Welcome Guest
Log In | Register )
You last visited December 11, 2016, 8:42 am
All times shown are
Eastern Time (GMT-5:00)

For those who use draw history...

Topic closed. 41 replies. Last post 12 years ago by JRLGGA.

Page 1 of 3
PrintE-mailLink

How far back is enough history?

Full History [ 20 ]  [26.67%]
Last 100 draws [ 9 ]  [12.00%]
Last 50 draws [ 8 ]  [10.67%]
Last 25 draws [ 7 ]  [9.33%]
Last 10 draws [ 17 ]  [22.67%]
last 3 draws [ 3 ]  [4.00%]
last draw only [ 3 ]  [4.00%]
Other (please post number) [ 8 ]  [10.67%]
Total Valid Votes [ 75 ]  
Discarded Votes [ 49 ]  
hypersoniq's avatar - 8ball
Pennsylvania
United States
Member #1340
April 6, 2003
2450 Posts
Offline
Posted: December 12, 2004, 11:05 am - IP Logged

Curious as to the number of draws back most people are using in calculations, regardless of the system employed.

If you could also post why you go back as far as you do that would be great ;-)

right now I have powerball history into 3 categories for calculations...

A. full history

B. last 100 draws

C. last 3 draws

they all get computed into a weighted average, but I am wondering if any one should have preference over another...

THanks to those who participate :-)

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

    Johnny5's avatar - japheth
    Ocala
    United States
    Member #3015
    December 13, 2003
    70 Posts
    Offline
    Posted: December 12, 2004, 11:07 am - IP Logged

    I use the past draws since any Ball Set change.

    Its about three years here in Florida.

    But for certain calculations I only use the last 30 to 40 draws.

    JOhnny5

    To play is not to win, but to win keeps you playing..       

      RJOh's avatar - chipmunk
      mid-Ohio
      United States
      Member #9
      March 24, 2001
      19831 Posts
      Offline
      Posted: December 12, 2004, 11:13 am - IP Logged

      I use the full history of the current matrix only.  Once numbers are added or reduced, the old history losses its value.

      RJOh

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

        CARBOB's avatar - FL LOTTERY_LOGO.png
        ORLANDO, FLORIDA
        United States
        Member #4924
        June 3, 2004
        5914 Posts
        Offline
        Posted: December 12, 2004, 11:21 am - IP Logged

        A little correction to Johnny5's post. Last ball change for Florida Cash 3 & Play 4 was May 2003, according to web site. I use 66 draws,12 draws, full draw history. I'm beginning to believe, no farther than last ball change, but I also believe that some combos will show in the same month. Example, all of the Cash 3 drawings, so far in Dec, except 12/11/04, hit in Dec, in past drawings. I  think you are going to get a lot of different opinons on this. 

        Carbob

          hypersoniq's avatar - 8ball
          Pennsylvania
          United States
          Member #1340
          April 6, 2003
          2450 Posts
          Offline
          Posted: December 12, 2004, 11:37 am - IP Logged

          My goal with blending stats from histories in various depths comes from observation...

          Full history seems to range the balls ok, last 100 seems to capture most trends... for instance in the first 100 draws of the new PB matrix, 20 was hot, now 31 is hot... this shows up in the last 100 draws. And last 3 seems to add a moving component so you aren't always playing a static number.

          I have an excel workbook with 3 sheets in 3 history depths that uses some A.N.N. weighting algorithms to boil it down to 1 single pick. When I added a midrange sheet (last 100) it seemed to make the numbers less inclined to move, mostly because the weight of the last 3 was reduced in the overall calculations. I am not set on the last 100 as a number, I would change it to last 50 or last 150 if it would help. Each week the matches seem to come from one of the components, either the full history, last 100 or last 3 is in the range of the drawing... once in awhile the composite is closest, but since that is the one I play, I would like to fine-tune it.

          In the end it is a q.p. with history as a guideline... just a pain to update and maintain... I want to get the formula wired then code up the whole mess into a VB.NET or C# prog that can handle last draw input and spit out the next pick. I am getting tired of wasting time on prediction... I am not saying it cannot be done, I am ready to admit to myself that I cannot do it. You gotta play to win, so I will keep my buck in the ring, but it would be nice to just generate the 1 pick right away then play it and forget it until the next draw.

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

            fja's avatar - gnome1

            United States
            Member #91
            January 19, 2002
            11937 Posts
            Online
            Posted: December 12, 2004, 11:43 am - IP Logged

            I use both full history and current year history & last 10 draws history....being that MM is only 269 draws old it hasn't even reach 1% of its draws.... 

            "Everybody has to believe in something...I believe I'll have another beer!"   = W.C.Fields                      

              lottoarchitect's avatar - waveform

              Greece
              Member #2815
              November 18, 2003
              502 Posts
              Offline
              Posted: December 12, 2004, 11:51 am - IP Logged

              Full history can be used to make backtests on theories to apply. It doesn't necessary mean it will work for the next draw of course but if you find an approach that provides good results constantly enough in the past, then you have a reason to believe that this approach will work again.

              Last 3 draws statistically offer good advice on numbers to thraw away and last 100 (it depends on the lottery size you play) offer a good guidance of due numbers. So, actually we need everything and not only a simple category.

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

                Avatar
                East Jordan, MI
                United States
                Member #2
                June 1, 2001
                675 Posts
                Offline
                Posted: December 12, 2004, 12:21 pm - IP Logged
                Quote: Originally posted by apagogeas on December 12, 2004

                Full history can be used to make backtests on theories to apply. It doesn't necessary mean it will work for the next draw of course but if you find an approach that provides good results constantly enough in the past, then you have a reason to believe that this approach will work again.




