Mega Millions discussion

Mega Millions discussion

Postby rich » Wed Oct 14, 2009 5:44 pm

I would like to open a discussion regarding strategies for the MM lottery. I have been using my own spreadsheets for a number of years and using EL Pro for about 2 months. The most successful strategy I have employed to date is based on several factors.

First is what I refer to as relativity. This employs the frequency each inidividual number is drawn relative to every other number in each position. For example: the number one in position 1 occurs with numbers 2 through n in postion 2, the numbers 3 through n in position 3, and so on. I have found this to work much more successfully when the statistical data (numbers in the previous draws) are listed in the order they are drawn rather than sequentially.

Second is what I refer to as proximity. This function is similar to checking a generated package against previous draws. The difference is that it also takes into consideration the 'skip' of each number that matches previously drawn numbers. Using this as a comparison, I am able to utilize fewer previous draws as my statistical base. Generally this number is between 36 to 117 previous draws.

From that point, I utilize the sum and standard deviation filters in ELPro then run a new comparison of the package against previous draws. You can imagine that the packages at this point are still rather large. Utilizing the optimizing algorithms in ELPro will minimize the package, however, they seem to randomize the "leave" more than I would like.

Hitting 3+0, 3+1, and 4+0 per 250 to 400 selection sets is pretty standard. Occassionaly I will see a 4+1 (with a great amount of celebration afterward). I am wondering if there can be a method for implementing skewness and kurtosis in the optimizing utility...
rich
 
Posts: 4
Joined: Fri Oct 09, 2009 9:13 pm

Postby kdsjeter » Wed Oct 14, 2009 9:06 pm

Are you playing with real money? If, so then are you saying you have a least 1 $10,000 hit?

Maybe if Stan sees your first post he maybe able to integrate skewness and kurtosis in EL Version 5.

Personally, I feel the Mega Millions and Powerball are "Ultra Tough Nuts" to break requiring lots of processing power to filter.

When I play Mega Millions I try to make sure to cover a lot of patterns such as:

1 # From 1 - 10
1 # From 11 - 19
1 # From 20 - 29
1 # From 30 - 39
1 # From 40 - 49


2 # From 1 - 10
1 # From 11 - 19
1 # From 20 - 29
1 # From 30 - 39


2 # From 11 - 19
1 # From 20 - 29
1 # From 30 - 39
1 # From 50 - 56

Of course there are many more. Good Luck.
kdsjeter
 
Posts: 162
Joined: Tue May 30, 2006 2:32 am

Postby stan » Wed Oct 14, 2009 9:38 pm

[quote=rich]
I am wondering if there can be a method for implementing skewness and kurtosis in the optimizing utility...
[/quote]

it depends what you mean by 'skewness' and 'kurtosis'. but if you're referring to the optimizer filters in expert lotto than their algorithm quite strict.
i haven't seen that code for some time but you should get different (possibly better) result after shuffling the package content first.
Expert Lotto Team
User avatar
stan
Site Admin
 
Posts: 6338
Joined: Thu Sep 23, 2004 1:01 pm

Thanks for getting back to me.

Postby rich » Thu Oct 15, 2009 9:28 pm

KDSJester,

Typically, I will run the spreadsheets/EL Pro for several drawings to get an idea of the limitations I will be faced with regarding expense. This also gives me some time to save money for the purpose of playing. Yes, however, 3/0, 3/1. and 4/0 are pretty standard. Like I said before, the 4+1 was followed by a great deal of celebration. That was some time ago.

MM and PB lotteries are really no more difficult than any other pick 5. Yes there are more numbers in the primary set and then there is that nasty little bonus ball. Here is a thought on the bonus ball/power ball. Typically I will end up with between 150 and 450 selection sets. I take the bonus balls/power balls from the previous 200 drawings and sort them in ascending order. Then I will start with the number that was drawn most recently and work backwards, putting the numbers from the sorted list with the selection sets I plan to play. Once I have gone through the sorted list of numbers, if I have more selection sets, I just start over again. Remember your trig here...N=log(1-n)/log(1-p)...or in other words, with a degree of 90% certainty, the number(s) that will be drawn have been drawn within the last 17 draws. When it comes to selecting a bonus ball with multiple selection sets, it really is all about coverage.

Stan,

By skewness, I am referring to the assymetry of the distribution around the mean. Positive skew would put the distribution to the right of the mean, negative to the left. By kurtosis, I am referring to peakedness of the distribution. I think though, that you have come pretty close to this with the sum/difference filters. The only formulae I have for these now is from the MS spreadsheet application(not naming any names here). I believe that openoffice.org has an open source math application now.

I am not quite sure what you mean by 'shuffling the package content first'. Wouldn't that result in a more random (wheeled) distribution? Perhaps I am misunderstanding what you meant. Would you please elaborate?

Thank you both again for getting back to me.
rich
 
Posts: 4
Joined: Fri Oct 09, 2009 9:13 pm

OOPS>>>

Postby rich » Thu Oct 15, 2009 9:37 pm

I should have stated, with regard to skewness, that negative skew puts the distribution to the RIGHT of the mean and positive to the LEFT.

Coffee would be wonderful right now...
rich
 
Posts: 4
Joined: Fri Oct 09, 2009 9:13 pm

Postby stan » Fri Oct 16, 2009 5:25 pm

[quote=rich]
By skewness, I am referring to the assymetry of the distribution around the mean. Positive skew would put the distribution to the right of the mean, negative to the left. By kurtosis, I am referring to peakedness of the distribution. I think though, that you have come pretty close to this with the sum/difference filters. The only formulae I have for these now is from the MS spreadsheet application(not naming any names here). I believe that openoffice.org has an open source math application now.
[/quote]
i'll check openoffice.org calc then.
the only problem with medians is that i need to either calculate all values twice or stored the calculated values somewhere. both ways take processing time and/or free memory.

I am not quite sure what you mean by 'shuffling the package content first'. Wouldn't that result in a more random (wheeled) distribution? Perhaps I am misunderstanding what you meant. Would you please elaborate?

we might be on different tracks here:)
i just wanted to point out that coverage filters (optimizer/minimizer/best tickets) yield slightly different results when the package content is shuffled first.
Expert Lotto Team
User avatar
stan
Site Admin
 
Posts: 6338
Joined: Thu Sep 23, 2004 1:01 pm

Good morning Stan.

Postby rich » Sat Oct 17, 2009 3:42 pm

we might be on different tracks here:)
i just wanted to point out that coverage filters (optimizer/minimizer/best tickets) yield slightly different results when the package content is shuffled first.

I am going to spend some time this weekend running the different sorting scenarios utilizing the shuffling method you're referring to. I will post some results afterward. I am also going to take a look at WolframAlpha.com. I understand they may have a math app as well.

Last night was rather disappointing. 231 selection sets yielded (1)3/1, 3(3/0), 2(2/1), and 3(0/1).

Have a great weekend.
rich
 
Posts: 4
Joined: Fri Oct 09, 2009 9:13 pm


Return to Tips + Tricks

Who is online

Users browsing this forum: No registered users and 3 guests

cron