what does a prediction formula look like?Prev Topic Next Topic

New Topic New Poll

hypersoniq

Pennsylvania
United States
Member #1,340
April 6, 2003
2,581 Posts
Offline

Jan 15, 2006, 2:30 pm

back to the core of development, what is a prediction algorithm? what does it look like? what are the steps needed to look at a list of numbers and try to extrapolate the next draw?
there must be something outside of the realm of lottery that can be tried...
links to sourcecode/math texts appreciated

Have an EXCELlent day!
RJOh

100

mid-Ohio
United States
Member #9
March 24, 2001
20,272 Posts
Offline

Jan 15, 2006, 3:09 pm

If you can define an algorithm that cover 75% of the previous winning combinations, then you probably have a 75% chance of extrapolating the next winning combination. The problem will be finding an algorithm that extrapolates a group of combinations that you can afford to cover.

Some years ago I read about a Canadian lottery investment group that claimed they had done that with one of their 649 lottery which included about 300,000 combinations and was looking for players to invest $10,000 in their syndicate. I checked the website and they had some impressive data about their previous successes, but nothing about their algorithm because it was proprietary. I think it was some kind of scam or their success would have made the news by now.

* you don't need to buy every combination, just the winning ones *
time*treat

United States
Member #13,130
March 30, 2005
2,171 Posts
Offline

Jan 15, 2006, 3:45 pm

My thoughts are that the answer depends on (at least) three things.
1) type of game. I think p3/p4 types are a different beast than the lotto type games. And the worst are those where you can't pick all your own numbers.

2) How often you are trying to win. A 'win-weekly' system for p3 may cause more grief than a once-a-month-and-done system for p4.
3) How much the spend per try. Can't have your cash run out today and your number fall tomorrow.

In neo-conned Amerika, bank robs you.
Alcohol, Tobacco, and Firearms should be the name of a convenience store, not a govnoment agency.
hypersoniq

Pennsylvania
United States
Member #1,340
April 6, 2003
2,581 Posts
Offline

Jan 15, 2006, 4:14 pm

I'm not looking to win anything just yet... I want to see the basic mechanics and mathematics of prediction in general... I think the interesting aspects of using a computer to help out would go smoother if we all understood what goes on behind the scenes... such as the formula used for simple extrapolation (still searching)
simple steps...
I would like to see the steps involved, break them down into individual steps, apply each step in excel to see how things progress... p3/p4/jackpot are different animals, but I think the basic fundimentals of looking at a list of numbers and generating what is most likely to be next is a way under-developed part of what everyone is doing with systems on the LP.

Have an EXCELlent day!
popeye

Canada
Member #2,352
September 19, 2003
42 Posts
Offline

Jan 16, 2006, 12:03 pm

It would be interesting to know what method Lottosync uses. I think it is based on Strange Attractors which makes a lot of sense since it is randomness used to find randomness. I think he uses the differences between numbers rather than the actual numbers and then uses the values found to determine the best possible numbers that are in the sequence. But how he uses those differences to generate the attractors is a mystery.
RJOh

100

mid-Ohio
United States
Member #9
March 24, 2001
20,272 Posts
Offline

Jan 16, 2006, 1:23 pm

If as hypersonig stated in his post, "he's not looking to win anything just yet" then Lottosync might be a good candidate for his research, obviously the Guru put a lot of thought into his prediction algorithm before writing his program.

* you don't need to buy every combination, just the winning ones *
time*treat

United States
Member #13,130
March 30, 2005
2,171 Posts
Offline

Jan 16, 2006, 2:11 pm

Some people can't understand how there can be more than one system that works on the same game, so what you wind up with is fighting this idea that there is ONE WAY, ONE METHOD. Some methods(filters) are in direct conflict with others. I have spent plenty of time developing filters that I eventually tossed, not because they were bad in themselves, but because they didn't play well with some of the other filters.

In neo-conned Amerika, bank robs you.
Alcohol, Tobacco, and Firearms should be the name of a convenience store, not a govnoment agency.
paurths

Switching between Fairfax, VA and Belgium
Belgium
Member #19,286
July 29, 2005
2,254 Posts
Offline

Jan 16, 2006, 2:27 pm

what does a prediction formula look like?
The question lots of us asks ourselves. You can step up to the drawingstable, get pen and paper and just stare at the wall. No formula there.
"This number follows that number", "use these groups to get hits from that group", "track the double return warning", "throw 3 darts at your screen", "if it's raining play this, if the sun shines play that unless it's snowing in Key West", etc. etc. etc. ... there are plenty of "systems", and some people seem to have success with them, so that's good for them.
The posts where the winnings are in big are all over the forum, so i guess there must be more systems then there are numbers. It seems everybody is making truckloads of money... whenever some number hits.
So, i for one have no idea what a prediction formula look like, i'm not amongst the truckloads-winners, but it seems to me that there are plenty of people here that do know. All you/we have to do now is sit and wait, until they come forward and type the formula down in this thread.
Good thread, hypersoniq, just like the other one in "lottery systems". Your work is highly appreciated!
Cheers,
Ricky
hypersoniq

Pennsylvania
United States
Member #1,340
April 6, 2003
2,581 Posts
Offline

Jan 16, 2006, 9:18 pm

I am doing initial research into pattern recognition... some sites say that it cannot be taught, but it can be enhanced by an understanding of math...
for example
9-18-27-A?-45-54-63-72-81-90-B?
knowing multiples of 9 (or just being able to add 9) makes it obvious that
A=36
and
B=99
here's another
A-B-C-C-D-A-B-C-C-D-A-B-C-C-??
??=D
in each case, however... what were the individual steps involved? we all just "got it" without following any steps because of all those goofy papers in early elementary school with first picture patterns then numbers...
I believe the first pattern employed recognition of rote math facts... we all learned that sequence at some point in our lives and it "clicked" instantly...
the second pattern is obviously instant classification and grouping

