Prediction Discussion: Round 2

Reply #52

Clipper — Thu, 04 Mar 2004 17:59:27 GMT

hypersonig: All this sounds like too much number crunching, etc to me lol I try to keep it plain and simple. Like you said before.....the Pick 3 game can be 3 games in one if you track by position.Consequently, part of my system is tracking by position (hopefully to obtain the EXACT to play). With 10 digits, half of them Odd and half Even you have a 50/50 chance of getting which way a digit is going to lean. Then try to figure out if the next digit will be more or less then the last.Y... [ More ]

Reply #51

Millennium — Thu, 04 Mar 2004 13:24:56 GMT

Well, I hope you are good at programming. I've peeked into books about programming for neural nets, fuzzy logic, pattern recognition, stochastic methods ... have the aspirin at hand.I can see patterns. But programming is Martian to me.And there I was reading The Constants of Nature yesterday afternoon, deep into how the number 137 is significant, and ... aw, you can guess the rest

Reply #50

hypersoniq — Wed, 03 Mar 2004 17:19:21 GMT

I think I know my problem... I cannot see patterns!People claim odd/even, v-trac etc.. sums root sums and such can help identify trends ,I cannot see these, sometimes after the fact but never in time.I will continue to keep my spreadsheets up to date in case something else comes along, but I don't forsee my pick 3 playing ever getting off of paper... I am going back to the programming side of things, as my analysis skills are woefully inadequate.My local library was kind enough to order a book f... [ More ]

Reply #49

Millennium — Tue, 02 Mar 2004 20:00:22 GMT

It's really like digging for coal to unearth this thread! Perhaps Todd will do us a favor and move this and the first part to Systems?I'll say it again: You are trying to do too much at one time.I've begun to read The Constants of Nature and that gave me the idea to tell you this:What you are trying to do is to come up with what is known in physics as a Theory of Everything. Physicists didn't start out trying to do that. They went one step at a time. I think you should break off the last year's... [ More ]

Reply #48

hypersoniq — Sun, 29 Feb 2004 00:16:52 GMT

maybe adding the other 12 subsystems is missing something...the full system is currently behind... by less than 2 straight hits.This means that if you played $20 per day since 1977, EVERY DRAWING ,you would have lost money... you would be in the red a grand total of $840I'm not really up to any more number crunching to see what tradeoff NOT playing forecasted triples AND eliminating repeated picks does to the bottom line... but apparently 19 straight hits in each subsystem is NOT enough, time to... [ More ]

Reply #47

hypersoniq — Sat, 28 Feb 2004 22:58:13 GMT

expanded composite system picks for 2/29/04 EVE PA pick 3... 217 appears to be the next due (Sub-System J)...257839953214203136238030217201255345459434861939032596

Reply #46

JAP69 — Sat, 28 Feb 2004 06:18:00 GMT

HypersonigTry this workout to see how it has done over the past27 years of the Pa lottery.Take last drawn number and change only one digit up ordown.______________105example 216____216 = 116-316-206-226-215-217______________327now take those numbers and place them in box order116-136-026-226-125-127. On the next draw see how manytimes did they hit in that exact order.now take those numbers and place them in reverse box order611-631-620-622-521-721.On the next draw see how manytimes did they hit

Reply #45

hypersoniq — Sat, 28 Feb 2004 02:53:47 GMT

Am I way off on this a

Reply #44

hypersoniq — Fri, 27 Feb 2004 14:22:58 GMT

I have identified 21 subsystems that individually have profited over 27 years, I just took the top 8 for now... I may just have to add the rest to see if better coverage can be obtained...

Reply #43

Millennium — Fri, 27 Feb 2004 14:05:54 GMT

Quote: Originally posted by doodoohead on February 24, 2004hyper, There is a guy named Sloane a mathematician at AT T Bell Labs, >snip<This is the first time I've heard of this service. I will have to investigate. Thanks much

Reply #42

Millennium — Fri, 27 Feb 2004 14:04:45 GMT

Quote: Originally posted by hypersoniq on February 25, 2004millennium, how many picks did you generate to get several matches per month?Too damned many! As low as one as high as 20+. This is per draw. Multiply that by the days in the month......oh, and some days, *no* prediction. I was just stymied. It happens.

