and lighting to get the best out of the hardware using trillions of calculations.

    Ever unlocked your phone with your finger or thumb or facial recognition.

    Ever asked Siri or Google to set an alarm or call someone?

    Ever touched the screen of a phone? AI is used to understand intent and guestures.

    Ever used a credit card whose transactions are monitored for fraud using AI?

    Ever used satnav, google or apple maps? Again AI is used to work out the best route based on current conditions.

    Ever walked down the street, a shopping mall or into a casino? Ai is used for surveillance to protect you from theives, and other persons of interest.

    Ever bought something online? AI is used to calculate and optimise the shipping rates (logistics).

    AI is working behind the scenes, it is already the holy grail.
     
    JacobBlaze and Jimske like this.
  13. Nathan Detroit

    Nathan Detroit Well-Known Member Founding Member

    Joined:
    Dec 25, 2014
    Likes:
    2,020
    Another day for some cats to go ape about . Thats all for this gimmick which it is is . Just a gimmick .



    Tje " A I " gimmick . The built in loss chaser .
     
  14. Nathan Detroit

    Nathan Detroit Well-Known Member Founding Member

    Joined:
    Dec 25, 2014
    Likes:
    2,020
    Disdain for the word " recovery "
     

  15. JacobBlaze

    JacobBlaze Active Member

    Joined:
    Jul 12, 2018
    Likes:
    164
    Location:
    Vegas
    Thanks for all the links and Info, but my brain will probably blow up when I try to decipher this lol
     
    TwoUp likes this.
  16. JacobBlaze

    JacobBlaze Active Member

    Joined:
    Jul 12, 2018
    Likes:
    164
    Location:
    Vegas
    Stochastic... Wow
     
  17. Punkcity

    Punkcity Well-Known Member

    Joined:
    Mar 31, 2020
    Likes:
    1,037
    Occupation:
    Sanctamonious bullshit baccarat advisor
    Location:
    In someone’s head MATE. Skipptophia.
    So that is what a snow flurry in a desert looks like.
    Dry
    Very dry humour lol.
     

Share This Page