It works!

[stan]

Stan, any suggestions?

you can try an alternate strategy which doesn't aim for the jackpot - make your history estimates as usual but allow for errors, e.g. by saying that 'at least 9 to 11 levels must pass'. then filter a representative set of tickets using your estimates, analyze tickets remaining in the package and take the first 10 or 15 most frequent numbers. then put them into a wheel or create your own combinations manually.

[MartinMan]

Ameriken,
I have had good luck using the Optimizer by looking for tickets that are 1 less than the number of positions drawn.
If your lottery is 5 numbers drawn, then Optimize for Quads etc.
If you get your numbers down to the amounts you have been showing, then it will keep a large percentage of the higher tier prizes. At least it has for me on a 5 of 39 and on a 5 of 56 lotto's that I play.
To test it just test your tickets, save them, then use the Optimizer and test them again.
I have been Very happy with the results on larger packages. I have not tried it on very small packages as it tends to Not remove many tickets on those.
But on larger packages, I generally see roughly a 50% reduction in tickets with high numbers of retained winners. Thats on 12 to 50 K numbers.
It has Not worked well for me when Optimizing for Trips or Pairs, as far as retaining winners. At least not in my games. Try them on yours. I think the problem is the much reduced number of tickets that you get.
It is just to hard for Anything to work consistently well when picking small amounts of tickets.

Regards,
Martin

[sherlock]

Hi Ameriken

Regarding your method(s) for arriving at your very impressive wnh filter estimates, could you create a pdf document containing a series of annotated screen shots that would 'talk' the reader through how you arrive at your estimates? I'm not very good with estimates for this particular filter, and would be very interested to see how you get such good results.

Regards

Sherlock.

[Bobijohn]

Hi Stan

Thanks for the quick and detailed response.

Very happy to hear that the "Optimizer", "Fast Minimizer" and "Minimizer" will "probably" be in v5.2

Regarding the Package Management routines (Perl Library issues etc) it is only fair for me to say that I am no hurry for those at this time since my Lottery endeavors have been severely curtailed due to other more pressing matters. However I will certainly write up a request at the appropriate time.

You have done great work on version 5 --> Thank you.

Bobijohn

[ameriken]

Thanks for your replies everyone. I'll try and figure out how to get a screen shot in, Sherlock. I'll give you the basics however.

If you set your winning numbers anyalyzer to show 'positive differences' and 'negative differences' it will tell you where the majority of values are, like how far up and how far down the sums move. For my games, 0 to 20 usually accounts for 60 to 75% of the changes.

Then in the menu I go to 'History' then 'Summary' and try to make a guess which way and how far the sums will move up or down. It's really trying to make a guess as to which way I think it is going to go and by how much. So, for my games, I may go -4 +16 for the first level if I think it has a better chance of going up, or -16 +4 if I think it will go down. There's no magic to it, it is really just a guess. I turn on the yellow 'average' line and that helps me when viewing the charts so I know where the center lies. I just look at past history, patterns, the direction it has been moving, and try to make a guess as to which way and how far I think the next sum will go. I also turn on the 'pink and blue' coloring in 'History and Differences' so I can get another visual as to how far up and how far down the sums are going and again by how much. If I see a value has decreased for 7 or 8 draws, I try and take advantage of the bias and predict it will go up in the next draw, and in a case like that I may go 0 +20.

If I can get 8 or 9 out of 11 correct, then I can usually score one or two 4/5 wins in the package, and 20 or 30 3/5's. I don't always hit and when I do, I sometimes lose good numbers through the filtering process.

So it's not an exact science, and I am far from perfect, but so far it is the best method I've ever used. All the other methods I've used are continual misses, whereas the WNH sums gives me hits.

I'll try and get those screen shots on here so you know what parts of the software I am using. Good luck!

[ameriken]

OK, I hope I did this print screen stuff right. These are the two screens I use. I don't know why it was cut off when I put it here, but on the right side would be the 'my estimates' screen where you actually adjust the levels up and down.

[scenium]

Hi Ameriken,

Your History Sums have quite a consistent range of ODD/SUM ( 6:5,5:6,4:7,7:4),
maybe you can try the History Sums ODD/SUM filter.

[ameriken]

Hi Ameriken

Regarding your method(s) for arriving at your very impressive wnh filter estimates, could you create a pdf document containing a series of annotated screen shots that would 'talk' the reader through how you arrive at your estimates? I'm not very good with estimates for this particular filter, and would be very interested to see how you get such good results.

Regards

Sherlock.

Shirlock, to do what you asked would take me a bit of time that I don't have, and there really is no special method I use in arriving at my predictions (other than what I mentioned in my last post). In fact, I often ask myself 'how did I do that yesterday and can I duplicate it'? I can't...it all comes down to a 'best guess'. I just look at the patterns in the chart and look to past patterns to see how they behaved and if ther eare any similarities. Then I try and make my best guess on which way it may go. I do miss a lot as well, so I as said earlier it is not scientific. Sometimes I will do two separate predictions for a draw and play #'s from both sets. I also keep track of all the predictions I make, and I save the packages so i can go back and 'debrief' myself and see where I was right and where I was wrong. When I know the winning numbers, I'll run the package again and try and filter it differently to see if I could have trapped more of the better combinations.

It's really a lot of experimenting and trial and error. And of course, some good ol' luck.

[daniel]

It's really a lot of experimenting and trial and error. And of course, some good ol' luck.

