Welcome Guest
Log In | Register )
You last visited December 4, 2016, 3:06 am
All times shown are
Eastern Time (GMT-5:00)

Statistics, weighting and scope....

Topic closed. 16 replies. Last post 12 years ago by hypersoniq.

Page 1 of 2
PrintE-mailLink
hypersoniq's avatar - 8ball
Pennsylvania
United States
Member #1340
April 6, 2003
2450 Posts
Offline
Posted: May 7, 2004, 9:48 pm - IP Logged

Powerball Statistics... there is not really a great deal of draw history available for the 5/53 + 1/45 game... 165 draws... but that is what we have to work with...

Let's take a look at a concept borrowed from the weighting of Neural Network Cells... Long Term and Short Term "Memory" AND Weighting inputs...

Let's jump right into real numbers...

in 165 games (long term memory) ,the most popular (by position...sorted order) 02-17-34-42-52 PB  35

in the last 16 games only (short term memory), same stat... different numbers (mostly)... 02-07-34-39-47 PB 6

The question is... how to combine them... obviously the 2 would be a steady pick for position 1, and the 34 for position 3... but what about the 2nd white ball? range from 7 to 17... average? I don't know... tomorrow I may try the average...

What about weighting? average of 7 and 17 is 12... but that keeps LTM and STM as equal... which is more important? long term memory or short term trends? I went last week with short term memory... my 1 line matched 2 numbers in position...21 and 25... that pays squat... but it was promising as the others were not far off... (picked PB to be 7... it was 8... had I done the proper average, it was 7.64... excel apparently doesn't round automatically with the INT function)... could have been $7...

My goal with this new workout strategy is to get one good line... I will also play a control QP just to see which does better... weight ratio is one variable I still need to work on... STM does make numbers change more dynamically than just stats on the entire history... it may not work but just might provide the extra oomph to catch 3 numbers this time... if it gets close again then wednesday I'll go with a +1 line above and a -1 line below.

That's what I'm working on right now... what's in your spreadsheet?

Playing more than one ticket per game is betting against yourself.

    hypersoniq's avatar - 8ball
    Pennsylvania
    United States
    Member #1340
    April 6, 2003
    2450 Posts
    Offline
    Posted: May 9, 2004, 7:47 am - IP Logged

    no big bux (but Someone in PA won it) ,but how is this for a ticket...

    02-09-17-33-43 PB 17

    3 hits IN POSITION and only one off on the 2, using the above method of weights and LTM & STM...

    that's $7 on $2 in tickets ;-)

    my control QP had nothing, but was closer on the PB with a 40

    results 03-09-17-37-43  PB 39

    Playing more than one ticket per game is betting against yourself.

      ayenowitall's avatar - rod serling4.jpg

      United States
      Member #4416
      April 22, 2004
      1075 Posts
      Offline
      Posted: May 10, 2004, 3:45 am - IP Logged

      hyper,

      I'm still working on trying to generate a useful number sequence, but I've applied weighting to the dates and frequencies in my standby method. It appears to have helped in narrowing down the numbers. I hit 3 of 6 on two lines of my LOTTO South plays at $5 each on Saturday. I played five lines , so I made $5. I'll see what the weighting does for my results on Wednesday night. I'll be playing a few more lines, but I'm satisfied that I have all the good numbers covered.

        hypersoniq's avatar - 8ball
        Pennsylvania
        United States
        Member #1340
        April 6, 2003
        2450 Posts
        Offline
        Posted: May 11, 2004, 4:50 am - IP Logged

        one-line tests coming up...

        A. tonight for Match6 (using weighted top performance numbers, only no control Quick Pick because they give you 2 QPs)

        B. wednesday's PB (using a combined weighted stat set that uses full history and wednesday-only data... also the weighted top performance line)

        C. Friday's Match6 (IF tuesday results are in the ballpark... otherwise I'll skip it)

        D. saturday's powerball (same method as wed. only with full history + sat. only data)

        Test cost... $10 (not bad to be in 4 jackpot games, considering Match6 is a $2 ticket)

        total lines... 3 per match6 draw(counting the 2 QPs included), 3 per PB draw (counting the control QP)

        test type... A & B = blackout (picks not posted)  C & D = Open (picks posted to "prediction" feature along with the QPs for control purposes... posted just prior to draw time[10 to 15 min. Match6, 30 min. PB])

        Test pass criteria... MY picks are closer than QPs... looking at overall range by position as well as individual exact matches

        Playing more than one ticket per game is betting against yourself.

          hypersoniq's avatar - 8ball
          Pennsylvania
          United States
          Member #1340
          April 6, 2003
          2450 Posts
          Offline
          Posted: May 11, 2004, 8:14 pm - IP Logged

          A. test results...

          My pick...

