Welcome Guest
You last visited September 25, 2017, 12:41 pm
All times shown are
Eastern Time (GMT-5:00)

# What is the minimum sample size needed to give you an edge?

Topic closed. 15 replies. Last post 4 years ago by JKING.

 Page 1 of 2

What is your minimum size for good predictions?

 10 games [ 8 ] [38.10%] 50 games [ 3 ] [14.29%] 100 games [ 3 ] [14.29%] 500 games [ 2 ] [9.52%] other [ 4 ] [19.05%] Previous games don't help [ 1 ] [4.76%] Total Valid Votes [ 21 ] Discarded Votes [ 1 ]

United States
Member #5599
July 13, 2004
1197 Posts
Offline
 Posted: December 19, 2013, 2:37 am - IP Logged

Everybody has thier own system. I am curious, what is do you think the minimum amount of previous/historical games is required for your system to work? Even if you don't have an in depth type of system, how many games back do you look at to make your prediction. Please, state the game type you are talking about (p3, p4, p5...etc).

As for me, I like to have about 150 historical games for a p5 game. It seems to give me pretty much a full range of the highs, lows and trends that will occur in a filter. Obviously, the larger the sample size the more accurate your info becomes. But, 150 seems to work for me.

You are a slave to the choices you have made.  jk

Even a blind squirrel will occasionally find an acorn.

There is no elevator to success, you will have to take the stairs.

mid-Ohio
United States
Member #9
March 24, 2001
20130 Posts
Offline
 Posted: December 19, 2013, 3:41 am - IP Logged

Everybody has thier own system. I am curious, what is do you think the minimum amount of previous/historical games is required for your system to work? Even if you don't have an in depth type of system, how many games back do you look at to make your prediction. Please, state the game type you are talking about (p3, p4, p5...etc).

As for me, I like to have about 150 historical games for a p5 game. It seems to give me pretty much a full range of the highs, lows and trends that will occur in a filter. Obviously, the larger the sample size the more accurate your info becomes. But, 150 seems to work for me.

My system is designed to pick numbers from historical data so it needs as a minimum enough history for all the numbers in the matrix to have hit at least twice.  I'm still waiting for all the numbers in the new MM matrix to hit at least once.

In the last MM drawing, all the winning numbers had hit in the previous 39 drawings but their sum of 98 was the lowest since the matrix was changed.

* you don't need to buy more tickets, just buy a winning one *

MD
United States
Member #1701
June 18, 2003
8557 Posts
Offline
 Posted: December 19, 2013, 4:05 am - IP Logged

Pick 3-4 i use only 10 draws each

Big John says. You don't hit the number. The number hits you!!!!

I'm not Big John, I'm Four4me, Big John's a friend.
u\$a
United States
Member #106665
February 22, 2011
19967 Posts
Offline
 Posted: December 19, 2013, 10:59 pm - IP Logged

0

Winning is great.
bel air maryland
United States
Member #90251
April 24, 2010
5953 Posts
Offline
 Posted: December 19, 2013, 11:12 pm - IP Logged

Everybody has thier own system. I am curious, what is do you think the minimum amount of previous/historical games is required for your system to work? Even if you don't have an in depth type of system, how many games back do you look at to make your prediction. Please, state the game type you are talking about (p3, p4, p5...etc).

As for me, I like to have about 150 historical games for a p5 game. It seems to give me pretty much a full range of the highs, lows and trends that will occur in a filter. Obviously, the larger the sample size the more accurate your info becomes. But, 150 seems to work for me.

Depends on the system I'm using at the time. I'm always looking to improve on what I'm doing so my systems are always

evolving. To answer the question, I've gone back 20 draws and as few as 2 for the P3 & P4 games. I look for whatever

the current trends are, so I don't pay much attention to whatever happened last month or 6 months ago.

"You can observe a lot just by watching." Yogi Berra, Hall of Fame baseball player.

