Welcome Guest
Log In | Register )
You last visited January 21, 2017, 2:30 am
All times shown are
Eastern Time (GMT-5:00)

Something Better

Topic closed. 6 replies. Last post 10 months ago by osmannica2001.

Page 1 of 1
PrintE-mailLink
Avatar

United States
Member #98539
October 9, 2010
57 Posts
Offline
Posted: February 10, 2016, 6:26 pm - IP Logged

I think that I may have found something that is even better than Neural Net Lottery Picker and Percep-Lotron. 

 

It is called Neural-Lotto.  You will have to Google it to see what I mean.  Do not forget the dash in the name when you Google it.   

 

It just went live in January of this year.

 

It has eight parameters.   They are as follows:

 

Learning Rate    --- from 0 to 1

 

Neurons / Layers  ---   from 10 to 990

 

Layers  --- from 1 to 999

 

Stepping --- from 1 to 200

 

History --- from 10 to 200

 

Epochs -- from 100 to 10,000

 

Algorithm --- from 1 to 5

 

Momentum -- from 0 to 1

 

I have been experimenting with the settings.  I have not yet had a chance to study the results. 

 

Now that I have told all on this forum / website of the existence of this thing, I hope that someone will reciprocate by posting the results of his or her experiments.  I am particularly interested in knowing the best possible settings for a Pick 5 game such as Mass Cash which I play.

    SergeM's avatar - slow icon.png
    Economy class
    Belgium
    Member #123700
    February 27, 2012
    4035 Posts
    Offline
    Posted: February 11, 2016, 12:01 pm - IP Logged

    Much too long!

    Chart -> Decision -> Play

      Avatar

      United States
      Member #168877
      September 20, 2015
      47 Posts
      Offline
      Posted: February 11, 2016, 12:19 pm - IP Logged

      best of lucks having your money sucked away. I recently submitted a recommendation to at least show what the gradient descent looks like based on your plays. Meanwhile, you're running in blind mode.

        Avatar

        United States
        Member #168877
        September 20, 2015
        47 Posts
        Offline
        Posted: February 11, 2016, 12:24 pm - IP Logged

        So far I have spent about $50 in plays since July of 2015. But, I still find it very complex to use as it is lacking many features. Don't get me wrong however, they have very fast servers.

          SergeM's avatar - slow icon.png
          Economy class
          Belgium
          Member #123700
          February 27, 2012
          4035 Posts
          Offline
          Posted: February 11, 2016, 12:27 pm - IP Logged

          Wasn't this the project of three bored ex-students who got drunk in Mexico?

            Avatar

            United States
            Member #98539
            October 9, 2010
            57 Posts
            Offline
            Posted: February 18, 2016, 8:54 pm - IP Logged

            best of lucks having your money sucked away. I recently submitted a recommendation to at least show what the gradient descent looks like based on your plays. Meanwhile, you're running in blind mode.

            Dear Osmannica2001,

             

            In plain English please, what is a gradient descent as it relates to this kind of software?

             

            Thank you very much in advance for any light that you can shed on this matter.

             

            Kosh

              Avatar

              United States
              Member #168877
              September 20, 2015
              47 Posts
              Offline
              Posted: March 10, 2016, 11:20 am - IP Logged

              Hmm, gradient descent is a monster topic itself. But in a few words; gradient descent is what Visually will allow you to see how an output is close to zero after reiterating training examples. Evidently, the closer your output is to zero, the higher chances your have of your prediction being 100% accurate.