QuoteIt is a simple thing to probe the gambling session to see if a string of wins is forming

Quotetracking or measuring "effectiveness" and adjusting, tuning, etc. until you are in-phase with what is transpiring can lead to a significant and real advantage.  I believe you

This is exactly what I and a few others do, with great effect for me atleast, I've played probably 2 to 3 thousand short sessions doing simple formation watching, increasing bet size when the time is 'right' my session targets are 12 to 15 units PLEASE don't confuse this with hit n run, I don't wait for an event to happen, I play what's coming out, I play all EC groups and as Bayes may agree, I am only touching the tip of the iceberg with what I do, I also only play on BV NZ the amount of units I aim for is set to cover the 10% they take each session.

If you can work with random then it isn't too difficult to stay ahead.

Superman, this approach might be of use. Once you master dozens & columns you can move on to the opportunities of collisions, 16 numbers. Their odds are very close to the even chance bets. I'm finding them very interesting.

Mucho fascinating! I confess I really like your term "effectiveness track". It resonates well with me, and I suspect fits well with the way our brains process incoming data too. I remember seeing a doco that described how we recognize  faces, often from a great distance and with great accuracy - apparently we spot a bunch of "differences from the norm or average" and these differences are what defines the individual for us, without any conscious effort at all - it all just happens automatically. So our brains are already wired to measure and refine and make judgement. I guess all you have to do is hook into that, which means personal disciplines to make sure you treat the incoming data with the right part of the brain, leaving all the other stuff to the side.  In other words, you have to "see it the right way".


I'm also fascinated that you are attempting to turn this into a rule-based strategy so that it CAN be simulated and therefore tested - a bold and ambitious venture I dare say!  But where would we be if we weren't willing to tackle the difficult tasks in life - I wish you much success and look forward to hearing how you get on.  As a programmer myself I know how much effort it takes.


Bryan

Perhaps it is worth mentioning at this point a little of my own background, because we do see a lot of references to programming and languages and tools and so on, and I think, to a degree, pretty much anybody can call themselves a programmer if they do a little work with excel or the like.  The advent (and proliferation) of tools such as excel with an accessible macro capability puts very powerful tools at the hands of the layman and this is a good thing in my opinion. Likewise other tools such as the roulette xtreme (I think it is called) which provides an environment for testing strategies that anybody can use brings the power of computer testing to the world - again, a good thing. 


But I'm old school.  I spent my career working on mainframes, DEC machines to be specific, and coded in what is known as 3GL's. I hate PC's (although I am forced to work with them) and generally speaking I find the quality of apps (what we used to call programs) pretty poor. In my opinion Bill Gates has a lot to answer for!  DOS was never the right choice for an operating system!  But, here I go again - raving!  Must be getting old or something.  Anyway, I am fortunate enough to own my very own VAX machine which is a brilliant number cruncher and I use this to write my own programs and run the simulations.  So I will not be very helpful in PC matters - sorry. 


However, I am trained in programming the old way - that is, we paid attention to efficiencies both in terms of processing and memory usage. We understood how to create readable modular code, reuseability, etc. and how to structure a logical pathway, with error handling and correction.  We understood boundary conditions, interfaces to devices and people, and how to timeshare resources.  We solved the problems of file sharing and locking, how to detect and avoid deadlocks and multiple-user file accessing - long before the PC pundits came along (who then, in their arrogance, attempted to reinvent the wheel thereby imposing pain and suffering on the poor user all over again, including dates in the wrong format, wrong size paper, numerical rounding that doesn't work, file sharing that behaves weirdly, and so on and so on that persists to this day).  All this was solved and fixed 25 years ago in my world.  Still, PC's do make good gaming machines!


So I might be able to help with a logic problem or structure suggestions or whatever, but I am not likely to be helpful in language specific issues - unless it is VAX Basic or VMS DCL.  That's me in a nutshell - slowly becoming a dinosaur, and resisting all the way! 


