Welcome Guest
Log In | Register )
You last visited December 8, 2016, 6:54 pm
All times shown are
Eastern Time (GMT-5:00)

G.A.T. Engine 2.0

Topic closed. 53 replies. Last post 5 years ago by RJOh.

Page 4 of 4
PrintE-mailLink
LANTERN's avatar - kilroy 28_173_reasonably_small.jpg
Tx
United States
Member #4570
May 4, 2004
5180 Posts
Offline
Posted: December 22, 2011, 7:18 pm - IP Logged

Think of a test run over 100 tested draws. Lets assume we have 20 hits produced, that means 20% success (our X). Then 100/X = 100/20 = 5 draws on average.

It would be ideal to have hits produced exactly every 5 draws but this is not the case generally. Some hits might be consecutive, some at 2 or 3 or 4 draws after a hit. Some can be at 6 or 7 or 8 draws after a hit. The least variation is the best possible to that. So, if we expect a hit in the range of 1-8 (can be observed by the red line at the graph), on average we expect a hit on 100/X = 5 draws. It might come sooner, might a bit later, so since we don't know when exactly this will occur, we have the 100/X rule on average. Might be good to allow 1-2 more draws for that hit, so to go up to 6-7 draws in this example. The best guide is always the red line in the graph.

AT

That was a very good explanation and what I needed to understand it.

On hit patterns for filters, there is always a skips section at which from 1 to X skips 50% of the hits happen at, past that point it might be more unreliable or what ever the word is, as a particular pattern can stay very much over-due for-ever it seems like sometimes, some people play almost or max overdue patterns, I like best the 1 to X 50% skip hit gap, as it might be more predictable and reliable.

Also, there is or there might be one or more skips sections of higher density hits along the skips progression line and a person should try to monitor or take note of that.

By that I mean that maybe more often hits happen at or near particular skips, but only looking at the long term stats of skips can tell for sure.

Thanks very very much!

BibleOnline  ParishesOnline  ChristianRadioOnline   MassOnline   Mass

"Ten measures of beauty descended to the world, nine were taken by Jerusalem."

    RJOh's avatar - chipmunk
    mid-Ohio
    United States
    Member #9
    March 24, 2001
    19829 Posts
    Online
    Posted: December 23, 2011, 9:37 am - IP Logged

    True and fair enough. Any tested draw which is to be predicted (and all following it) is of course not known prior the prediction of that draw and the statistics generated simply say how the engine managed in predicting that "unknown draw" to the engine till that point. There would be no other way to make that hits comparison test, so to display statistics if the tested draw isn't there for that comparison to be made! From this point of view, i am not afraid or being fraud about my claims or the hits displayed, since this is exactly the logic how a predictor should operate and definitely GAT does not show plausible hits; the results are a real prediction test. If that was the case, then my existing customers will rebel at the very next draw seeing an incosistency of the hits the engine indicates it predicted before and after the draw. So even if what you say makes sense, it doesn't make sense to do simply because I'd be full of lawsuits for a fraud product by all my existing users, which means in 1-2 days top from today. I believe you understand that I am not so stupid to make anything like that, just to tease people, so you can trust what you see is exactly what you get in GAT.

     

    Garyo, GAT gives any requested amount. If you pick e.g. 10 numbers, GAT will give you a break down on what to expect from that set of numbers in terms of correct numbers predicted. This is the whole point in GAT. It is up to you how to use these 10 numbers. So you base your strategy based on your budget and ask from GAT how many numbers you want from it. It will give you the most promising set of whichever amount of numbers requested, along with statistics following that prediction. The compatibility is needed in some setups because GAT needs to write a small file at its installation folder at every run and also to save the history draws created at the installation folder. If you don't allow the compatibility, then the OS blocks these operations. It is a general "problem" of these OS in their attempt to protect their systems from trojans viruses etc, not a particular problem with GAT. I prefer that so everything needed is in one place instead of here and there in various folders scattered. if you feel there is something "bad" about this, don't install GAT.

    As for the 2x-a few thousand, I have already explained that 3 times. Observe the Imp. ratio.

    So even if what you say makes sense, it doesn't make sense to do simply because I'd be full of lawsuits for a fraud product by all my existing users

    I wouldn't consider your program a fraud if it used the data to come up with a pattern of number selection which would have resulted in a number of wins above the norm if used for previous drawings and assumed it would do the same for future drawings.  But if its numbers selection has nothing to do with previous drawings of the data file you use then those numbers should do equally well in all similar games and that is hard to believe.

     * you don't need to buy more tickets, just buy a winning ticket * 
       
                 Evil Looking       

      lottoarchitect's avatar - waveform

      Greece
      Member #2815
      November 18, 2003
      502 Posts
      Offline
      Posted: December 23, 2011, 12:30 pm - IP Logged

      So even if what you say makes sense, it doesn't make sense to do simply because I'd be full of lawsuits for a fraud product by all my existing users

      I wouldn't consider your program a fraud if it used the data to come up with a pattern of number selection which would have resulted in a number of wins above the norm if used for previous drawings and assumed it would do the same for future drawings.  But if its numbers selection has nothing to do with previous drawings of the data file you use then those numbers should do equally well in all similar games and that is hard to believe.

