huge database strategy

[Dennis Bassboss]

Try Lotto-Hat ....It has it too!!!

 

[daleks]

came across a group out of M.I.T. who are into chaos theory...in the late '70's and early 80's they devised several algorithms which allowed them to make money at the roulette wheel in vegas, the roulette wheel essentially being a closed system.....they went on to the stock market in the 80's, and did quite well predicting market trends, betting on futures and options.....using a similar-to-the-roulette wheel approach gave them nothing but failure.....they succeeded by eliminating most of the models they had made, and concentrated on short periods of time only......

in lotto games, i wonder whether most filters and long past histories can be eliminated, concentrating rather on linear models and others which occur only in, say, the past 20 or 25 draws.....would that not eliminate a lot of the noise, which one is uncertain most of the time how to treat anyway,,,,and get rid of the occasional fat tails displayed in a bell curve and caused by uncertain noise.........

 

[winhunter]

Daleks...

WINHunter helps you examine whether or not more or less history works. For most of the filters, less history is better. For some filters, skipping the first few drawings works better. The ONLY purpose of the large/complete history, is to see how well that configuration would have performed if used over the entire history. You can limit WINHunter to testing just the most recent history though, so past trend changes are ignored.



Andrew

 

[daleks]

thanks, winhunter..........in the super 7 lotto especially, i have noticed groups of, say, 15 to 18 or 20 numbers active over a period of perhaps several months, to fall by the wayside a few at a time as others take their place.......have not seen that happen in either of the 6/49 draws.........i do understand the need for the large data base and the verification of models it can provide.....i am becoming a bit discouraged of the ability of any model to predict.....such discouragement is not, however, neither chronic nor endemic......merely serves to provide encouragement to tweak further..........

 

[Rebeckah]

Re: Idea

....if you could predict the direction vector of the index based on previous vectors (by charting it), you could eliminate many lines that way! Also, what if you could predict the target position, which would be a target index value that would make it fall in between two previous draw index values (this has to be the case, unless you match a previous draw EXACTLY, which is UNLIKELY.

I think this is an EXCELLENT strategy for eliminating many combinations in this strategy I brought up. This way your combo lines would be between two index values giving you maybe 10,000 possible winning combination sets to pick from instead of millions. Yea, that's a very good filter. My thinking is that if you could apply ALOT of filters to the master list then it could be scaled down to something playable? I'm really too confused & unorganized to work out this strategy right now, but I hope if somebody does it that they'll post their results. I would think you could get a guaranteed jackpot win from it. ? Right now I'm trying to learn WinHunter, MDIEditor, I'm looking at LOTWIn and upgrading from MSWorks spreadsheet to Quatro Pro 10, still trying to get Lottery Director, AND make charts by hand. PLUS I developed a whole different strategy which required making a new database and all new charts like it was a new game. busy, busy.

 

[Rebeckah]

Re: LotWin

well, the problem I've run into with LOtWin is that the database requires the year to be YYYY and my databases are YY, so I'm looking for another database so I won't have to change them by hand. So I haven't gotten very far since I'm stuck on the database problem. Sure I could use another game, but I'm sure you all know you can't really get a feel for something unless you see how it works with your game #s.

& Dennis, THANKYOU for the stats on my game! I appreciate it.

 

[Rebeckah]

in lotto games, i wonder whether most filters and long past histories can be eliminated, concentrating rather on linear models and others which occur only in, say, the past 20 or 25 draws.....would that not eliminate a lot of the noise, which one is uncertain most of the time how to treat anyway,,,,and get rid of the occasional fat tails displayed in a bell curve and caused by uncertain noise.........

I think it was from the lottery connection store website that I read their theory that you shouldn't look at data for more than 12 games back. I disagree with this on a few counts. First of all, what if a number is doing a repeat skip hit, and the skip period is 9? If you're only looking at 12 games you'd miss that this # has come up 10/20 total hits at a skip of 9. There's many trends that ebb and flow over different *skip* periods. an E/O ratio may trend in the 20s, meaning heavy on the evens for 20 games, then heavy on the odds for 20 games, then heavy on the evens for 20, etc. But you'd never even LEARN this trend if you only look at 12 games back. Looking at a short term trend IS helpful tho, for discerning short term trends. Like if a # is hot. To qoute Dennis: look at EVERYTHING.

I have considered altering databases to remove the antipodes, but still feel uncomfortable about it. It would just give me a big average soup. Which may be good, but may be too cloudy. I don't know daleks, I think it may be helpful to do as an adjunct to looking at the whole history as well. Like you have 2 seperate databases for one game. If you try it will you let us know how you do with it? I'm working on too many other things right now to do it meself.