Bryan

I've been working on creating foundations to stand on. I want to create functionality that allows the software to ask questions like I would ask the question. I instantly relate to dozen 1, numbers 1 to 12. Throughout the software this will be an object that has the ability to know if one of its numbers hit on the current spin. It will pass on an ability to know if it was part of an occurring pattern. It will know if it's a sleeper. It will know if it belongs to any global effects that exist long term & short term. In other words I'm going to create a Roulette playing genius mind. Software can evaluate so many functions that hundreds of conditions can be considered for each number and grouping. I doubt that a person could find the truly abstract and elegant permutations that randomness has to offer. Thank goodness there are only 37/38 numbers and only 12 options for outside bets to consider for pattern recognition. Yes, this is a huge task.

Welcome Sqzbox



Looking forward to much that you have to offer. The resources within this Forum are formidable. Perhaps your musicianship will enable an 'effective' inter-action of right and left brain activity -lol.


Have you thought about a more 'efficient bet' in your studies?


Best XXVV

Thank you Sir X  !  So - an "efficient bet" eh! Are you poking the hornets nest?  I tell you what - I'll trot over to the private forum and post a little challenge.  Maybe we can get a rewarding discussion going.   I'll use the private forum because I really don't want to waste my time with naysayers and general negativity - I'm happy to be told that my ideas are stuff, but only if the responses are well-reasoned and well presented.  I am a great believer in Aut tace aut loquere meliora silencio.


Gizmotron - I love your approach.  You are basically, if I understand you correctly, making each of the possible 128 legal bets an object in its own right. This should be do-able, and, I think, a very smart way to do it. In the early days of research into AI I recall a study done on attempting to simulate the way a school of fish moves.  Computer power was not what it is today and the scientists involved found that, while the simulation worked, it required huge amounts of resource and was very slow in real-time when run.  So they changed their approach and assigned each fish in the school (this is a virtual school of course) its own process (or object in today's lingua). Each process had 2 simple rules to follow - always keep moving and swim as close to your neighbour as you can without touching. This worked astonishingly well apparently and performance was greatly enhanced. Anyway, I believe that your approach is eminently sensible and would be easily extensible, once perfected, into other possible bets such as neighbours, tiers, wheel sectors, and the like.  Good luck!


regards
Bryan

Sqzbox, I was just thinking how this opens the door to extensibility. If every programmer has a weakness then mine is namespace overloading and recursion. But there is a beautiful thing I like about these kind of features. I can break the rules and store the attributes in a line delimited kind of a text object. Each object can have an unlimited number of easy to read lines. That way I won't go crazy with huge lists in hierarchies and/or arrays. Function calls, with their attributes can be position based or XML based. I built a non-traditional parallel linear parser for an experimental markup language experiment. I can use it to power  a read/write process for each objects list of characteristics and attributes. If I wanted to, and I really don't, at this time, I could make these objects available for each spin. These would be mind numbing data sheets, if math types wanted to see them that is. It's actually coming together.

Gizmo - are you testing me?    While I am a great believer of KISS there is no doubt that at times a complex problem requires a complex solution.  What you are attempting is really, it seems to me, to develop an expert system.  There are commercially available packages that allow for the development and running of such things - have you thought about utilising one of these?  Or even, tackling the problem in this way?  Developing an expert system in the defined sense of the term does require a change in approach - "knowledge" becomes key, rather than "information". However, defining that knowledge can be tricky.  I guess by going through the development process that you are, helps clarify all this in your mind.  I know that for myself often the "doing" helps clarify the requirement and the solution, and sometimes starting again from scratch with this enhanced understanding can be quite beneficial.  It sounds like the main issue will be the proper definition of the objects properties and doing this in the way you suggest does allow for easy adjustment - which has to be good right?  But if this is truly the main issue, then perhaps your time is better spent there and building this knowledge into an already existing expert system engine rather than doing all that yourself.  Just a thought.