Reply #41

hypersoniq — Thu, 26 Feb 2004 23:41:35 GMT

1977....2 Winners B,F Losers A,C,D,E,G,H1978...10 Winners A(3),C,D,F(2),G,H(2) Losers B,E1979....6 Winners A,C(2),D(2),E Losers B,F,G,H1980....5 Winners A(2),F,H(2) Losers B,C,D,E,G1981....2 Winners C,G Losers A,B,D,E,F,H1982....5 Winners A,B,C,E,G Losers D,F,G1983....9 Winners B,C,D,E(4),F(2) Losers A,G,H1984....7 Winners A(2),B(2),E(2),G L... [ More ]

Reply #40

hypersoniq — Thu, 26 Feb 2004 19:34:45 GMT

As to what I found for the initial +1,+1,+1 system...ST BX ss P1 P2 P315 60 -9 913 947 94115 52 -8 951 861 92111 47 -7 1016 925 92912 45 -6 929 885 94613 43 -5 892 929 91612 41 -4 923 949 95510 34 -3 922 906 939 4 36 -2 873 901 88512 52 -1 915 888 88910 50 0 942 953 941 5 47 1 957 955 910 7 52 2 930 935 84810 39 3 955 922 905 6 42 4 903 865 95911 41... [ More ]

Reply #39

hypersoniq — Thu, 26 Feb 2004 19:17:59 GMT

The first rough draft at analysis of the above described composite pick 3 system (including instructions on performing the rundowns)... breaking down by winning number and subsystem for each year...Subsystems are identified by alpha codeAdd the number from the last draw to the positions as indicated (always add, including negativenumbers, take absolute values, drop the 10 column on numbers greater than 9 (e.g. 15 = 5) and lessthan -10 (e.g. -14 = 4) )A = -7,-8,-9B = +3,+7,-9C = -1,-5,+7D = 0,

Reply #38

hypersoniq — Thu, 26 Feb 2004 01:17:22 GMT

millennium, how many picks did you generate to get several matches per month

Reply #37

hypersoniq — Thu, 26 Feb 2004 01:15:55 GMT

TESTpicks for thursday 2/26/2004 EVE draw in PA...0-6-20-1-26-3-47-0-80-4-90-5-89-9-10-3-3to keep the thread less cluttered, this is the last time I will put picks in here... I'll keep any test picks in the Predictions feature of LP.

Reply #36

hypersoniq — Thu, 26 Feb 2004 01:01:28 GMT

wow! great info!I will definitely look into that...

Reply #35

doodoohead — Wed, 25 Feb 2004 01:47:58 GMT

hyper, There is a guy named Sloane a mathematician at AT T Bell Labs, well known for his work in graph theory, combinatorics and geometry. He is also the author of the Handbook of Integer Sequences, a collectionof some 2,300 sequences.Sloane's sequence database is also accessible by electronic mail. If you send a message to the Internet address sequences@research.att.com with the text lookup 1 1 2 5 14 42 132 429, you will receive a reply (typically a few minutes later) identifying the sequenc... [ More ]

Reply #34

hypersoniq — Wed, 25 Feb 2004 01:21:18 GMT

Quote: Originally posted by Millennium on February 24, 2004 you're chasing a chimera. we all are...Give the fact that any ONE of the picks would have left you ahead, I don't think chambrs make a difference because they hit individually over all the years...Even this exercise has not answered the riddle of prediction, but it may be giving us limits and parameters to attempt to predict... Still not enough data to tell (or time to analyze it this week....

Reply #33

hypersoniq — Wed, 25 Feb 2004 01:16:32 GMT

Quote: Originally posted by Dr Pepper on February 24, 2004 Interesting lack of response. see above...also, feel free to jump in and help out...

Reply #32

hypersoniq — Wed, 25 Feb 2004 01:15:41 GMT

Quote: Originally posted by doodoohead on February 24, 2004 Hyper, would you ask your next question and let me try to answer it. Maybe we can make this easier. Try to be as specific as you can. One question at a time. Perhaps we don't have to go all the way around the tree to get to the same place. Thanks. half of the battle is knowing what questions to ask... I will take you up on that as soon as I have properly figured out what it is exactly I need to know. Thanks

Reply #31

Millennium — Tue, 24 Feb 2004 19:51:02 GMT

Speaking of chimeras, after reading the number of hits you counted over the past 27 years, I went and counted how many times I correctly predicted for the past year. It was more than once a month.But guess what? That means *nothing*.Any prediction system is going to spit out several numbers. How many of those numbers can you afford to play each day/every draw? Of all the correct predictions, I only played *one*. The others I simply didn't *believe* and so did not play. (Yes, I have been very sor... [ More ]

Reply #30

Millennium — Tue, 24 Feb 2004 19:30:32 GMT

Hyper: Over the course of those 27 years, have they used the same chamber system (I mean identical design; NY's has changed)? If not, you're chasing a chimera.

