Algorithms for number selection?

RL:

Yes, you are so right and this is exatly what has been getting me frustrated!  After the draw comes out it seems so simple but before the draw just too many numbers.  You might have an idea of studying the igits due instead.  Might be worth looking into as it can't hurt as what I'm doing now doesn't help.

Will keep grinding away at it.

Thanks for your help as it makes me feel good that someone else sees my logic in trying.

M

No lottery jackpot winner has ever regretted the time and effort he/she spent trying to win it.

I tried a new strategy today with my PB pick. I did an individual calculation for each number...I picked the first number based on the first number drawn last time and compared it to the last 10 (up one or down one based on an average, etc)....then picked the rest of the numbers the same way.  One of the numbers seemed to be almost always connected with multiples of another number. Gave me an interesting set of numbers I would never have picked.

that is pretty cool and I also think that numbers are connected in ways that we just somehow dont really see. I was once working on on the distance of numbers and releating them to decades its something most people will blow off as a waste of time but I do see releationships between numbers in this way.

Keep up your research its something i would be really interested in.

Writing an algorithm will be easy once someone has done the work of coming up with the variables that are related to the winning numbers in the drawings at least 75% of the time. 

Also those variables much target a group of numbers small enough that most of the possible combinations could fit in the budgets of a small pool of players or a single player who doesn't mind spending a little money.  Even if they don't win a jackpot, most of the cost sould be covered by the smaller winnings such that they can afford to continue trying to win a jackpot until they actually do.

Apparently, there is no program that does that.

writing an algorithm that a computer can run is not necessarily easy even if you know the variables...

i observe numerous variables...but i dont know how to write a algorithm...

not in VBA, not in C+, not in Java,

and variables move in trends also...so there is no set of variables that are reliable 75% if time...it is up to the player to "time" the variables appearance.

the more important question is how often does a specific set of variables produce a jackpot in say 3 months or 6 months time...z(on average)

and how many variables is too little...

how many variables is too much...

what are more important variables...

i have seen the simplistic approach go right out the window in so many threads...to pi in pick 3 to lexographic clustering...

i must admit i am also guitly of losing the simplistic approach quite often...but i think i found it again...sometimes it takes a long losing streak and allowing your self to become detached from the result of the draw for the answer to appear...and many times reappear...

You don't have to use a computer to come up with variables or use them, it just that using a computer might make the process faster.  Besides if you come up with something that shows promise, the potential rewards could be an incentive to learn what ever more you think you need to know to use a computer to do the work.

I'm experimenting with ways to come up with variables that are reliable 75% of the time using the local game (6/49) but the targeted number groups are too large to get a winner in 10 lines as was obvious in my Saturday predictions. 

I actually played 20 lines and had hits in lines 17 and 19 that will cover 20% of my cost.  I checked the local games of Indiana (6/48) and Michigan (6/47) and would have covered 40-60% of my cost scoring the same hits in their games.

Covering the cost with the small wins is an important part of the strategy of being able to afford playing the number of lines using an algorithm would require until a big win.

The variables I'm using aren't fixed for every drawing but are similar when the conditions leading up to next drawings are similar so I have to go back through all the previous drawings to find similar conditions and the numbers that worked then.  Needless to say, a good data base of previous drawings for the game you're playing is required.  Of the two multi-state games MegaMillions and PowerBall, MM has the best history to exploit since it has 719 drawings with its present matrix.  I'll be concentrating on MM again when its jackpot is higher.

My experiment is a work in process and I hope I will be able to post a jackpot winning prediction one of these days, winning a jackpot would be a bonus.

After observing some of the types of wins that have occured lately with MM and PB where one person or group played the same combination several times with different bonus numbers and winners having several other lines that won seconds I not sure that's true.  Apparently there are strategies that some players have developed and maybe even programmed but they are proprietary.

Hello, RJOH a lotto 49/6 divided by 20 positions to see the suit 1,2,3 1,2,4 1,2,5 1,2,6 ..... = 4,5,6 are 20 possible positions One could vera likelihood of repetition in each of the 20 positions   Example suit 05,12,45 = the best position to bring this suit? Then they could see which of 20 possible lottery 49/6, you could play as fixed For a calculation of probabilities of the positions is not on the forum The frequency of repetition of certain suits for use in the same positions as fixed!

Hello dr san,

I consider the validity of ones posts by the ratings of their predictions.  Since you don't make predictions, your posts lack any significance to me.  Good luck to you.

LB

I agree, there are the simple and then there are some that are very complex.  I just added a couple

suggestion algorithms for the lower level filters and was suprised that it's hitting 80% and up for

my 5-39.  The problem is that I still have to find the 20% which are not correct so I don't know if

there is any avantage to using it.  I don't much believe in prediction but I am open minded to the

possibility that more can be done.  I started a new algorithm for my group settings and it's proving

to be a hair puller.   I am getting ready to release a big-game version of my program but have been

trying to add a few features to make setups for the big games a little easier.  Both MM and PB are tough

games for any system programmer and I will have to wait and see how it does.  From my experience the

simple approach seems to always do better.   I am a little suprised that no one has offered any real

information on this subject.  I have found a few PDF's online that look interesting but it will take me a

while to code them.  Copy, paste into google should get you there if your interested.

#1 Minimax vs bayes by D. Blackwell, Berkeley

#2 A simple randomized algorithm for consistent sequential prediction of ergodic time series.

It would be nice if could find some prebuilts that I could download and start experimenting.

RL

RL

You don't have to believe in prediction to use the prediction board, it can be an excellent independent testing facility for evaluating theories and strategies for winning lottery games.  Using the prediction board is as close as you'll ever come to applying those theories and strategies in a real lottery without actually buying tickets.

RJ

There is a simple approach. When you flip a coin, what is the next outcome?