Welcome Guest
Log In | Register )
You last visited December 4, 2016, 11:27 pm
All times shown are
Eastern Time (GMT-5:00)

Gap Analysis

Topic closed. 5 replies. Last post 12 years ago by hypersoniq.

Page 1 of 1
PrintE-mailLink
JKING's avatar - Kaleidoscope 3.gif

United States
Member #5599
July 13, 2004
1184 Posts
Offline
Posted: July 18, 2004, 10:25 am - IP Logged

  After looking at the horizontal gap history, for fantasy 5), for over 300 games, I'm confused how this kind of compaison can yield any useful information. Yet, it is mentioned in several forums as an important element in analyzing your numbers. What am I missing...help!

Horizontal gap summary table for Fantasy 5

gap    L1    L2    L3     L4






11703410205
21166560187
34573200138
43568443150
51154546125
65374518105
7222411176
8118502190
9010332669
1004373071
1100243357
1200153752
130093140
140053035
150012829
160003131
170002020
180001111
190001717
2000099
210001010
2200033
2300044
2400000
2500011
2600022
2700011
2800011
2900000
3000011
3853853853851540



L1 indicates the lowest sorted gap amount

L4 indicates the highest sorted gap amount

    RJOh's avatar - chipmunk
    mid-Ohio
    United States
    Member #9
    March 24, 2001
    19823 Posts
    Offline
    Posted: July 18, 2004, 10:59 am - IP Logged

    In the RNG program that I use to pick my numbers, gap and range are used as parameters to control how numbers are distributed.  For example in picking numbers to play in MegaMillions,  I might enter parameters of 20-51 for ranges and 0-20 for gaps and discard any combinations which had 2 or more adjacent gaps of 1(0) such as 15-16-17-35-40.  Gaps, ranges and sums are the easiest kinds of data to calculate so they are mentioned all the time weather they are useful is a question yet to be answered.

    RJOh

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

      JKING's avatar - Kaleidoscope 3.gif

      United States
      Member #5599
      July 13, 2004
      1184 Posts
      Offline
      Posted: July 18, 2004, 12:39 pm - IP Logged

      So, looking at the the revised data below:

      I might want to pick the highlighted area of the total amount column to pick numbers that happen (330/386)*100 =86% of the time, as an example.



      L1L2L3L4TOTTOTTOT
                VALAMT
                   
                 
      247162910
      114192520
      124121930
      26682240
      22451350
      148223560
      13691970
      233162480
      3313173691
      145616102
      2232330110
      1132126121
      3781028138
      35111938147
      66101335154
      4581229167
      24101329179
      1389211820
      57817371911
      37910292012
      14510202117
      111017292213
      125917239
      33523342413
      24419292524
      12415222621
      12320262720
      46714312821
      151213312926
      471010313016
      24413233126
      1489223217
      24510213318
      1388203419
      131415333514
      691013383613
      2451728379
      112913387
      2241220390
      1231824TOT385



      only partial raw data view

        hypersoniq's avatar - 8ball
        Pennsylvania
        United States
        Member #1340
        April 6, 2003
        2450 Posts
        Offline
        Posted: July 18, 2004, 12:42 pm - IP Logged

        gaps are a profiling tool... you can gather the past gap information and make sure your pick(s) "fit" the profile in hopes of a win... I once wrote a program to filter all the possible combos in the PA cash 5 to leave only those that matched the "profile" for gap, range and sum... the number that was drawn wasn't in the 200,000+ combos...

        the only thing I took from the study of profiling tools (gaps, sums, odd/even, high/low, delta etc..) is that numbers that stand out statistically are rarely drawn together. I once took this to the extreme with cash 5's 3000+ draw history, by identifying the most popular numbers in each position, then filtering the history to find their most popular companions... the whole thing only generated 5 picks, and never did better than matching 3 a few times... I didn't need to try it for real, backtesting is easier and cheaper ;-)

        the major flaw... if one of the popular numbers was not drawn, your picks were basically useless....

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

          JKING's avatar - Kaleidoscope 3.gif

          United States
          Member #5599
          July 13, 2004
          1184 Posts
          Offline
          Posted: July 18, 2004, 1:05 pm - IP Logged

          That's pretty much the game I'm playing right now, profiling, however, I never key off any key numbers, only the numbers that best fit multiple profiles. I may in the long run end up with the same conclusion you have. But for now it's a work in progress. As far as results, I'm doing faily well in the two and three hit range, but have noticed I would have had 4 hits several times if I had been betting more. The biggest problem I see is that error %'s are additive, so if you have 10 filters/profiles each with a 5% error, then you end up with picks that are either 50% right or wrong.

          Thanks for the input and put backtracking routines on my list of things to do.

            hypersoniq's avatar - 8ball
            Pennsylvania
            United States
            Member #1340
            April 6, 2003
            2450 Posts
            Offline
            Posted: July 18, 2004, 1:23 pm - IP Logged

            good luck.

            Also, never give up just because someone else has had bad results... the journey through picking numbers is one everyone (especially someone developing their own systems) should make on their own.

            the other suggestion I have is to keep past failed systems on hand, so you can review where you have been.

            As far as filters go, that error can be corrected out, but it's in hindsight on past results... I was once working on an excel spreadsheet for powerball and made a cell reference error... what happened was that the actual results were only a few numbers off of my picks... consistently. I was quite thrilled until I realized I was referencing the wrong data set... (checking from actual results rather than my generated picks... of COURSE that way shows hits galore).

            now that I have been down several roads, I might go and revisit profiling to apply to the one pick I now generate... couldn't hurt...

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