Reply #29

Dr Pepper — Tue, 24 Feb 2004 19:23:07 GMT

That was a tempting o

Reply #28

hypersoniq — Tue, 24 Feb 2004 16:56:42 GMT

mad? no... just excited because something finally shows a profit versus backtesting... I have seen NO other system even backtested, let alone strong enough to generate a profit over 27 years... of course the profit is small right now, but that's where the analysis comes in... determining not only what to play when, but also what and when NOT to play.... that profit number will only go up...last time around we had alot of great ideas ,but no back-testing... no evidence that there was any merit...... [ More ]

Reply #27

doodoohead — Tue, 24 Feb 2004 16:45:16 GMT

Hyper, would you ask your next question and let me try to answer it. Maybe we can make this easier. Try to be as specific as you can. One question at a time. Perhaps we don't have to go all the way around the tree to get to the same place. Thanks.

Reply #26

Millennium — Tue, 24 Feb 2004 13:55:54 GMT

Quote: Originally posted by hypersoniq on February 23, 2004total cost to play (without refinements, this is for everyday so far) $73,616total winnings to date $83,000... not bad!Earth to hypersoniq: Please return!Go re-read what you wrote there.Have you gone mad

Reply #25

hypersoniq — Tue, 24 Feb 2004 13:42:10 GMT

once I have defined an easy update protocol, I will begin placing the 8 picks in the Predictions section... look for some actual test data by the weekend...Nothing speaks louder than proof, right? ;-)unfortunately there are still relatively long stretches (some over 6 months) dry (no straight hits)... You will undoubtedly see alot of boxes (there were over 450 in the past 27 years)... but we're only looking at the STRAIGHTS.If this goes the way it has over the game history... once Todd starts... [ More ]

Reply #24

hypersoniq — Tue, 24 Feb 2004 13:21:55 GMT

The analysis of the hits will take quite some time... if need be I can add all of the 19's, which also beat the system (at least for the next 250 or so draws)... I will be breaking down the wins by day/month/Quarter... trying to determine any indicators that may make one of the particular numbers better to play than another... I'll post data when I have it... maybe a fresh set of eyes can spot something I am missing... 17 GAIN years, 10 LOSS years... I am over 1/2 way to the goal on timeframe, a... [ More ]

Reply #23

hypersoniq — Tue, 24 Feb 2004 00:28:49 GMT

A rudimentary breakdown by year using 8 numbers per day EVERY DAY....total you would be AHEAD is $9,384Caution tho, system is still in progress... Here is the breakdown...1977...2 LOSS (need 6)1978...10 GAIN1979...6 GAIN1980...5 LOSS1981...2 LOSS1982...5 LOSS1983...8 GAIN1984...7 GAIN1985...10 GAIN1986...2 LOSS1987...7 GAIN1988...5 LOSS1989...3 LOSS1990...6 GAIN1991...7 GAIN1992...7 GAIN1993...8 GAIN1994...6 GAIN1995...6 GAIN1996...9 GAIN1997...10 GAIN1998...3 LOSS1999...7 GA... [ More ]

Reply #22

hypersoniq — Mon, 23 Feb 2004 23:39:37 GMT

finally done part 1, all 6,859 combinations... whew... I just put 5 side by side so far, the 2 with 22 hits and the 3 with 21, here is the rough stats...IF you played each of the 5 numbers straight for $1 every day over 27 years in PA's evening pick 3, you would now have a PROFIT of $7,490 AND the ONLY year you would not have had at least 1 straight hit was 1998! It's not there yet, but I think these results are pretty impressive...too many loss years, it doesn't gel with the timeframe requi... [ More ]

Reply #21

hypersoniq — Mon, 23 Feb 2004 20:29:23 GMT

near done with the first part... so far it looks like the best performers have matched 22 drawings, the majority have matched 19, a few at 20 and 21... I may try checking intervals with just the 20+ group first... the only all-positive one in the 20's is +2-+5-+4 with 20 hits. I never looked at subtraction before... interesting results so far...

Reply #20

hypersoniq — Mon, 23 Feb 2004 19:58:28 GMT

thanks for that...I hope the hits line up in some semi-guessable sort of way...

Reply #19

KyngeRycharde — Mon, 23 Feb 2004 03:42:46 GMT

Hypersonig:I think that I may have inadvertantly given you the wrong formula, for which I apologize.First, allow me to quote from page 190 of Beating The Wheel , in which the author is describing a mathematical formula to detect bias in a roulette wheel. (I see no reason why the formula cannot be adapted to testing a system.) On the other hand, it is important to point out that the exceedingly competent Belgian gambler and mathematician Pierre Basieux prefers the more conservative [n /37] + [3.... [ More ]

Reply #18

hypersoniq — Sun, 22 Feb 2004 17:38:33 GMT

Thanks Rick!57.17... got my work cut out... best so far is 22, but I just found a second 22 straight hitter...

