6/49 Discussion For May 05, 2004

[gsobier]

Here is what my program reports for the next draw along with a colour
code of hits/bonus number we had...
(the bonus number could show up anywhere)

(either four, three, or five hits most of the time from here)
01 02 04 05 07 08
09 10 11 12 14 16
17 18 21 22 25 26
28 30 31 35 37 38
39 44 45 46 48 49

No more than three from here is most likely:
(2 hits came from here in last draw)
03 06 13 15 19 20
23 24 27 29 32 33
34 36 40 41 42 43
47

For the last 1000 draws, we had the following results:
[4 and 2]: 340
[3 and 3]: 265
[5 and 1]: 234
[2 and 4]: 087
[6 and 0]: 049
[1 and 5]: 022
[0 and 6]: 003

Good luck to all players!

Regards,
George

 

[Sheba]

Jackpot for May 5 is :
2 Million..

Sheba__ ~~~
Everybody's Got a Dream...JACKPOT!
--------------------------------------------------------------------------

 

[Karnac]

May 5 ... 649

The O/E bias for the last 10 draws is 30/40 for a +10 Even bias.....in continuing this trend 10 28 40 42 44 46 48 have not appeared in the last 5 draws making them attractive if the Even number trend continues. Combine these with 8 22 24 36 which have appeared most often in the last 5 games .These numbers along with other even numbers of your choosing would be a possible play in an even bias wheel. Add 21 25 39 45 47 which are the most recent odd numbers and you can cover an even/odd parameter .

The O/E bias for the last 10 draws is 30/40 for a +10 Even bias.....in opposing this trend 1 3 5 7 9 17 19 31 33 have not appeared in the last 5 draws making them attractive if an Odd number trend begins. Combine these with 21 25 39 45 47 which have appeared most often in the last 5 games . These numbers along with other Odd numbers of your choosing would be a possible play in an odd bias wheel. Add 24 8 22 36 which are the most recent even numbers and you can cover an odd/even parameter .

Good Luck all.......

 

[hot4]

Will we consider last draw (8,21,24,26,34,38/14) important enough to change our last assumptions?
I'm feeling that such assumptions must not be discarded at all.

I go on liking 18 with 22 and/or 39 for example.

 

[Sheba]

Should see 3 or more from these:
03-08-12-14-17-19-25-27-32-34-40-43-45
With the 08-14-34 as possible repeaters.


Sheba__ ~~~
Everybody's Got a Dream...JACKPOT!
--------------------------------------------------------------------------

 

[Sheba]

<snip>
(either four, three, or five hits most of the time from here)
01 02 04 05 07 08
09 10 11 12 14 16
17 18 21 22 25 26
28 30 31 35 37 38
39 44 45 46 48 49

No more than three from here is most likely:
(2 hits came from here in last draw)
03 06 13 15 19 20
23 24 27 29 32 33
34 36 40 41 42 43
47
<snip>
Good luck to all players!

Regards,
George

George..
From your first group I really Like these:
18-22-30-39-46 and from your second group
I like these: 03-19-27-32-34-40-43....
and 34 is Looking really good as repeater...


Sheba__ ~~~
Everybody's Got a Dream...JACKPOT!
--------------------------------------------------------------------------

 

[Beaker]

Some gimmies lining up

Something from 3-4 ,
seeing a lot of 13-15-40 and
a few from 43-45-47

 

[Beaker]

Should see 3 or more from these:
03-08-12-14-17-19-25-27-32-34-40-43-45
With the 08-14-34 as possible repeaters.


Sheba__ ~~~
Everybody's Got a Dream...JACKPOT!
--------------------------------------------------------------------------

With some minor modifications to this set, you'll win - BIG

 

[Maggie]

What I'm getting is mostly numbers under 10 and in the 40's. Not too much of 20's and 30's.

 

[Dennis Bassboss]

Just starting here....I'm looking at 11-16 .....

 

[Sheba]

Good Luck People in Wednesday's Draw..

Sheba__~~~
Everybody's Got a Dream...JACKPOT!
---------------------------------------------------------------

 

[Nick Koutras]

Question for Ideas!

Let say we have a collection of Numbers (a Group).

Q. What information you/we should know
(from the database) to arrive to some concrete conclusions as of what may happen on the next 1-3 draws about that Group?

 

[Dennis Bassboss]

Question for Ideas!

Let say we have a collection of Numbers (a Group).

Q. What information you/we should know
(from the database) to arrive to some concrete conclusions as of what may happen on the next 1-3 draws about that Group?

In my opinion...And it is only an opinion here...
For prediction purposes in a relative short time like you say 1-3 draws here's my steps...
1-Look at which ones are currently hot from that group...
2-Look at recents hot pairs and triples and try to find some good matches with what is currently hot.. from outside the group itself..(so you have to also watch for cold pairs doing that, a very hot number often comes in with another number that is medium or very cold though the pair is in itself very cold...)
3- Third thing to know would be to see what numbers are coming in pairs and triples mostly in the recents draws concerning that initial group...In other words which come together mostly from within the group itself...You can spot some low or high trends in that at times..
4- After pairing the best couples or triples look at what they've announced in the past at different times in the history...
What you search for then is to find what was mostly announced and is no longer mostly announced or has not been mostly announced for a long time..
What has been hot will be hot again at one point from an announcing stand point...
5- You match that with what is announced mostly in the recent history by these pairs and triples..
From there on you can uses the usuals look out...
So you also look for hot repeaters and consecutives...You try to have a balance of odd-even according to the actual trend for this...You also have a nice spread between your numbers...Therefore the decades and the LDs also etc...

