Baltimore, MD United States Member #67219 November 20, 2008 105 Posts Offline

Posted: December 5, 2008, 3:21 pm - IP Logged

I was just joking around Rj, Ive paid attention to and respect your dedicated MM work, I also play only MM so developing a system is something Im interested in. Good luck tonight!

Alexandria,VA United States Member #51062 March 26, 2007 367 Posts Offline

Posted: December 5, 2008, 4:53 pm - IP Logged

Greetings all: Sorry for doing two posts today within the system topic for draw 12/5/08…. But I tried to develop a new system which automatically selected twenty numbers and then populated them into rows of five per row. Since everything is done at a “push of a button” – I am hoping to remove the human element out of it. The twenty numbers are chosen based on “out” status meaning how many draws out the number is. They are first divided into four categories Cold Numbers (greater 20 out), Luke Warm Numbers (11-19 Out), Warm Numbers (6-10 Out) and Hot Numbers (0-5 Out). Then the numbers are selected four for Cold, four for Luke Warm, five for Warm and seven for hot. Then the select numbers are sorted and then wheeled. Since I did not purchase any of these lines, my luck is all five will be in this set….Ha ha

Here are the twenty Numbers 1-2-3-5-7 9-12-19-20-25 33-34-36-37-39 40-45-47-52-56

mid-Ohio United States Member #9 March 24, 2001 19826 Posts Offline

Posted: December 5, 2008, 10:34 pm - IP Logged

These are my picks for tonight drawing 12/05/08. All the combinations use 5 numbers from the previous drawings without skipping over one already picked,"clean look back", and the bonus numbers were the most popular ones in the last 360 drawings.

Picked up 4 of 5 WB's in my pool, and 2 2+0's and a 0+1. Most of the hits from the numbers tonight were on tickets I filtered out, but all 20 I removed (I start with 40) didn't have more than 1+0 hits. 10 of the 20 I filtered out had the 1+0 hits.

Not sure if I mentioned before, but I slightly changed my picking algorithms 2 draws ago. Now I'm using a variant on the followers (that I mentioned earlier in this thread). I'm not using what I call a 'Followers + Siblings' method. What I do is first compute the followers, and pick one ball that has the highest tally of followers for all 5 (combined) of the previously drawn WB's. Then I run another calculation that looks at all past draws and finds which balls typically show *with* that ball the most. For instance, when all followers for 21,27,46,52 and 55 are added together, the ball #53 has the highest occurances of showing after any of those balls have hit. I then pick 44 balls (for 45 total) which have been seen the most with ball #53. Their order is as follows for 12/09:

This method with 45 picks matches 3 or better in the pool 78.6% of the time (since the matrix change). This algorithm matches all 5 balls drawn 19.6% of the time. For MB's I choose the top 20 'Due' balls, where the algorithm hits the MB in the selected pool 44.4% of the time with 20 picks. I used to use the top 20 repeat positions for this pick (we use Repeats in Powerball, credit to RJ for that algorithm), but picking the top 20 repeat positions has only matched the MB draw 34.1% of the time since the matrix change in Mega Millions.

Not sure if I mentioned before, but I slightly changed my picking algorithms 2 draws ago. Now I'm using a variant on the followers (that I mentioned earlier in this thread). I'm not using what I call a 'Followers + Siblings' method. What I do is first compute the followers, and pick one ball that has the highest tally of followers for all 5 (combined) of the previously drawn WB's. Then I run another calculation that looks at all past draws and finds which balls typically show *with* that ball the most. For instance, when all followers for 21,27,46,52 and 55 are added together, the ball #53 has the highest occurances of showing after any of those balls have hit. I then pick 44 balls (for 45 total) which have been seen the most with ball #53. Their order is as follows for 12/09:

This method with 45 picks matches 3 or better in the pool 78.6% of the time (since the matrix change). This algorithm matches all 5 balls drawn 19.6% of the time. For MB's I choose the top 20 'Due' balls, where the algorithm hits the MB in the selected pool 44.4% of the time with 20 picks. I used to use the top 20 repeat positions for this pick (we use Repeats in Powerball, credit to RJ for that algorithm), but picking the top 20 repeat positions has only matched the MB draw 34.1% of the time since the matrix change in Mega Millions.

mid-Ohio United States Member #9 March 24, 2001 19826 Posts Offline

Posted: December 7, 2008, 1:35 am - IP Logged

For those that like our posts for the next draw early, I'll post mine now. ;) Just worked them up, hot off the press..

Stew, That was quick. Do your software use the same settings every time? I usually check both my pick and reject files from the last drawing and try to tweak my parameters settings before coming up with my picks for the next drawing.

I calculated some parameters settings for Saturday Ohio Classic Lotto game (6/49) and ran off 50 lines which I posted on the prediction board. They matched 3 and 4 which would have paid $72, but I've never had the desire to play that many lines in a local jackpot game. Once in a way I might play 20 lines and I'm lucky to match 3 for $2, I win more spending the same amount playing MegaMillions. I did calculate the parameters a little differently for the 6/49 game and may try the same thing with the MegaMillion this Tuesday. Now that it getting close to $200M, it time to try everything.

