Welcome Guest
You last visited January 24, 2017, 8:11 am
All times shown are
Eastern Time (GMT-5:00)

# Global Lottery Solutions (G.L.S)

Topic closed. 195 replies. Last post 7 years ago by StrikeSentinel.

 Page 9 of 14
Park City, UT
United States
Member #69864
January 18, 2009
995 Posts
Offline
 Posted: February 3, 2010, 4:38 pm - IP Logged

The code is not that smart right now.  In your example above the draw 46-15-17-28-16-32-22 is a draw for Draw A results.  14-28-11-04-15-42-46 is in Draw B results.  You then take the unmatched numbers from both draws.

A.  46-15-32-22  B. 28-11-15-46.   If they have only 1 common number between them then remove the common number and combine them into a prediction.  If they have 0 common or more that 1 common number then I don't create a prediction from them because you would not have 6 unique numbers left.  In this example the number 46 links them together so you remove 46 from the two 4 number groups and that leaves you with two 3 number groups and you combine them together for a prediction.

You could search for match4 in both A and B that would leave you with 3 and 3 and if they did not have any common numbers then generate a prediction with the 6 remaining balls.

If you want to search for more than Match3 then it becomes more tricky in how you combine them together to form a prediction.  I'm willing to add it but someone has to give the exact algorithm to do it for all the ways you want them combined together.  There is no gray area in programming everything needs to be specified.

Other than myself and Moses, 4 other people have the software Turtle, Carbob, ChetserDennis, and martor54.  I don't mine giving out the software to anyone, but I thought Moses didn't want me to give it to anyone that just joined and did not provide feedback or help.  In your case you did not just join and it appears you want to help so I don't have a problem providing it to you.  It would be easy to just post a link to my website like I was doing previously and let anyone have it, but I did do this software for Moses so I want to respect his wishes with who gets access to the software.

Jimmy

UK
United Kingdom
Member #9650
December 15, 2004
222 Posts
Offline
 Posted: February 4, 2010, 4:06 pm - IP Logged

The code is not that smart right now.  In your example above the draw 46-15-17-28-16-32-22 is a draw for Draw A results.  14-28-11-04-15-42-46 is in Draw B results.  You then take the unmatched numbers from both draws.

A.  46-15-32-22  B. 28-11-15-46.   If they have only 1 common number between them then remove the common number and combine them into a prediction.  If they have 0 common or more that 1 common number then I don't create a prediction from them because you would not have 6 unique numbers left.  In this example the number 46 links them together so you remove 46 from the two 4 number groups and that leaves you with two 3 number groups and you combine them together for a prediction.

You could search for match4 in both A and B that would leave you with 3 and 3 and if they did not have any common numbers then generate a prediction with the 6 remaining balls.

If you want to search for more than Match3 then it becomes more tricky in how you combine them together to form a prediction.  I'm willing to add it but someone has to give the exact algorithm to do it for all the ways you want them combined together.  There is no gray area in programming everything needs to be specified.

Other than myself and Moses, 4 other people have the software Turtle, Carbob, ChetserDennis, and martor54.  I don't mine giving out the software to anyone, but I thought Moses didn't want me to give it to anyone that just joined and did not provide feedback or help.  In your case you did not just join and it appears you want to help so I don't have a problem providing it to you.  It would be easy to just post a link to my website like I was doing previously and let anyone have it, but I did do this software for Moses so I want to respect his wishes with who gets access to the software.

Jimmy

Hello all

Jimmy is right, I asked him to not allow some members to download the program if they are newly re-registered! My reason is, if too many members download the program then there is going to be more questions rather than solutions! I want to spend my time to find more information as I have got the tool that I need rather than answering questions but it is guaranteed that you will get the final result! But however I think Jimmy’s idea is even better as he is asking for specifications if you want the program upgraded so this means you have to provide feedback.

My concentration is on this issue why the program can only produce match 5 and match 5+ BB but no match 6 or even 7s, I think I already found the answer which could be the error from my end. To find the answer I need to do lots of experiments with program!