      I'm not sure why you say GAT has nothing to do with previous drawings. It uses past draws equal to stat. data size defined to make a prediction. So, e.g. it uses draws 10-30 to evaluate draw 31 (which is a test draw), then draw 31 becomes part of that known set of draws used, so we now have the set 11-31 to predict draw 32 and this goes on till the end of the tested draws. Therefore, when predicting for the new draw to come, the same logic has been used through the whole set of tested draws and the same logic continues for the future draws not drawn yet. Therefore anything observed in the tested draws, the same behaviour more or less expected for the future draws. So if a GAT has managed to find e.g. 10 times a 4+ hit over the tested draws, we can expect this to continue in the future. It may produce 8 or 12 or something close to that but we don't expect this to be 0 or 30 hits in the future. GATs don't die suddenly and observing their behaviour over a set of e.g. 100 tested draws tells a lot of what to expect from its predictions.

      If you have something to do, at least do it well...

        RJOh's avatar - chipmunk
        mid-Ohio
        United States
        Member #9
        March 24, 2001
        19829 Posts
        Online
        Posted: December 24, 2011, 10:05 am - IP Logged

        I'm not sure why you say GAT has nothing to do with previous drawings. It uses past draws equal to stat. data size defined to make a prediction. So, e.g. it uses draws 10-30 to evaluate draw 31 (which is a test draw), then draw 31 becomes part of that known set of draws used, so we now have the set 11-31 to predict draw 32 and this goes on till the end of the tested draws. Therefore, when predicting for the new draw to come, the same logic has been used through the whole set of tested draws and the same logic continues for the future draws not drawn yet. Therefore anything observed in the tested draws, the same behaviour more or less expected for the future draws. So if a GAT has managed to find e.g. 10 times a 4+ hit over the tested draws, we can expect this to continue in the future. It may produce 8 or 12 or something close to that but we don't expect this to be 0 or 30 hits in the future. GATs don't die suddenly and observing their behaviour over a set of e.g. 100 tested draws tells a lot of what to expect from its predictions.

        I'm not sure why you say GAT has nothing to do with previous drawings.

        When I suggested it would be easy to construct a winning history using the data in the file since it had all the winning combinations and your response was there would be no other way to make a hit comparison test if the file was not there as if its only purpose was for comparison and not for coming up with anything associated with its winning history.

         * you don't need to buy more tickets, just buy a winning ticket * 
           
                     Evil Looking       

          RL-RANDOMLOGIC's avatar - usafce

          United States
          Member #59354
          March 13, 2008
          3972 Posts
          Offline
          Posted: December 24, 2011, 10:50 am - IP Logged

          Any well done analysis of past draws will produce some very good information, The problem

          is that good information and the next drawing are not related, well maybe 5th or 6th cousins.

           

          Merry Christmas ALL

          RL

          Working on my Ph.D.  "University of hard Knocks"

          I will consider the opinion that my winnings are a product of chance if you are willing to consider

          they are not.  Many great discoveries come while searching for something else

          USAF https://en.wikipedia.org/wiki/Prime_Base_Engineer_Emergency_Force

            US Flag Trump / 2016 & 2020  

            lottoarchitect's avatar - waveform

            Greece
            Member #2815
            November 18, 2003
            502 Posts
            Offline
            Posted: December 24, 2011, 12:07 pm - IP Logged

            Any well done analysis of past draws will produce some very good information, The problem

            is that good information and the next drawing are not related, well maybe 5th or 6th cousins.

             

            Merry Christmas ALL

            RL

            In a true random event this can be said and being undisputable. In the case of lotteries, when managing to have 5x or 10x improvements over probability, this tells me they are related somehow and it is what I call reduced randomness and actually try to exploit. If you call it 5th or 6th cousins, so be it but they are related. Otherwise, there is absolutely no reason why I get such hit improvements, but they are there.

            If you have something to do, at least do it well...

              Avatar
              bgonçalves
              Brasil
              Member #92564
              June 9, 2010
              2123 Posts
              Offline
              Posted: December 24, 2011, 1:15 pm - IP Logged

              Hello, the question is, how many will repeat the previous result,! For example, a lottery
              49 / 6, if I bet with 4,5,6 equals the previous results, 98%, will not win!
              Why give up to a maximum of three numbers, one could use it somehow as a base filter. Not living with this filter wheel has a maximum of 3 numbers from previous results,
              The best prediction has to be playing 10 to 20. For a single event, it is impossible, but within this range, yes! The Math Factor and random, part of the forecast, then a lottery 49 / 6 to try to predict up to 4 numbers

                Avatar
                bgonçalves
                Brasil
                Member #92564
                June 9, 2010
                2123 Posts
                Offline
                Posted: December 24, 2011, 4:06 pm - IP Logged

                Hello, lottoarquitet in studying the history of output, for sweepstakes and sweepstakes odd pair, get the number of hits that results to be obtained by choosing X number of groups more outgoing and late in the last N draws odd / even in previous 10 draws. example
                124 = ......... result pair
                125 =......... finished result in odd

                  RJOh's avatar - chipmunk
                  mid-Ohio
                  United States
                  Member #9
                  March 24, 2001
                  19829 Posts
                  Online
                  Posted: December 25, 2011, 12:54 pm - IP Logged

                  Any well done analysis of past draws will produce some very good information, The problem

                  is that good information and the next drawing are not related, well maybe 5th or 6th cousins.

                   

                  Merry Christmas ALL

                  RL

                  I Agree!  And many times that good information leads different people to different conclusions.

                   * you don't need to buy more tickets, just buy a winning ticket * 
                     
                               Evil Looking