Statistics, weighting and scope....

Reply #16

hypersoniq — Fri, 11 Jun 2004 01:17:49 GMT

oh ,yeah... I chose the above stats because the first 3 allow mw to accurately determine skew and the last is a nice little self-contained complex algorithm built into excel.

Reply #15

hypersoniq — Fri, 11 Jun 2004 01:16:52 GMT

my problem is twofold...1. including stats with no reason for them being there2. not backtesting due to complexity of pick generationI have fixed #1... I just have MODE, Mean, Average and Forecast now, both full history and optimized last 5-10 draws... AND I have finally concocted a backtest sheet that I am still working on to see how this complex pick method has done over time AND to develop an error-correction mask for the overall pick. I have to test friday (M6) and Saturday (PB) but I finall... [ More ]

Reply #14

MrSteel01 — Fri, 11 Jun 2004 00:50:10 GMT

Quote: Originally posted by hypersoniq on May 14, 2004C. test results...My pick...03-05-14-19-31-41Results...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=113Results...01-07-19-23-33-44=126Whats the problem

Reply #13

hypersoniq — Sun, 06 Jun 2004 14:02:29 GMT

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... [ More ]

Reply #12

hypersoniq — Sat, 05 Jun 2004 21:12:49 GMT

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 e... [ More ]

Reply #11

johnph77 — Mon, 17 May 2004 01:34:14 GMT

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

Reply #10

hypersoniq — Sun, 16 May 2004 13:44:14 GMT

results...Test cost $10Return on picked numbers... $0Return on QPs... $9 (Game A. 3/6 + 4/18 Game C. match 4/18)net loss $1system 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

Reply #9

hypersoniq — Sun, 16 May 2004 13:25:00 GMT

D. test results...My picks...Best Performer (full history)... 03-15-28-41-49 PB 28after 10M draws (Average)... 08-18-27-33-43 PB 19Results...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...

Reply #8

hypersoniq — Sun, 16 May 2004 02:14:19 GMT

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 Powerball03-15-28-41-49 PB 28note 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

Reply #7

hypersoniq — Fri, 14 May 2004 23:09:13 GMT

C. test results...My pick...03-05-14-19-31-41Results...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...

Reply #6

hypersoniq — Fri, 14 May 2004 22:56:00 GMT

C. Test pick... 4 min. before draw time...03-05-14-19-31-41 (PA Match 6

Reply #5

hypersoniq — Fri, 14 May 2004 00:44:09 GMT

B. test results...My picks...Best Performer (full history)... 03-15-26-41-49 PB 27Combined (Full + wednesday only)... 07-16-27-35-46 PB 23Results...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

Reply #4

hypersoniq — Wed, 12 May 2004 00:14:22 GMT

A. test results...My pick...04-07-16-18-32-41Results...06-07-15-31-38-43(-2)(Hit)(+1)(-20)(-6)(-2) [individual numbers](Hit)(Hit)(Hit)(Miss)(Hit)(Hit) [decades

Reply #3

hypersoniq — Tue, 11 May 2004 08:50:56 GMT

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 jackp... [ More ]

Reply #2

ayenowitall — Mon, 10 May 2004 07:45:17 GMT

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.... [ More ]

Reply #1

hypersoniq — Sun, 09 May 2004 11:47:14 GMT

no big bux (but Someone in PA won it) ,but how is this for a ticket...02-09-17-33-43 PB 173 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 40results 03-09-17-37-43 PB 39

Statistics, weighting and scope....

hypersoniq — Sat, 08 May 2004 01:48:59 GMT

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 35in the last 16 games only (short term memory), same stat... [ More ]