Moses

Ontario
Member #70051
January 23, 2009
76 Posts
Offline
 Posted: February 4, 2010, 7:10 pm - IP Logged

Hello all

Jimmy is right, I asked him to not allow some members to download the program if they are newly re-registered! My reason is, if too many members download the program then there is going to be more questions rather than solutions! I want to spend my time to find more information as I have got the tool that I need rather than answering questions but it is guaranteed that you will get the final result! But however I think Jimmy’s idea is even better as he is asking for specifications if you want the program upgraded so this means you have to provide feedback.

My concentration is on this issue why the program can only produce match 5 and match 5+ BB but no match 6 or even 7s, I think I already found the answer which could be the error from my end. To find the answer I need to do lots of experiments with program!

Moses

--------------------------

Maybe a couple of things:

Standardized File Format: Mo might be using the other data for other things, but if Jim's program is only using the draw numbers for prediction, then it would be easier for user's to input their data, and not create extra, non-important stuff.

Link Numbers: Just a curious note as to why the link numbers are not included in prediction. There are many times that ball numbers repeat themselves in the next draw. To eliminate them from prediction could prevent you from getting a higher prize. However it might increase the number of predictions.

Data Analysis: Results may vary on this point. As stated earlier in a previous post, Mo is using 6-7 lotteries to derive is findings. Other may need more or less.  As you can see from Canada's POV, there might be more than 5000 draws to sift through, and to gather all that data might be troublesome. So, one might end up with more than the standard 100 - 500 matches to begin with, This in turn will increase the number of predictions - perhaps more than what Jim's program allows at the moment.

There's still more investigating to be done.  To be continued.....

CD

Park City, UT
United States
Member #69864
January 18, 2009
995 Posts
Offline
 Posted: February 4, 2010, 8:44 pm - IP Logged

Hello all

Jimmy is right, I asked him to not allow some members to download the program if they are newly re-registered! My reason is, if too many members download the program then there is going to be more questions rather than solutions! I want to spend my time to find more information as I have got the tool that I need rather than answering questions but it is guaranteed that you will get the final result! But however I think Jimmy’s idea is even better as he is asking for specifications if you want the program upgraded so this means you have to provide feedback.

My concentration is on this issue why the program can only produce match 5 and match 5+ BB but no match 6 or even 7s, I think I already found the answer which could be the error from my end. To find the answer I need to do lots of experiments with program!

Moses

Hello,

Moses said "why the program can only produce match 5 and match 5+ BB but no match 6 or even 7s,".  Even though the lotto is choosing 7 balls, we the users are only choosing 6 balls in our predictions so the best we can hope to match against the actual results draw is 6.  You would think match6 would happen as often as match5BB, because we are comparing our 6 numbers against 7 numbers but so far I have seen no Match6.

Jimmy

Park City, UT
United States
Member #69864
January 18, 2009
995 Posts
Offline
 Posted: February 4, 2010, 9:07 pm - IP Logged

--------------------------

Maybe a couple of things:

Standardized File Format: Mo might be using the other data for other things, but if Jim's program is only using the draw numbers for prediction, then it would be easier for user's to input their data, and not create extra, non-important stuff.

Link Numbers: Just a curious note as to why the link numbers are not included in prediction. There are many times that ball numbers repeat themselves in the next draw. To eliminate them from prediction could prevent you from getting a higher prize. However it might increase the number of predictions.

Data Analysis: Results may vary on this point. As stated earlier in a previous post, Mo is using 6-7 lotteries to derive is findings. Other may need more or less.  As you can see from Canada's POV, there might be more than 5000 draws to sift through, and to gather all that data might be troublesome. So, one might end up with more than the standard 100 - 500 matches to begin with, This in turn will increase the number of predictions - perhaps more than what Jim's program allows at the moment.

There's still more investigating to be done.  To be continued.....

CD

Hello,

Chetser said the following:

"Link Numbers: Just a curious note as to why the link numbers are not included in prediction. There are many times that ball numbers repeat themselves in the next draw. To eliminate them from prediction could prevent you from getting a higher prize. However it might increase the number of predictions."

So remember we enter 2 draws, Draw A and Draw B.  Lets use an easy example and say Draw A is 01-02-03-04-05-06-07 and Draw B is 07-49-48-47-46-45-44

We search thru our database for match3 against Draw A and we are going to use the unmatched numbers to form half of a prediction.  What this means in this example is that the 3 numbers from Draw A must come from pool of numbers 08-49.