Reply #17

KyngeRycharde — Sun, 22 Feb 2004 17:19:32 GMT

Quote: Originally posted by hypersoniq on February 22, 2004 probability experts... are we better than random with 20 straight hits in 9,202 drawings? if not, what's the magic number to beat random playing? Hypersonig:I came across a formula in Beating The Wheel by Russell T. Barnhart on pages 188-190 which may apply in answering this question.There are two formulas. The first expresses the upper and lower limits at the 80% level, the second at the 95% level.Adapted for the Pick 3 lottery....#1... [ More ]

Reply #16

hypersoniq — Sun, 22 Feb 2004 16:59:54 GMT

Quote: Originally posted by MikeK on February 22, 2004 Based on your formula fore example the -5 can never create a potential 6, 7, 8 or 9. Is that intentional and if so why?it's like a filter using real numbers, trying to pick the zones where hits may be coming from... sort of like when you would use HLL or OEO to filter results... only making all possible combos and then placing them side by side to see where the hits are spatially...there seem to be many fans of the 111 system, I'm just tryin... [ More ]

Reply #15

hypersoniq — Sun, 22 Feb 2004 16:52:37 GMT

the system doesn't pick the same number every time... the straights are varied.the most frequent hitting combo has been -7,-8,-9... it produced straights 22 times, the numbers were...Draw#......Date......Number0361.....04/25/1978....1470396.....06/05/1978....1590517.....10/24/1978....1120756.....07/31/1979....7751044....07/02/1980....2551123....10/02/1980....3651722....09/03/1982....5222197....03/12/1984....1202427....12/05/1984....0262486....02/13/1985....4823361....12/03/1987....3354413....01... [ More ]

Reply #14

MikeK — Sun, 22 Feb 2004 16:08:40 GMT

Hyper,I see your calcucations now:-9-8-7-6-5-4-3-2-1012345678909, (($A2+B$1)-10), ABS($A2+B$1)) >99, (($A2+C$1)-10), ABS($A2+C$1)) >89, (($A2+D$1)-10), ABS($A2+D$1)) >79, (($A2+E$1)-10), ABS($A2+E$1)) >69, (($A2+F$1)-10), ABS($A2+F$1)) >59, (($A2+G$1)-10), ABS($A2+G$1)) >49, (($A2+H$1)-10), ABS($A2+H$1)) >39, (($A2+I$1)-10), ABS($A2+I$1)) >29, (($A2+J$1)-10), ABS($A2+J$1)) >19, (($A2+K$1)-10), ABS($A2+K$1)) >09, (($A2+L$1)-10), ABS($A2+L$1)) >19, (($A2+M$1)-10), ABS($A2+M$1)) >29, (($A2+N$1)-10... [ More ]

Reply #13

hypersoniq — Sun, 22 Feb 2004 15:31:36 GMT

if a number results in a number higher than 10, the 10's column is cut (the ABSolute value of a negative number is kept)...by the above I should have the same amount of straights and boxes for -7,-8,-9 AND +3,+2,+1... yet they are different (-7,-8,-9 yields 22 straights and 57 boxes... 3,2,1 yields 12 straights and 48 boxes)it works the way you mention above for that specific number, but as the numbers change, the hits by position change for the other numbers...I am over 1/2 way done, and have... [ More ]

Reply #12

MikeK — Sun, 22 Feb 2004 14:56:03 GMT

Hyper,The number 6,859 combinations (19*19*19) is definteily too high. The number is 1000 (10*10*10).Why? Because 5859 of the combinations that are already part of the initial 1000.-1 = +9, -2 = +8, -3= +7, -4=+6, -5=+5, -6=+4, -7=+3, -8=+2, -9=+1By adding 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 in each position you create all possible options.Good Luck,

Reply #11

hypersoniq — Sun, 22 Feb 2004 14:29:45 GMT

gilla, it might be a different set of numbers for each data set...there are, given the numbers -9 thru 9 (counting 0), 6,859 combinations to try...I have completed 2,888 and found 13 winning combos so far...at this point, you could have played ALL 13 picks every day straight for $1 for 27 years(9,202 draws, total cost of $119,626) and would be ahead of the game by $6,374. now, after all the winning combos are uncovered, I will put them side-by-side to see if some plays could have been avoided ba... [ More ]