          04-07-16-18-32-41

          Results...

          06-07-15-31-38-43

          (-2)(Hit)(+1)(-20)(-6)(-2) [individual numbers]

          (Hit)(Hit)(Hit)(Miss)(Hit)(Hit) [decades]

          Playing more than one ticket per game is betting against yourself.

            hypersoniq's avatar - 8ball
            Pennsylvania
            United States
            Member #1340
            April 6, 2003
            2450 Posts
            Offline
            Posted: May 13, 2004, 8:44 pm - IP Logged

            B. test results...

            My picks...

            Best Performer (full history)... 03-15-26-41-49  PB 27

            Combined (Full + wednesday only)... 07-16-27-35-46  PB 23

            Results...

            11-31-35-42-45  PB 35

            (-8)(-16)(-9)(-1)(-4)  PB (-8) [individual numbers 1st game]

            (-4)(-15)(-8)(-7)(+1)  PB (-12) [individual numbers 2nd game]

            (Miss)(Miss)(Miss)(Hit)(Hit)  PB (Miss) [decades 1st game]

            (Miss)(Miss)(Miss)(Miss)(Hit)  PB (Miss) [decades 2nd game]

             

            not looking to bright with this latest idea... Next try is test C. tomorrow

            Playing more than one ticket per game is betting against yourself.

              hypersoniq's avatar - 8ball
              Pennsylvania
              United States
              Member #1340
              April 6, 2003
              2450 Posts
              Offline
              Posted: May 14, 2004, 6:56 pm - IP Logged

              C. Test pick... 4 min. before draw time...

              03-05-14-19-31-41  (PA Match 6)

              Playing more than one ticket per game is betting against yourself.

                hypersoniq's avatar - 8ball
                Pennsylvania
                United States
                Member #1340
                April 6, 2003
                2450 Posts
                Offline
                Posted: May 14, 2004, 7:09 pm - IP Logged

                C. test results...

                My pick...

                03-05-14-19-31-41

                Results...

                01-07-19-23-33-44

                (+2)(-2)(-5)(-4)(-2)(-3) [individual numbers]

                (Hit)(Hit)(Hit)(Miss)(Hit)(Hit) [decades]

                 

                not so hot on a per number basis, but it did get 5of6 right on the decades again... the same 5of6... There may be a use for it yet...

                Final Test D. tomorrow night for PB...

                Playing more than one ticket per game is betting against yourself.

                  hypersoniq's avatar - 8ball
                  Pennsylvania
                  United States
                  Member #1340
                  April 6, 2003
                  2450 Posts
                  Offline
                  Posted: May 15, 2004, 10:14 pm - IP Logged

                  D. Pick (almost 1 hour before draw time)

                  Since saturday-only data is too wild... I just went with the one weight-adjusted LTM/STM combination... for 5/15/2004 MUSL Powerball

                  03-15-28-41-49  PB 28

                  note how similar they came out compared to test B. (only the 28 varies in the 3rd position and the PB... replacing 27 from wed...) I need something that catches more wildness... For wednesday I will make more changes to the overall formula to give even more weight to short term stats...

                  Also played another ticket based on the average number to come up AFTER a 10 million dollar drawing...

                  08-18-27-33-43  PB 19

                  I should have skipped the QP in favor of the second "last minute" line... so now I have 3...

                  Playing more than one ticket per game is betting against yourself.

                    hypersoniq's avatar - 8ball
                    Pennsylvania
                    United States
                    Member #1340
                    April 6, 2003
                    2450 Posts
                    Offline
                    Posted: May 16, 2004, 9:25 am - IP Logged

                    D. test results...

                    My picks...

                    Best Performer (full history)... 03-15-28-41-49  PB 28

                    after 10M draws (Average)... 08-18-27-33-43  PB 19

                    Results...