                I agree that using the full history to test static number theory is the most logical approach. For testing relative number theory; the smaller the sample the better.



                If you're lucky enough to have a state which is forward in revealing its ball change schedule. The exact number of draws to use for relative number theory would correspond to the set of draws beginning with the last ball change.



                I use a new hybrid theory, that I developed myself, which combines both.
                  WIN  D's avatar - q05Q0
                  Stone Mountain*Georgia
                  United States
                  Member #828
                  November 2, 2002
                  10491 Posts
                  Offline
                  Posted: December 12, 2004, 12:25 pm - IP Logged

                    Hyper.....It depends on the questions that we want to ask doesn't it?


                   


                   "If we knew the right questions to ask .....we would probably already know most of the answer."  


                   The data needed depends on the data question.   


                  I have noticed that Archaeologist dont throw away crap.........literally. In fact finding fossilized crap is one of their favorite things. (It speaks to them!) LOL             

                   

                   

                  The only real failure .....is the failure to try.                               

                                                Luck is a very rare thing....... Odds not so much. 

                                                Odds never change .....but probability does. 

                                                                                                         Win d    

                    four4me's avatar - gate1
                    MD
                    United States
                    Member #1701
                    June 18, 2003
                    8366 Posts
                    Offline
                    Posted: December 12, 2004, 12:40 pm - IP Logged

                    I'v tried it all . all the history for pick 3/4/5 games didn't do much for me.  The last 5 days worth of history for pick 3/4 games worked best.  At leats 4 times a week a pick 3 or 4 number would be drawn from numbers in that range. For match 5 i use the last 10 draws and it seems to work for me. In mega and power ball i use about 2/3 months worth of history using half hot numbers 1/4 luke warm and 1/4 cold numbers. this works for me. It's kinda inportant to me to use less history because it's possible they are using the same balls for the games or the same group of stored balls. when the numbers change drastically meaning not drawing many of the same numbers i figgure they changed the balls to a different set of them. So i wait for a couple of draws and go with the new flow of numbers.

                      0zerothru9nine's avatar - Lottery-018.jpg
                      NEW JERSEY
                      United States
                      Member #2525
                      October 14, 2003
                      39 Posts
                      Offline
                      Posted: December 12, 2004, 4:09 pm - IP Logged

                      Full history for a complete picture.  90 days pattern if you watch the flow closely example  NJ Last  hot watch  963 combo 395 followed. I cashed in  on both.  To answer your question. that's just one of many ways to arrive and the winnings combination. Most of all don't forget to listen to yourself.

                        Avatar
                        Greenwich, CT
                        United States
                        Member #4793
                        May 24, 2004
                        1822 Posts
                        Offline
                        Posted: December 12, 2004, 5:20 pm - IP Logged

                        I used to use full history, last 30 and last 10 draws.  Now I use full history, last 20 and last 5, because you're always trying to catch trends, and I think shorter history lets you catch them when they occur.

                          Maverick's avatar - yinyang
                          USA
                          United States
                          Member #8242
                          October 29, 2004
                          1133 Posts
                          Offline
                          Posted: December 12, 2004, 5:59 pm - IP Logged
                          Quote: Originally posted by hypersoniq on December 12, 2004


                          I am getting tired of wasting time on prediction... I am not saying it cannot be done, I am ready to admit to myself that I cannot do it. I will keep my buck in the ring, but it would be nice to just generate the 1 pick right away then play it and forget it until the next draw.


                          I agree.
                            four4me's avatar - gate1
                            MD
                            United States
                            Member #1701
                            June 18, 2003
                            8366 Posts
                            Offline
                            Posted: December 12, 2004, 6:17 pm - IP Logged

                            I think using the full history all the way back to the beginning of the game when the changed the amount of balls only tells you how many times each ball was drawn. This information isn't going to tell you what's due. Looking at powerball and megamillions last twenty or so drawing tells me more about what balls are currently being drawn. Not what happened last year. Plugging the last 20/30 games into a system would probably produce better results. The only thing that gets in the way is when they draw a number that is cold and hadn't hit in those 20/30 draws. So upon making your list of numbers to plug into your system. keep track of cold balls and include them if you have a hunch that they are going to come out. Sometimes I make two lists of numbers. The last 20/30 games whose numbers have been drawn and a list of numbers that haven't been drawn in that time frame. Then I wheel them separately. I check them against past draws to make sure I'm not picking a whole set of numbers that has already been draw. Then I wheel the hot numbers with the cold numbers and check those sets. Then I pick my plays from the 2 lists this way I have covered all my bases so to speak. 

                              Avatar
                              Melbourne
                              Australia
                              Member #5731
                              July 18, 2004
                              36 Posts
                              Offline
                              Posted: December 12, 2004, 7:50 pm - IP Logged

                              If I might add, one can't talk about numbers being cold or hot or due unless you take into account previous draw history.... And why refer to 100 past draws or 30 past draws, why not 107 past draws or 29? Nothing special about 100 or 30. Or 107 and 29 for that matter. Seems to me one is either a proponent of using past draws or one is not. If one is not a proponent of past draws then all numbers are standing at the line with an equal chance at the next draw, and the whole thing is a complete random walk. End of story. Except if there is any bias in the lottery. hmmm. If one is a proponent of analysing past draws then the theory must start with something like 'over time' every number ought to appear the same number of times as every other number. But what about before we get to the end point ie we are half way there to this mysterious end point called 'over time'? If some numbers have distinctly lagged then those lagging numbers have to catch up, thus an entire data analysis may be appropriate, the difficulty of course always being Where are we? along the path to reaching that time postion called over time?



                              For me I study the whole of data and then I look at the last 4 draws and think of Bernouli (who has very kindly prepared some probablity tables) to see what numbers to reject.



                              relowe