I spent yesterday canabalising Sims that were developed over the past few years. Many of them are good at extracting sleeping dozens from the stream of spins for an example. I'm an expert at using pull parsing as a way of extracting & populating structured data. I know how to think your way through reading randomness, my way, and using it to manipulate the effectiveness track. If I were to attempt to use another method I would get lost in the demands of the intelligent system. I don't want to use cookie cutter 4th or 5th generation solutions. I will develop this much faster if I just solve actual thinking issues by making it the same process as human association and visual dexterity. You see I know this topic. I play this in real time using a pen and index card at a live casino. All I have to do is task the computer to see my list of characteristics in the flow of spins, select the best occurring trend, check it against the effectiveness track, and apply the next bet. There is no way I would want to cram all that into a bunch of classes in a language that I probably detest. It would be like using excelScript. I'll show an example of a structured data document. Maybe you will see how easy it will be to see what I mean?

Example of extensible data structure:
// -- this could be held in a global variable

// -- contains object data set for the dozens set
// -- used for best trend selection

<dozSetData>
  // -- 30 spins deep
  <dozStream deep="30">3,13,20,12,34,0,5,12,16,...etc</dozStream>
 
  <dozSleeper dozOne="3" dozTwo="6" dozThree="2">2</dozSleeper>
  <dozSingles numSingles="3">yes</dozSingles>
 
  // -- global effect sleepers
  <dozGlobalSleeperGroup occur="yes">5</dozGlobalSleeperGroup>
  <dozGlobalSleeperD1>6</dozGlobalSleeperD1>
  <dozGlobalSleeperD2>4</dozGlobalSleeperD2>
  <dozGlobalSleeperD3>5</dozGlobalSleeperD3>
 
  // -- global effect singles
  <dozGlobalSinglesGroup occur="yes">5</dozGlobalSinglesGroup>
  <dozGlobalSingleD1>2</dozGlobalSingleD1>
  <dozGlobalSingleD2>7</dozGlobalSingleD2>
  <dozGlobalSingleD3>4</dozGlobalSingleD3>
 
</dozSetData>



// -- contains object data set for the dozens set
// -- used for best trend selection

<dozSetData>
  // -- 30 spins deep
  <dozStream deep="30">3,13,20,12,34,0,5,12,16,...etc</dozStream>
 
  <dozSleeper dozOne="3" dozTwo="6" dozThree="2">2</dozSleeper>
  <dozSingles numSingles="3">yes</dozSingles>
 
  // -- global effect sleepers
  <dozGlobalSleeperGroup occur="yes">5</dozGlobalSleeperGroup>
  <dozGlobalSleeperD1>6</dozGlobalSleeperD1>
  <dozGlobalSleeperD2>4</dozGlobalSleeperD2>
  <dozGlobalSleeperD3>5</dozGlobalSleeperD3>
 
  // -- global effect singles
  <dozGlobalSinglesGroup occur="yes">5</dozGlobalSinglesGroup>
  <dozGlobalSingleD1>2</dozGlobalSingleD1>
  <dozGlobalSingleD2>7</dozGlobalSingleD2>
  <dozGlobalSingleD3>4</dozGlobalSingleD3>
 
</dozSetData>



Fair enough.  I suppose, when you get right down to it, it is really just a bunch of rules anyway.

Quote from: sqzbox on January 02, 2013, 05:27:57 AM
Fair enough.  I suppose, when you get right down to it, it is really just a bunch of rules anyway.

It is. It even includes a layer for subjective qualities. I'm including a layer for percentage of perfection and a layer for difficulty of session. It will compare these traits for the best qualifiers. In the end it is nothing more than a complicated set of rules. An agile set at that. Everything I have experienced in playing says this won't work. If I can catch a moving target with a simple computer sim then that will be amazing.

As fascinating as this is, and it is, isn't this getting a little off topic..? This was made a sticky by esoito to provide good advice for the inexperienced newcomers, it was intended to enlighten them of the pitfalls and to try to steer them away from losing their first bankroll all in one night..??

TC -- you're right!

But I was reluctant to say anything whilst there was so much of interest being posted here.

And it's a good example to the newcomers of the quality posts available on the forum.

Eventually the thread will run out of steam (most do in the end!)  and then Victor or I can consider splitting it.