Hi ameriken,
For predicting WNH you might try the shareware aiNet as it uses a form of Probabilistic
Neural Network which requires no learning phase, is supposed to be good with noisy data, and
provides an estimate for the rate of error in it's predictions.
The downside is the amount of time it would take!

http://www.filedropper.com/ainet124manual

http://www.filedropper.com/ainet125

If you have a spare $595 to spend you could try these..

http://www.wardsystems.com/predictor.asp

http://www.wardsystems.com/classifier.asp

daniel

[sherlock]

Hi Ameriken

Thanks very much for your explanation on the techniques you use for estimating the wnh level directions and ranges, and
also for the screenshots.

If you are using renges of 20, and attaining 8 to 11 levels correct, then that is a really good result.

I'll work through your comments and see how I get on. I'll also try exporting the wnh data to metatrader 4, and see
if that helps with the analysis - MT 4 does have a lot of visual charting tools that can help measure
deviation, range of movement,turning points, etc.

Many thanks

Sherlock

[ameriken]

Hi Sherlock, I haven't tried the meta thingie yet. I think Stan said something that gave me the idea that it wouldn't be for me. I may check it out sometime.

Anyhow, good luck in your predictions. It's not perfect and doesn't work everytime, but like I keep saying, I've gotten better numbers out of the WNH sums than anything else I've tried. So that is all I do now. BTW, I see I made an error, I get 7 to 10 levels, not 11, havent done that yet.

[ameriken]

Here are the packages I played tonite. Not a bad nite, and again, better than I've done using any other method.

1) In MM, WNH gave me 92 #'s and there waere 2 3 +0's in the package. Fortunately I played one of them. I didn't match the powerball, bummer, if I did I would have won $150. I use a totally different method in picking my bonus #'s and missed tonite.

2) In Matchplay, WNH have me 796 lines that had 11 4/6's and 144 3/6's. I picked 10 lines and got 1 3/6. I usually don't like this game that much: for $2 you get 3 lines. You get to pick the first line, but the other 2 are QP's. I never get 3 or more matches in a QP line, only in my SP's. This is why I don't play QP's.

3) Cash 5 I got a 450 line package, which contained 21 3/6's. Tonite I was fortunate to have 2 in my final package of 30 numbers played. Not really a bad night and again, better than anything I've done before.

MegaMillions Winning #'s- 10-23-39-41-45 +15

1 3 8 44 56 3
1 3 27 47 56 17
1 6 13 22 44 21
1 17 43 44 56 22
1 22 23 31 52 46
3 9 20 47 56 3
3 9 27 36 56 17
3 16 41 45 48 21
3 17 41 45 48 22
3 20 46 47 56 46
3 27 36 46 56 3
4 12 23 26 52 17
6 7 39 48 54 21
6 23 30 34 44 22
8 27 30 34 56 46
9 18 36 40 54 3
9 20 23 30 50 17
9 27 30 36 50 21
13 27 31 42 45 22
17 28 41 45 48 46
18 30 33 36 45 3
20 27 32 45 52 17
23 27 28 41 45 21
27 30 34 47 56 22
27 34 42 43 50 46



Matchplay: Winning #'s 7-16-22-24-31-32

3 6 15 22 31 33
3 11 22 26 33 36
3 16 19 28 31 37
4 12 22 23 26 31
4 14 23 25 31 34
6 13 14 22 25 31
7 11 19 28 36 37
7 14 23 27 31 34
10 15 23 25 31 36
13 16 19 23 31 32


Cash 5: Winning #'s 9-14-19-20-27

1 4 13 19 30
1 4 19 25 30
1 4 22 23 26
1 7 21 27 30
1 10 21 29 30
2 4 13 19 30
2 7 9 22 27
3 4 8 22 32
3 4 13 22 24
3 4 19 22 32
4 7 8 16 30
4 7 16 22 30
4 9 11 20 28
4 9 16 18 19
4 9 16 22 25
4 10 11 30 32
4 11 12 22 30
4 11 16 30 31
4 11 18 22 30
4 11 21 25 30
4 11 22 30 31
4 12 22 23 31
4 16 18 19 30
4 16 19 28 30
5 9 19 22 27
7 8 13 27 30
7 9 21 22 27
8 9 18 21 27
9 13 22 27 28
9 18 19 27 32


I still can't believe how the WNH filter works. I am totally convinced that over time, one can win a jackpot using this, and I think this is the only prediction method out there where you can actually find the winning numbers. EL 5 continues to enhance my lottery play. Great job Stan!

[ameriken]

This is how it's supposed to work: a package of 141 lines, which contained 3- 4/5 lins and 21- 3/5's.

Filtered it down to 45 tickets, and got 1 4/5, 10 3/5's, and 12 2/5's for a total of $312 won on a $45 purchase.

[ameriken]

04-19-11

Winning #'s 10-14-21-26-32

1 2 17 21 24
1 3 5 26 29
1 3 7 25 26
1 3 7 26 29
1 3 18 26 29
1 3 19 26 31
1 3 22 26 31
1 3 26 28 29
1 6 10 21 24
1 7 10 19 26
1 7 10 21 26
1 7 13 21 26
1 8 18 21 26
1 8 21 26 32
1 9 21 26 32
1 10 11 21 26
1 10 16 21 26
1 10 17 19 24
1 10 17 21 24
1 10 18 21 26
1 17 18 21 24
1 17 21 22 24
1 17 21 24 27
1 17 21 26 28
2 10 17 24 29
2 10 18 21 26
3 7 8 25 26
3 7 10 26 29
3 8 26 29 31
3 10 18 26 29
6 7 10 12 26
6 8 26 29 31
6 10 18 26 29
6 10 19 29 31
7 8 17 26 29
7 9 12 17 27
7 10 14 26 29
7 10 17 26 29
7 12 17 22 26
7 12 17 25 27
9 10 21 26 32
10 11 16 26 29
10 12 18 21 26
10 12 21 26 28
4 12 22 23 26

[Neener]

!!!Congratulations !!!