Reply #10

gilla — Sun, 22 Feb 2004 13:31:37 GMT

Just started doing that hypersoniq, going to each number and adding/subtracting to get number. ex. for ga.Lotto Pro 2003 Page 12/22/2004(C)1992-2003 Data SolutionsGeorgia Pick-3 Drawing List02538) 1/9/2004 5 6 8 02539) 1/10/2004 2 6 4 304 02540) 1/11/2004 9 8 0 32402541) 1/12/2004 5 3 5 45502542) 1/13/2004 4 6 3 132 02543) 1/14/2004 8 0 0 443 02544) 1/15/2004 7 7 3 13302545) 1/16/2004 7 1 2 04102546) 1/17/2004 2 0 6 51402547) 1/1... [ More ]

Reply #9

hypersoniq — Sun, 22 Feb 2004 13:11:37 GMT

Quote: Originally posted by hypersoniq on February 22, 2004 REVISED count of system beaters... hit 19 = 9hit 20 = 2hit 22 = 1 had to correct by eliminating a few double entries... but I am only to the -5.x.x part and already have caught 12 system beaters...

Reply #8

hypersoniq — Sun, 22 Feb 2004 12:27:45 GMT

new high score... 22 straights for (-7,-8,-9) this one puts you ahead by $1,798 !pick for that one for tonight is 5-6-8count of system beaters...hit 19 = 10hit 20 = 3hit 22 = 1 comments, questions, observations, contributions, refinements, corrections

Reply #7

hypersoniq — Sun, 22 Feb 2004 11:33:13 GMT

so far I have uncovered 3 subtraction systems that win over 27 years as mentioned above... I will collect data on each, perhaps the frequencies of hits will determine when to change between them...by comparison, the most frequently occurring combo in the PA game is 293 with 18 straight appearances... had you played this from the beginning, you would be out $398...probability experts... are we better than random with 20 straight hits in 9,202 drawings? if not, what's the magic number to beat rand... [ More ]

Reply #6

goose — Sun, 22 Feb 2004 11:27:46 GMT

formula is the key no matter what kind of analysis you use

Reply #5

hypersoniq — Sun, 22 Feb 2004 11:25:25 GMT

the position concept seems to be headed in the right direction, until now, the best single-system performance I have backtested has been the 7th rundown of the julian date+number drawn, which produced 15 straight hits in 27 years... now I have uncovered one that caught 20 straights in 27 years... the significance is mind-boggling...IF you discount playing boxes and played using this system for $1 straight back in 1977 EVERY DAY until today ,you would be AHEAD by $798 (assuming PA's numbers and... [ More ]

Reply #4

goose — Sun, 22 Feb 2004 01:34:19 GMT

position working system is part of my research and its to be finished within half a month but its for all pick5 and pick6formulas aint that useful for pick3 and pick4.i dont know how to crack it yet

Reply #3

JAP69 — Sat, 21 Feb 2004 23:39:06 GMT

Quote: Originally posted by JAP69 on February 21, 2004 What your description of what you want a system to do I already have. I can enter one draw or the whole database of previous draws to pull straight position hits. It has pulled the winning number on one pull or has taken as many as thirty or more or not at all for that day. It hits straight and if it misses the straight position it may hit box. I use it for the three game,four game,jackpot games. No figurin, no ciphering, just pull the amoun... [ More ]

Reply #2

hypersoniq — Sat, 21 Feb 2004 23:15:36 GMT

To give you an idea of what I am currently working on...1. I have been able to discount the Longest Out by position as a viable way to select numbers based on the above criteria... It satisfied the requirement of working isolated by position, but only produced 13 straight hits over 27 years, and there were gaps of 5 years twice... Last year was a good year for that system ,producing 2 straights after a 5+ year dry spell, so if you play that way in PA, it might catch a hit yet this year.... One o... [ More ]

Reply #1

JAP69 — Sat, 21 Feb 2004 23:15:09 GMT

What your description of what you want a system to do I already have. I can enter one draw or the whole database of previous draws to pull straight position hits. It has pulled the winning number on one pull or has taken as many as thirty or more or not at all for that day. It hits straight and if it misses the straight position it may hit box. I use it for the three game,four game,jackpot games. No figurin, no ciphering, just pull the amount of plays you want and play em.

Prediction Discussion: Round 2

hypersoniq — Sat, 21 Feb 2004 22:31:21 GMT

This is a continuation of the now massive thread found here...https://www.lotterypost.com/thread/76507.htmactually, it's more of a goal redefinement also...We need to more clearly focus on what it is that needs to be done moving forward in the prediction realm... Perhaps prediction is the wrong term to be using... Educated guess would be more appropriate...THe GOAL: To develop a pick 3 positional prediction (straight shooter) system that will win more than is invested over a defined time period... [ More ]