Have an EXCELlent day!
RJOh

100

mid-Ohio
United States
Member #9
March 24, 2001
20,272 Posts
Offline

Jan 17, 2006, 1:22 am

In defining and checking a prediction formula or system, it must pick a group of combinations that constantly do better than an equal amount of combinations picked randomly. Eventually it should be self supporting, that is winning more than it cost to play the combinations it picks.
I've have been working on programs to back-test lottery results and look for parameters to pick combinations for future drawings. I'm also working on a RNG that pick combinations using those parameters and as I come up with different parameters or formula I have to updated it to use them.
In my research, I've found a group of parameters that cover 40% of the possible MegaMillions combinations which included 50% of the winning combinations since the last change to 5/56+1/46. If I could find a group of parameters or a formula that covered 20% of the possible combinations and still included 50% or more of the winning combinations then I might think I was getting close to something that had a chance of predicting a winner.

* you don't need to buy every combination, just the winning ones *
hareless

Anna, TX
United States
Member #26,719
November 21, 2005
263 Posts
Offline

Jan 17, 2006, 2:07 am

I have been thinking, would it not be easier to find a prediction formula for a computerized drawing rather than a drawing with real balls? I am sure that the states that run computerized drawings do everything to keep that from happening, but under it all, there will be a pattern.
One time I was taking an attention test, which I had to sit at a computer and look at a black screen. Everyime it flashed I was to click this button. I was told that there was no pattern and it was completely random. Of course at that point it became a challange and I sat there and counted. between flashes. I hit usually right before, at the same time and very little after the flash. The counts varied by 5 secs either way.
The testor was baffled that the test could not give me a score. I told her it was because most of the time I was clicking the same time it flashed, once I discovered the pattern and how it changed. She said that was impossible because there was no pattern. I told her there is no such thing.

Anyway, I was also wondering, maybe we are looking for the answer all wrong. what about creating a drawing? Then use the code in the drawing against itself to create a prediction. I hope that made sense.
Im not sure if I am expressing that right.
You know, like "to understand a criminal mind you must become a criminal?

Melissa
hareless

Anna, TX
United States
Member #26,719
November 21, 2005
263 Posts
Offline

Jan 17, 2006, 2:15 am

Quote: Originally posted by RJOh on January 15, 2006
If you can define an algorithm that cover 75% of the previous winning combinations, then you probably have a 75% chance of extrapolating the next winning combination. The problem will be finding an algorithm that extrapolates a group of combinations that you can afford to cover.

Some years ago I read about a Canadian lottery investment group that claimed they had done that with one of their 649 lottery which included about 300,000 combinations and was looking for players to invest $10,000 in their syndicate. I checked the website and they had some impressive data about their previous successes, but nothing about their algorithm because it was proprietary. I think it was some kind of scam or their success would have made the news by now.
Would that algorithm work better undefined (in a sense)? Maybe a evolving, changing algorithm?
Melissa
RJOh

100

mid-Ohio
United States
Member #9
March 24, 2001
20,272 Posts
Offline

Jan 17, 2006, 5:33 am

hareless,
Before a system or algorithm can evolve, it has to exist, that's the reason I test and try to improve what I already have. I have no idea how to create algorithm that would evolve on its own, I'll leave that to the people who study AI.

* you don't need to buy every combination, just the winning ones *
hypersoniq

Pennsylvania
United States
Member #1,340
April 6, 2003
2,581 Posts
Offline

Jan 17, 2006, 6:01 am

AI is a good place to look... machine learning "for dummies" apparently hasn't been published yet, I have been looking...
mostly AI for pattern recognition leads to "black-box" artificial neural networks (ANNs)
I tried to implement this once with my old powerball statistics excel sheet... there were four particular excel functions I was using on the history: MODE(), MEDIAN(), AVERAGE() and FORECAST()
I had an average function that would take each pick and add a "weight" component... this weight was the number of times over history that the particular stat was correct. the formula ended up looking like this
=ROUND((((L351*C351)+(L352*C352)+(L353*C353)+(L354*C354))/C356),0)
where L351 was the "pick" using the MODE command, C351 was the number of times MODE was right over the draw history
L352 was MEDIAN, L353 was AVERAGE and L 354 was FORECAST
C356 was the SUM of C351-C354 this kept the weights in line (so the pick would stay within draw boundaries)
this formula was copied for each position (5WBs and 1 PB)
for an example
pick weight
2 24
7 25
1 31
8 27
sum of weights=107
becomes =ROUND((((2*24)+(7*25)+(1*31)+(8*27))/107),0)
answer is 4
of course this just ends up being an elaborate "sliding average" and did not pass backtesting, but it did represent one of the core neural net concepts of adding a weight to an input

Have an EXCELlent day!
RJOh

100

mid-Ohio
United States
Member #9
March 24, 2001
20,272 Posts
Offline

Jan 19, 2006, 12:07 pm

I didn't know you could put AI modules in Excel or write one using Excel. I would have thought that rather than just being a "sliding average" it would have produced or suggested an action based on historical data.

* you don't need to buy every combination, just the winning ones *

New Topic New Poll

Subscribe to this topic

what does a prediction formula look like?Prev TopicNext Topic

what does a prediction formula look like?Prev Topic Next Topic