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

Not sure if I mentioned before, but I slightly changed my picking algorithms 2 draws ago. Now I'm using a variant on the followers (that I mentioned earlier in this thread). I'm not using what I call a 'Followers + Siblings' method. What I do is first compute the followers, and pick one ball that has the highest tally of followers for all 5 (combined) of the previously drawn WB's. Then I run another calculation that looks at all past draws and finds which balls typically show *with* that ball the most. For instance, when all followers for 21,27,46,52 and 55 are added together, the ball #53 has the highest occurances of showing after any of those balls have hit. I then pick 44 balls (for 45 total) which have been seen the most with ball #53. Their order is as follows for 12/09:

This method with 45 picks matches 3 or better in the pool 78.6% of the time (since the matrix change). This algorithm matches all 5 balls drawn 19.6% of the time. For MB's I choose the top 20 'Due' balls, where the algorithm hits the MB in the selected pool 44.4% of the time with 20 picks. I used to use the top 20 repeat positions for this pick (we use Repeats in Powerball, credit to RJ for that algorithm), but picking the top 20 repeat positions has only matched the MB draw 34.1% of the time since the matrix change in Mega Millions.

Would it be possible for you to express that as "Percentage versus Random Play" ? I gave the formula a few pages back.

Over a number of draws, calculate the average hit. ( 3 or better in the pool 78.6% of the time and 5 balls drawn 19.6% of the time.) I'm gonna pluck a number out of the air and say the average was 4.1.

(4.1 / 5) / (45 / 55) = 100% which is precisely equal to random play. (Not sure whether you're playing a 55 or 56 WB game).

I've been experimenting with your Siblings of Followers method on my local 5 from 45 powerball. The best I've gotten so far is using the top 4 pairs of the top 5 followers. Discard the followers. Adding in my latest discoveries, which hit 0.625 from 2.625 at 214% and 0.875 from 4.125 at 190%, over 8 consecutive draws I'm hitting an average 2.875 from 16.75 at 154%. Over those 8 draws, best was 4 from 14 (257%), and worst was 2 from 18 (100%).

I truly hope that what I've written makes at least some sense.

CT United States Member #61398 May 21, 2008 781 Posts Offline

Posted: December 7, 2008, 10:43 am - IP Logged

Quote: Originally posted by RJOh on December 7, 2008

For those that like our posts for the next draw early, I'll post mine now. ;) Just worked them up, hot off the press..

Stew, That was quick. Do your software use the same settings every time? I usually check both my pick and reject files from the last drawing and try to tweak my parameters settings before coming up with my picks for the next drawing.

I calculated some parameters settings for Saturday Ohio Classic Lotto game (6/49) and ran off 50 lines which I posted on the prediction board. They matched 3 and 4 which would have paid $72, but I've never had the desire to play that many lines in a local jackpot game. Once in a way I might play 20 lines and I'm lucky to match 3 for $2, I win more spending the same amount playing MegaMillions. I did calculate the parameters a little differently for the 6/49 game and may try the same thing with the MegaMillion this Tuesday. Now that it getting close to $200M, it time to try everything.

RJ, the only settings I would change are filter settings (for after tickets are created) and which algorithm to use to generate picks. So basically each set of numbers would produce 1 possible ticket array. Sometimes I change the pattern, but not too often. The pattern is based only on an even distrubition, not any paramaters set depending on draw history.

CT United States Member #61398 May 21, 2008 781 Posts Offline

Posted: December 7, 2008, 11:27 am - IP Logged

Quote: Originally posted by commanda on December 7, 2008

Stew,

Would it be possible for you to express that as "Percentage versus Random Play" ? I gave the formula a few pages back.

Over a number of draws, calculate the average hit. ( 3 or better in the pool 78.6% of the time and 5 balls drawn 19.6% of the time.) I'm gonna pluck a number out of the air and say the average was 4.1.

(4.1 / 5) / (45 / 55) = 100% which is precisely equal to random play. (Not sure whether you're playing a 55 or 56 WB game).

I've been experimenting with your Siblings of Followers method on my local 5 from 45 powerball. The best I've gotten so far is using the top 4 pairs of the top 5 followers. Discard the followers. Adding in my latest discoveries, which hit 0.625 from 2.625 at 214% and 0.875 from 4.125 at 190%, over 8 consecutive draws I'm hitting an average 2.875 from 16.75 at 154%. Over those 8 draws, best was 4 from 14 (257%), and worst was 2 from 18 (100%).

I truly hope that what I've written makes at least some sense.

Amanda

Calculating 'Percentage versus Random Play' does play out to almost exactly 100% over the entire life of MM history, but shows much better in smaller doses. Calcualting for the past 6 draws shows an averge of 4.5, giving us (4.5/5)/(45/56) = 112%. Using smaller amounts of draws like this shows anywhere from 68% (April 8 -May 2) to 120% (August 12 to August 29). The highest possible percentage for picking 45 numbers with this equation is (5/5)/(45/56) = 124%.