 

[Beaker]

Question for Ideas!

Let say we have a collection of Numbers (a Group).

Q. What information you/we should know
(from the database) to arrive to some concrete conclusions as of what may happen on the next 1-3 draws about that Group?

Well, you have to classify the last draw in some way ( or many ways) - you need to develop a signature.

This is the key - what are the components of the signature.

I'm looking at cluster analysis - as powerball suggested - and Nick you posted a good article about that.

So, the idea is what does the last draw cluster with. This gives us an idea of what happened around draws that cluster together.

I'm only looking at 200 draws - a software limit - but, for example,
the last draw clusters with the following - best to least ( nearest neighbour) with bonus.

1926 7 10 21 23 28 31 39
2063 8 9 22 29 33 37 49
2020 14 18 22 23 31 32 45
2027 10 14 15 27 36 41 47
1981 14 15 17 25 36 38 45

Now, clustering can tell us what is similar but is not good as a predictive tool. I think we can evolve that based on what we know.

So, Nick one good question about the data is where in the history do we have similar signature.

 

[Nick Koutras]

In my opinion...And it is only an opinion here...
For prediction purposes in a relative short time like you say 1-3 draws here's my steps...
1-Look at which ones are currently hot from that group...
2-Look at recents hot pairs and triples and try to find some good matches with what is currently hot.. from outside the group itself..(so you have to also watch for cold pairs doing that, a very hot number often comes in with another number that is medium or very cold though the pair is in itself very cold...)
3- Third thing to know would be to see what numbers are coming in pairs and triples mostly in the recents draws concerning that initial group...In other words which come together mostly from within the group itself...You can spot some low or high trends in that at times..
4- After pairing the best couples or triples look at what they've announced in the past at different times in the history...
What you search for then is to find what was mostly announced and is no longer mostly announced or has not been mostly announced for a long time..
What has been hot will be hot again at one point from an announcing stand point...
5- You match that with what is announced mostly in the recent history by these pairs and triples..
From there on you can uses the usuals look out...
So you also look for hot repeaters and consecutives...You try to have a balance of odd-even according to the actual trend for this...You also have a nice spread between your numbers...Therefore the decades and the LDs also etc...

Wow Dennis,
You have described a complete selection procedure.

At the moment my ideas as twosome:

Firstly, display some historical statistical results about the Group

Define here the information that the user wishes to know.
..........



Secondly, define some theoretical procedure to identify/ select/ eliminate or anything
that can help us make some decision that can help us include/ exclude some other numbers.
.........



If you or anybody else can fill the two sections with some ideas
I'm willing to put them in a procedure and apply them to a spreadsheet.


I'll post some outputs, soon as I finish the programming part.

 

[Nick Koutras]

Well, you have to classify the last draw in some way ( or many ways) - you need to develop a signature.

This is the key - what are the components of the signature.

I'm looking at cluster analysis - as powerball suggested - and Nick you posted a good article about that.

So, the idea is what does the last draw cluster with. This gives us an idea of what happened around draws that cluster together.

I'm only looking at 200 draws - a software limit - but, for example,
the last draw clusters with the following - best to least ( nearest neighbour) with bonus.

1926 7 10 21 23 28 31 39
2063 8 9 22 29 33 37 49
2020 14 18 22 23 31 32 45
2027 10 14 15 27 36 41 47
1981 14 15 17 25 36 38 45

Now, clustering can tell us what is similar but is not good as a predictive tool. I think we can evolve that based on what we know.

So, Nick one good question about the data is where in the history do we have similar signature.

I use a different form of clusters as follows.

Store the Hits of the last ~4-5 draws
until the size is LottoBalls/2 as a binary string, Hit =1 not a hit=0

Then I compare these strings for maximum similarity and that is indicated as Maximum Efficiency.

I have used this technique to generate the
Neurons for the Announces Method.

 

[Beaker]

I use a different form of clusters as follows.

Store the Hits of the last ~4-5 draws
until the size is LottoBalls/2 as a binary string, Hit =1 not a hit=0

Then I compare these strings for maximum similarity and that is indicated as Maximum Efficiency.

I have used this technique to generate the
Neurons for the Announces Method.

Nick, based on what I am seeing, your method is fine

One input that seems to be very good is the sum of the numbers. Take a look at that for yourself - you may want to incorporate that value.

Take a look at the sums for the following cluster:

2116	8	14	21	24	26	34	38	165
1988	5	15	23	24	28	30	41	166
1970	1	14	21	26	30	33	38	163
2003	1	7	25	32	33	34	35	167
1948	3	11	19	29	34	35	36	167

 

[Nick Koutras]

Nick, based on what I am seeing, your method is fine

One input that seems to be very good is the sum of the numbers. Take a look at that for yourself - you may want to incorporate that value.

Take a look at the sums for the following cluster:

2116	8	14	21	24	26	34	38	165
1988	5	15	23	24	28	30	41	166
1970	1	14	21	26	30	33	38	163
2003	1	7	25	32	33	34	35	167
1948	3	11	19	29	34	35	36	167

Ok,
I'll think of a way to incorporate the Sum & SD of a draw to Announces

 

[shirazbai]

I am absolutely clueless at this point. This is going to keep me busy for weeks.

 

[luckystrike]

i THINK # 14 MAY REPEAT THIS DRAW...WHICH WOULD LEAVE ME WITH...# 13..&..14.. AS MY DN'S...BUT I ALSO LIKE THE LOOKS OF 15...

ANY THOUGHTS OUT THERE?????