We then search thru our database for match3 againg Draw B and we are going to use the unmatched number to form half of a prediction.  What this means in this example is that the number from Draw B must come from pool of number 01-06, and 08-43.  So in this case the number 07 will never be in our predictions.  So any common numbers in Draw A and Draw B will be excluded in our prediction for Future Draw C.  This could be a serious flaw that we need to fix.

The other thing is to not to get to hung up on the number of predictions.  Lets say we progress further with this and sometime in the future can guarantee a Match5BB in 100,000 predictions for every draw.  This is a huge accomplishment because the odds to match 5+BB in UK lotto is 13,983,816/6 which is 1 in 2,330,636.  So 1 in 100,000 is huge.  Of course what is more important is if we can't guarantee it is to at least be profitable.  So I would like to see our Match3 hits go up because I think that is the hidden secret to being profitable and the rest is gravy.

Jimmy

New Member
Johannesburg
South Africa
Member #64022
August 16, 2008
6 Posts
Offline
 Posted: February 5, 2010, 3:28 am - IP Logged

Hi Moses and Jimmy

I understand your reasons for not distributing the software especially with the number of questions that were raised in all forums about your methodology.

I used and participated in discussions on Patrick's program (which included a Moses section) and others for a long time before the development stopped.

I look forward to working with your software and backtesting it against our local and the UK drawings.

UK
United Kingdom
Member #9650
December 15, 2004
222 Posts
Offline
 Posted: February 5, 2010, 8:33 am - IP Logged

--------------------------

Maybe a couple of things:

Standardized File Format: Mo might be using the other data for other things, but if Jim's program is only using the draw numbers for prediction, then it would be easier for user's to input their data, and not create extra, non-important stuff.

Link Numbers: Just a curious note as to why the link numbers are not included in prediction. There are many times that ball numbers repeat themselves in the next draw. To eliminate them from prediction could prevent you from getting a higher prize. However it might increase the number of predictions.

Data Analysis: Results may vary on this point. As stated earlier in a previous post, Mo is using 6-7 lotteries to derive is findings. Other may need more or less.  As you can see from Canada's POV, there might be more than 5000 draws to sift through, and to gather all that data might be troublesome. So, one might end up with more than the standard 100 - 500 matches to begin with, This in turn will increase the number of predictions - perhaps more than what Jim's program allows at the moment.

There's still more investigating to be done.  To be continued.....

CD

Hello Chester

Yes, you’re right again and this is the area I am concentrating now! After combining the two draws and eliminating the numbers which already took match and producing results with none-matched numbers means in our predictions we are not going to have a repeated number(s) from draw and two draws ago!

..So, this is one of the reasons we don’t produce more of match 5+BB (that is what I meant when I said match 7s which is 5+ BB but happens more than once) is because of repeated numbers which recently we are experiencing in lotto results such as following;

27/01/2010, 27,13,01,04,07,11,12 --- lotto Wed

30/01/2010, 02,04,19,40,23,03,22 --- lotto Sat

03/02/2010, 11,03,12,28,21,20,31 --- lotto Wed

As you can see there are three numbers sharing in the newest draw from the past two results, this is why we don’t tend to produce match 5+ BB more or even match 6s!

I have an idea however which I will consult it with Jimmy first to get his opinion before I disclose it here!

BTW, this was going almost unnoticed, last Sat lotto result is a 6 numbers repeat which occurred in daily draw

30/01/2010, 02,04,19,40,23,03,22 --- lotto Sat

05/07/2007, 04 22 03 02 13 19 23 --- daily draw

What is the odd for this match 6 occurring in 6000 results? Not to forget that the latest lotto result has only one number in 30s and the rest are below again!

For past 7 UK results there only two numbers in 30 (30/31) and four numbers in 40s (40/42/43/43)

Do you think there is a pressure on Camelot?

Moses

P.S. Jimmy if you don’t mind to give the software to the members then you can but by request on the forum please so we know who is receiving it!

Thanks

Park City, UT
United States
Member #69864
January 18, 2009
995 Posts
Offline
 Posted: February 5, 2010, 12:01 pm - IP Logged

Hello Chester

Yes, you’re right again and this is the area I am concentrating now! After combining the two draws and eliminating the numbers which already took match and producing results with none-matched numbers means in our predictions we are not going to have a repeated number(s) from draw and two draws ago!