The numbers will tell you what numbers to play. Pay attention to to the numbers.

United States
Member #128790
June 2, 2012
5431 Posts
Offline
 Posted: December 19, 2013, 11:13 pm - IP Logged

I believe what's happening now reflects the current trend.

Data from 100-150 games ago is not important to me.

I like the last two draws as a reference.

Galena,MO
United States
Member #129328
June 15, 2012
101 Posts
Offline
 Posted: December 26, 2013, 7:58 am - IP Logged

I use a 5 day chart on p3 and p4 in missouri,most of the time it shows the trends.I have noticed at times however that 6-7 days will also provide a hot or key number.

The Dream Master says:Keep on dream'en,Dream'in of a win.........Because Dreams do come true

Lottery Dreamer
United States
Member #75493
June 4, 2009
46771 Posts
Offline
 Posted: December 26, 2013, 5:23 pm - IP Logged

Everybody has thier own system. I am curious, what is do you think the minimum amount of previous/historical games is required for your system to work? Even if you don't have an in depth type of system, how many games back do you look at to make your prediction. Please, state the game type you are talking about (p3, p4, p5...etc).

As for me, I like to have about 150 historical games for a p5 game. It seems to give me pretty much a full range of the highs, lows and trends that will occur in a filter. Obviously, the larger the sample size the more accurate your info becomes. But, 150 seems to work for me.

For my predicitions I personally sample back 30 days.

Kejana

" Share your predictions "

Whiskey Island
United States
Member #90216
April 24, 2010
13142 Posts
Online
 Posted: December 26, 2013, 5:43 pm - IP Logged

All depends on the games pick 3 and Pick 4 last 20 to 50 draws . Jackpot games last 100 !

United States
Member #91718
May 23, 2010
690 Posts
Offline
 Posted: December 27, 2013, 12:22 am - IP Logged

Maybe two weeks for a sample size, I don't see a need to go back more than that unless running a test.

Michigan
United States
Member #81740
October 28, 2009
50361 Posts
Offline
 Posted: December 27, 2013, 9:11 am - IP Logged

Depends on the game and what I am looking for.

For pick 3 usually bewtween the last 10 to 30 draws.

Stand for the Flag,  Kneel for the Cross.

Park City, UT
United States
Member #69864
January 18, 2009
1004 Posts
Offline
 Posted: December 30, 2013, 8:27 pm - IP Logged

For pick3 and pick4 I like to use multiple sample sizes.  Multiple samples allow you to recursively keep eliminating combinations until you get to your playable amount.  I can use 7 sample sizes and still be 47.8% accurate.

Jimmy

BETWEEN OAKRIDGE AND WRIGHT-PATTERSON AFB
United States
Member #1647
June 10, 2003
2830 Posts
Offline
 Posted: December 30, 2013, 9:20 pm - IP Logged

Pick 3 5 or less
Pick 4 10 or less
Pick 5/quinto 25 or less

Pi is the way...

turning \$30 or less into thousands everyday!!

here we go steelers ..here we go---------stairway to seven

TIME FOR THE LOMBARDI TROPHY TO  COME  HOME..

United States
Member #124493
March 14, 2012
7023 Posts
Offline
 Posted: December 30, 2013, 9:29 pm - IP Logged

In a pinch one game is all that is needed.

United States
Member #59354
March 13, 2008
4352 Posts
Offline
 Posted: December 30, 2013, 9:57 pm - IP Logged

My newest program uses the lexie values from the last 5 to 15 games and calculates a sort or average value

which is fed into a algorithm.  The algorithm then adds and/or subtracts specific values to the averaged lexie

then converts the lexies back into sets.  Seems to work great with most games, daily p3 and p4 along with the

P-5 and p-6 number games.  As of yet does not work too well with the big games like MM and PB but see no

reason why the algorithm can't be ajusted for these games, just going to take some time.

RL

....

 Page 1 of 2