 

[Brad]

Re: YY vs YYYY

Rebeckah,

I've run into that prob many times bfr, what works well is using Find and Replace functions with an editor such as WordPad, one can doctor up a data base fairly quickly, esp if only the YY has to be adjusted ... 10 yrs=10 changes, matter of a few minutes. Spread sheets work well too, but I'm probably not telling you anything new here.

Cheers

 

[winhunter]

Re: Re: Idea

I think this is an EXCELLENT strategy for eliminating many combinations in this strategy I brought up. This way your combo lines would be between two index values giving you maybe 10,000 possible winning combination sets to pick from instead of millions. Yea, that's a very good filter. My thinking is that if you could apply ALOT of filters to the master list then it could be scaled down to something playable? I'm really too confused & unorganized to work out this strategy right now, but I hope if somebody does it that they'll post their results. I would think you could get a guaranteed jackpot win from it.

Unfortunately, WINHunter is NOT a line filter... About the only thing that could be surmised from this type of filtering would be to score the numbers that occured within that range of draws, and then make a selection from that. And IM not really sure if that would really make sense, or even be feesible.




Andrew

 

[Rebeckah]

Re: Re: YY vs YYYY

Rebeckah,

I've run into that prob many times bfr, what works well is using Find and Replace functions with an editor such as WordPad, one can doctor up a data base fairly quickly, esp if only the YY has to be adjusted ... 10 yrs=10 changes, matter of a few minutes. Spread sheets work well too, but I'm probably not telling you anything new here.

Cheers

ah, Brad thank you. my mind has gotten clouded by the learning of new systems so much so that I had forgotten a few of the basics it seems.... hopefully that'll do it. I also realized I have to get a sw to reverse the draw order, I'll have to Ion's revigui, or whatever its called. Is that the easiest for this task?

 

[Brad]

Yes Rebeckah, that's CoolRevGui you mentioned (for others reading this) and it works well for reversing data ... you can thank thornc for developing it and Ion for providing the download ...

Cheers

 

[Rebeckah]

For a 6/54 lottery here are the expected values and the current results after 270 draws in % for the Texas draws...
––––––––––-expected––-actual values
1-1-1-1-1-1––54.1–––––––-60.4
2-1-1-1-1–––36.9––––––––32.1
2-2-1-1––––-4.9–––––––––5.6
2-2-2––––––-0.1––––––––-0.0
3-1-1-1––––-3.3–––––––––1.5
3-2–––––––-0.4–––––––––-0.4
3-3–––––––-0.0–––––––––-0.0
4-1-1––––––0.2–––––––––-0.0
4-2–––––––-0.0–––––––––-0.0
5-1–––––––-0.0–––––––––-0.0
6–––––––––0.0–––––––––-0.0

Dennis, I don't know where you got this from, but here's what I came up with: TX Lotto 6/54 last 270 games

Order Hits Percentage
1122------ 87------ 32%
11122---- 79---- 29%
1113---- 30---- 11%
123---- 26---- 9%
1112---- 20---- 7%
122---- 9---- 3%
222---- 4---- 1%
111111---- 4---- 1%
114---- 3---- 1%
23---- 2---- 0%
24---- 2---- 0%
33---- 2---- 0%
112---- 1---- 0%
1123---- 1---- 0%
111112---- 1---- 0%

 

[Dennis Bassboss]

Dennis, I don't know where you got this from, but here's what I came up with: TX Lotto 6/54 last 270 games

Order Hits Percentage
1122------ 87------ 32%
11122---- 79---- 29%
1113---- 30---- 11%
123---- 26---- 9%
1112---- 20---- 7%
122---- 9---- 3%
222---- 4---- 1%
111111---- 4---- 1%
114---- 3---- 1%
23---- 2---- 0%
24---- 2---- 0%
33---- 2---- 0%
112---- 1---- 0%
1123---- 1---- 0%
111112---- 1---- 0%

Sorry I just read that thread ...When I posted that I was under the impression that you were looking for the consecutives theorical values for a 6/54 lotto...I was rushing all week and I posted that in a hurry while in a little break without really looking at the rest of this thread...My mistake...

 

[Rebeckah]

Hold on Dennis, I think there may have been two different things going on. I went back & looked at Gilles post & it was for positional decades. Like: 1hit in one decade, 2 hits in one decade, one hit in the next, etc. thus giving a sequence 012121 or something similar. I think your data may be for decades themselves in sequence? Which would be why you have 4-5 & 6 in there? I'll post the stats I get for those once I run them.

 

[Rebeckah]

it's cool dude, I'm just glad we caught it.

hope you have a relaxing weekend.