..So, this is one of the reasons we don’t produce more of match 5+BB (that is what I meant when I said match 7s which is 5+ BB but happens more than once) is because of repeated numbers which recently we are experiencing in lotto results such as following;

27/01/2010, 27,13,01,04,07,11,12 --- lotto Wed

30/01/2010, 02,04,19,40,23,03,22 --- lotto Sat

03/02/2010, 11,03,12,28,21,20,31 --- lotto Wed

As you can see there are three numbers sharing in the newest draw from the past two results, this is why we don’t tend to produce match 5+ BB more or even match 6s!

I have an idea however which I will consult it with Jimmy first to get his opinion before I disclose it here!

BTW, this was going almost unnoticed, last Sat lotto result is a 6 numbers repeat which occurred in daily draw

30/01/2010, 02,04,19,40,23,03,22 --- lotto Sat

05/07/2007, 04 22 03 02 13 19 23 --- daily draw

What is the odd for this match 6 occurring in 6000 results? Not to forget that the latest lotto result has only one number in 30s and the rest are below again!

For past 7 UK results there only two numbers in 30 (30/31) and four numbers in 40s (40/42/43/43)

Do you think there is a pressure on Camelot?

Moses

P.S. Jimmy if you don’t mind to give the software to the members then you can but by request on the forum please so we know who is receiving it!

Thanks

Moses said

Yes, you’re right again and this is the area I am concentrating now! After combining the two draws and eliminating the numbers which already took match and producing results with none-matched numbers means in our predictions we are not going to have a repeated number(s) from draw and two draws ago!

Well this statement is not true.  A prediction is made as follows A1-A2-A3-B1-B2-B3.  Assume you use the following as DrawA  01-02-03-04-05-06-07  then A1-A2-A3 will never include the numbers 1-7, but can only be 8-49.  Assume you use the following for Draw B 43-44-45-46-47-48-49 then B1-B2-B3 will never include the numbers 43-49, but can only be 1-42.  So in this case all 49 numbers are still in play but the numbers 8-42 are more powerful because they can be in A1-A2-A3-B1-B2-B3 (think intersection of sets using Math), whereas 1-7 can only be in B1-B1-B3 and 43-49 can only be in A1-A2-A3.  But if you have common numbers in both DrawA and DrawB then currently they are excluded, which I noted a couple of posts ago.  Please think about this if you don't understand this then keep asking questions.  I think it is imperative people understand how predictions are generated.

Jimmy

Park City, UT
United States
Member #69864
January 18, 2009
995 Posts
Offline
 Posted: February 5, 2010, 12:04 pm - IP Logged

Hi Moses and Jimmy

I understand your reasons for not distributing the software especially with the number of questions that were raised in all forums about your methodology.

I used and participated in discussions on Patrick's program (which included a Moses section) and others for a long time before the development stopped.

I look forward to working with your software and backtesting it against our local and the UK drawings.

Jimmy

United Kingdom
Member #85601
January 19, 2010
11 Posts
Offline
 Posted: February 5, 2010, 12:16 pm - IP Logged

Hi Moses / Jimmy / Chester,

As you yourselves have pointed out, the problem of repeated numbers from the previous draw is very important. Here is the situation I have found with the UK Main Lotto to date. Some may find it useful in making their own strategies.

1 main # repeated from the previous draw                  :      603     40.94%

2 main #'s repeated from the previous draw                :      215    14.60%

3 main #'s repeated from the previous draw                :        22      1.49%  ... ... ...  TOTAL (1+2+3) = 57.03%

4 main #'s repeated from the previous draw                :          1      0.07%

5 main #'s repeated from the previous draw                :          0      0.00%

6 main #'6 repeated from the previous draw                :          0      0.00%

Bonus # from the previous draw repeated as main #   :      179    12.15%

Main # from the previous draw repeated as Bonus #   :      174    11.81%

Bonus # from the previous draw repeated as Bonus # :       36      2.44%  ... ... ...  GRAND TOTAL = 83.43%

I think I did not make any errors. (At least, I hope.)