                    01-26-27-43-46  PB 31

                    (+2)(-11)(+1)(-2)(+3)  PB (-3) [individual numbers 1st game]

                    (+7)(-8)(Hit)(-10)(-3)  PB (-12) [individual numbers 2nd game]

                    (Hit)(Miss)(Hit)(Hit)(Hit)  PB (Miss) [decades 1st game]

                    (Hit)(Miss)(Hit)(Miss)(Hit)  PB (Miss) [decades 2nd game]

                    The "after 10M" matched 2 but only one in position...

                    Playing more than one ticket per game is betting against yourself.

                      hypersoniq's avatar - 8ball
                      Pennsylvania
                      United States
                      Member #1340
                      April 6, 2003
                      2450 Posts
                      Offline
                      Posted: May 16, 2004, 9:44 am - IP Logged

                      results...

                      Test cost $10

                      Return on picked numbers... $0

                      Return on QPs... $9 (Game A. 3/6 + 4/18 & Game C. match 4/18)

                      net loss $1

                      system loss $10... noting I picked payed any money... bottom line... QPs are not a bad bet to pick up along with regular game tix... I must work more at that one line match 6 pick because that's where all the $$$ came from that helped defray the cost...

                      Good luck!

                      Playing more than one ticket per game is betting against yourself.

                        johnph77's avatar - avatar
                        CA
                        United States
                        Member #2987
                        December 10, 2003
                        832 Posts
                        Offline
                        Posted: May 16, 2004, 9:34 pm - IP Logged

                        hypersoniq -

                        Makes you want to believe Charlie Brown of Peanuts - "Tell your statistics to shut up." Thanks for the report and brutal honesty.

                        gl

                        john

                        Blessed Saint Leibowitz, keep 'em dreamin' down there..... 

                        Next week's convention for Psychics and Prognosticators has been cancelled due to unforeseen circumstances.

                         =^.^=

                          hypersoniq's avatar - 8ball
                          Pennsylvania
                          United States
                          Member #1340
                          April 6, 2003
                          2450 Posts
                          Offline
                          Posted: June 5, 2004, 5:12 pm - IP Logged

                          I finally got around to updating my match 6 sheet to include short term trends in it's one final pick... first test will be on tuesday... the beauty of the system is it is fully automated, including weighing!

                          the other major project I have to do is remove the INT keyword from all of my excel sheets, I think it would be more accurate to let the numbers flow and round the final pick off...

                          then I must do this with my ENTIRE powerball sheet, that is my june pet side project...

                          in the end I will have eliminated all other sources of numbers and be down to just 1 ticket per jackpot game in PA... a grand total cost of $6 per week or $8 if I decide to go with the power play ( I might just do that too)... I will be in for 2 draws of the match 6 and 2 draws of the powerball for under $10, win or lose... this was my initial goal anyway... ONE pick per game...

                          Playing more than one ticket per game is betting against yourself.

                            hypersoniq's avatar - 8ball
                            Pennsylvania
                            United States
                            Member #1340
                            April 6, 2003
                            2450 Posts
                            Offline
                            Posted: June 6, 2004, 10:02 am - IP Logged

                            finally done the powerball sheet, the one pick will be tracked and played for a few draws... I am not liking how the numbers stay close together, I am wondering if adding more weight to the short term data would be better...

                            draw data produces interesting numbers, including duplicates... that's why I skip draw data.

                            One test coming tuesday for Match6 and one wednesday for PB. I have got to come up with a way to backtest... now the formulas are complex and some calculations call data from multiple sheets...do-able but tedious...

                            Playing more than one ticket per game is betting against yourself.

                              Cleveland,Oh
                              United States
                              Member #1856
                              July 16, 2003
                              531 Posts
                              Offline
                              Posted: June 10, 2004, 8:50 pm - IP Logged
                              Quote: Originally posted by hypersoniq on May 14, 2004



                              C. test results...

                              My pick...

                              03-05-14-19-31-41

                              Results...

                              01-07-19-23-33-44

                              (+2)(-2)(-5)(-4)(-2)(-3) [individual numbers]

                              (Hit)(Hit)(Hit)(Miss)(Hit)(Hit) [decades]

                               

                              not so hot on a per number basis, but it did get 5of6 right on the decades again... the same 5of6... There may be a use for it yet...

                              Final Test D. tomorrow night for PB...





                              My pick...

                              03-05-14-19-31-41=113

                              Results...

                              01-07-19-23-33-44=126

                              Whats the problem!