It is easy to take into acount the link number and make it part of of a 6-number combination. Obviously, the only problem is that we will have 7x the amount of combinations as C(7,6) = 7. Also, the amount of time it would take to check numbers on the playslips, put them through the terminals, let alone the money involved, makes it virtually impossible to put into practice. We should, perhaps, use other extra methods in conjunction with Moses's elegant system, such as skips, median skips and standard deviation.

Best regards,

martor54

P.S. I have made the above breakdown with a bit of program made in QBasic that works as a stand alone. It works on any 6+B out of 49 database. Whoever wants it, give me a "shout" and I'll send it to you.

UK
United Kingdom
Member #9650
December 15, 2004
222 Posts
Offline
 Posted: February 5, 2010, 12:25 pm - IP Logged

Moses said

Yes, you’re right again and this is the area I am concentrating now! After combining the two draws and eliminating the numbers which already took match and producing results with none-matched numbers means in our predictions we are not going to have a repeated number(s) from draw and two draws ago!

Well this statement is not true.  A prediction is made as follows A1-A2-A3-B1-B2-B3.  Assume you use the following as DrawA  01-02-03-04-05-06-07  then A1-A2-A3 will never include the numbers 1-7, but can only be 8-49.  Assume you use the following for Draw B 43-44-45-46-47-48-49 then B1-B2-B3 will never include the numbers 43-49, but can only be 1-42.  So in this case all 49 numbers are still in play but the numbers 8-42 are more powerful because they can be in A1-A2-A3-B1-B2-B3 (think intersection of sets using Math), whereas 1-7 can only be in B1-B1-B3 and 43-49 can only be in A1-A2-A3.  But if you have common numbers in both DrawA and DrawB then currently they are excluded, which I noted a couple of posts ago.  Please think about this if you don't understand this then keep asking questions.  I think it is imperative people understand how predictions are generated.

Jimmy

I think, I'll do one test and put the  result from both files A and B here so we can all walk together! The way I understand it (may be I am wrong) when you mix both files (A and B) to generate predictions then none of the numbers from either file is used as link, if

01 02 03 04 05 06 07 = File A then none of these numbers are in link = ?

07 49 48 47 46 45 44 = File B then none of these numbers are in link = ?  However that 07 is shared in both, I'll do more tests and will be back!

Ontario
Member #70051
January 23, 2009
76 Posts
Offline
 Posted: February 5, 2010, 12:29 pm - IP Logged

-------------

Hello Guys,

You may have some matches regardless. If the number doesn't show up in Draw A's match list (because the number was marked for elimination), it might appear in Draw B's match list - unless the number was higlighted on both list. A minor problem perhaps.

My question was initally pointing to the unmatched numbers. Unfortunately we lose the link number when matching/creating the prediction lines.

Right now, I'm just rambling. Mo and Jim will sort this out eventually.

CD

Park City, UT
United States
Member #69864
January 18, 2009
995 Posts
Offline
 Posted: February 5, 2010, 2:53 pm - IP Logged

-------------

Hello Guys,

You may have some matches regardless. If the number doesn't show up in Draw A's match list (because the number was marked for elimination), it might appear in Draw B's match list - unless the number was higlighted on both list. A minor problem perhaps.

My question was initally pointing to the unmatched numbers. Unfortunately we lose the link number when matching/creating the prediction lines.

Right now, I'm just rambling. Mo and Jim will sort this out eventually.

CD

No your not rambling this is important to discuss.  There is no right or wrong way in making predictions.  The way I initially implemented the prediction algorithm was how I intrepreted what Moses wanted.  If I did it wrong then it can be fixed.  Maybe we want to have the prediction algorithm configurable where we can do it a number of different ways.  What your saying about the link ball is correct we are throwing it out and not using it in that prediction.  Instead of throwing it out we can wheel the 7 numbers into 6 number combinations, so we would have 7 times the predictions.  Thats okay if we have too many predictions, we want to capture the result first then work on reducing the predictions.  Right now we have seen some success using this tool, but from my experiments it is not 100% so we do need to make some adjustments.  I just do not know what those adjustments are without further tinkering.

So these are the enhancements that I'm going to make hopefully this weekend.  I will be out of town all of next week on business, so what I get done this weekend will have to do for the short term.

1) Allow one to specify the format of their draw file so that it can read in properly any pick5, pick6, or pick6+BB lottery game.
2) Allow one to specify how to combine numbers into a prediction to make it more flexible.  This one is definitely more complex in trying to make it work for pick5, pick6, and pick6+BB because they are different.  So if people have suggestions on how they want to generate predictions then make your requirements known.  Again at this stage there is no right or wrong way in generating a prediction, all feedback is welcome, and I will implement any prediction system you ask for.

Anyway I think Chetser you and I are on the same page, everything you said is true in how prediction system is currently implemented.  It sounds like we need to have more flexibility in how predictions are generated so people can experiment to see which method works best for their own lottery.

Jimmy

Park City, UT
United States
Member #69864
January 18, 2009
995 Posts
Offline
 Posted: February 5, 2010, 3:16 pm - IP Logged

I think, I'll do one test and put the  result from both files A and B here so we can all walk together! The way I understand it (may be I am wrong) when you mix both files (A and B) to generate predictions then none of the numbers from either file is used as link, if

01 02 03 04 05 06 07 = File A then none of these numbers are in link = ?

07 49 48 47 46 45 44 = File B then none of these numbers are in link = ?  However that 07 is shared in both, I'll do more tests and will be back!

I'm looking strictly for match3 using DrawA and DrawB, not match3 or greater so the way the prediction algorithm is currently implemented is why you never see any of the balls in DrawA or DrawB as a link ball.

Just so everybody is clear my definition of link balls is as follows:

There are 4 unmatched numbers left over after DrawA filter is done.  There are 4 unmatched numbers left over after DrawB filter is done.  So you have 8 balls left to make a prediction.  If six balls are unique and the other two are the same, then we make a prediction using the 6 unique numbers and the number common between the two 4-ball groups is the link ball.  In other words the link ball linked the two 4-number balls groups together.  I have tried to explain this a number of times now, is everyone catching on?  We are throwing out the link ball, I initially asked if we wanted to throw out the link ball, if we don't then I can fix it,  just tell me how to do it and still make a 6-ball prediction.  As Martor suggested earlier we can wheel the 7 numbers to make 6 predictions instead of one prediction so that we are not throwing away the link ball.

Sometimes my posts might not come across right, but I want to thank all the contributions that Chetser, Martor and others have made.    Its only by hashing out the details will we ever get a good working solution.

Jimmy

Ontario
Member #70051
January 23, 2009
76 Posts
Offline
 Posted: February 5, 2010, 10:53 pm - IP Logged

I'm looking strictly for match3 using DrawA and DrawB, not match3 or greater so the way the prediction algorithm is currently implemented is why you never see any of the balls in DrawA or DrawB as a link ball.

Just so everybody is clear my definition of link balls is as follows:

There are 4 unmatched numbers left over after DrawA filter is done.  There are 4 unmatched numbers left over after DrawB filter is done.  So you have 8 balls left to make a prediction.  If six balls are unique and the other two are the same, then we make a prediction using the 6 unique numbers and the number common between the two 4-ball groups is the link ball.  In other words the link ball linked the two 4-number balls groups together.  I have tried to explain this a number of times now, is everyone catching on?  We are throwing out the link ball, I initially asked if we wanted to throw out the link ball, if we don't then I can fix it,  just tell me how to do it and still make a 6-ball prediction.  As Martor suggested earlier we can wheel the 7 numbers to make 6 predictions instead of one prediction so that we are not throwing away the link ball.

Sometimes my posts might not come across right, but I want to thank all the contributions that Chetser, Martor and others have made.    Its only by hashing out the details will we ever get a good working solution.

Jimmy

--------------------------

@martor54
Thanks for the info. This gives us some idea in general on how important those previous draw numbers are.

@jimjwright
Your explanation is fine. We're just pointing out different aspects of the same problem. Wheeling 7 numbers, as martor54 suggests, is fine, but I think your ROI prize scenarios would be compromised due to the fact that more predictions are created, unless the top tier prizes can be captured on a regular basis.

I do like the initial procedure we currently have since the prediction footprint is relatively small. And anyone would gladly accept a 5+BB in their list.

More testing...
CD

 Page 9 of 14