# Reducing

Since finding a pattern that worked 64% (10 sets, 10 days) of the time, I really started hitting that hard. I kept doing the workout, one by one, to find a better pattern. I stumbled on two. The first is a twenty number system that hits 91% (tested 5 years) of the time. The other is a ten number system, if played for ten days, would hit 72% (tested over five years) of the time. So finding the percentage could get that high for a sure fire, no bias needed method. I wanted to see what was the best. I wrote a script that would test all three positions. Sadly, my system won't run it (quad core, 12 GB RAM...). I got an out of memory error after 11 hours... I scaled it back a bit to just do two positions and got this:

0: 089, 039, 058, 019, 049, 088, 059, 048, 099, 079
1: 109, 129, 102, 182, 169, 176, 112, 178, 137, 147
2: 268, 208, 210, 218, 230, 260, 258, 200, 263, 288
3: 396, 376, 316, 346, 397, 395, 385, 315, 398, 391
4: 460, 450, 436, 430, 466, 456, 457, 417, 492, 432
5: 597, 567, 569, 537, 547, 506, 583, 513, 543, 541
6: 684, 634, 614, 638, 654, 624, 618, 613, 644, 698
7: 734, 784, 714, 738, 731, 764, 754, 716, 702, 774
8: 861, 801, 851, 836, 856, 866, 847, 897, 826, 846
9: 935, 925, 955, 933, 923, 965, 928, 948, 912, 922

The way it works is, you wait for a double to hit and then look for that number's sets. So if a double 2 hit, you could expect 268, 208, 210, 218, 230, 260, 258, 200, 263, 288 to hit. It might take 1 draw or 20. Keeping in mind that these are box sets. My goal is to find the best reduction vs. highest hit percentage. Then work on putting them in order. I don't want to find a good order process only to find a better reduction and then have to start all over.

The above numbers is a test. The reason I'm posting is I'm curious if the same set works for other states. I just don't have time to test a bunch of states myself. If anyone feels like testing their state, please post the results here. I don't need exact percentages but if you could give me a general idea of how good/bad it worked for your state, I'd appreciate it.

Nice !i like  it and i can double check on certain parameters.So is  Good  chart for someone who knows how to use it !not so helpfull for regualr player since is box play ,draw times 0-10 draws,most the time less then 7 draws i noticed here in Nj after double hits.

Keep up the good work !and build yourself a more powerfull computer .

Tested 20 draw range for ten States and is not working, the box bet makes it a losing strategy. What's the ideal behind those patterns? Try analyzing positions 1 and 2 as ' pattern finder' for just 20 picks for only exact bets.

Nice Job,

Florida,

0: 089, 039, 058, 019, 049, 088, 059, 048, 099, 079
1: 109, 129, 102, 182, 169, 176, 112, 178, 137, 147
2: 268, 208, 210, 218, 230, 260, 258, 200, 263, 288
3: 396, 376, 316, 346, 397, 395, 385, 315, 398, 391
4: 460, 450, 436, 430, 466, 456, 457, 417, 492, 432
5: 597, 567, 569, 537, 547, 506, 583, 513, 543, 541
6: 684, 634, 614, 638, 654, 624, 618, 613, 644, 698
7: 734, 784, 714, 738, 731, 764, 754, 716, 702, 774
8: 861, 801, 851, 836, 856, 866, 847, 897, 826, 846
9: 935, 925, 955, 933, 923, 965, 928, 948, 912, 922

 Wed, Dec 3, 2014 3-1-6?
 Thu, Nov 27, 2014 3-5-1?
 Thu, Nov 13, 2014 1-3-5? Wed, Nov 12, 2014 3-3-9?
 Sat, Nov 15, 2014 3-3-4?

 Wed, Dec 3, 2014 3-1-6
 Fri, Nov 21, 2014 5-6-4?
 Sun, Nov 16, 2014 4-6-6?

 Tue, Dec 9, 2014 1-4-7?
 Fri, Dec 12, 2014 0-1-9
 Fri, Nov 21, 2014 0-1-1?
 Mon, Nov 24, 2014 1-1-8?

 Sun, Dec 7, 2014 9-5-6?
 Mon, Dec 8, 2014 8-3-5?
 Sun, Nov 30, 2014 5-4-5?

 Tue, Dec 9, 2014 1-4-7
 Tue, Dec 2, 2014 4-4-0?
 Sat, Nov 29, 2014 0-4-4?

 Mon, Dec 15, 2014 2-8-6?
 Sat, Dec 6, 2014 2-8-0
 Wed, Dec 3, 2014 2-2-5
 Mon, Dec 1, 2014 2-2-8?

 Tue, Dec 9, 2014 1-4-7?
 Fri, Dec 5, 2014 7-7-7?

Right, I'm not playing it because it hits more boxed than straight. I built it based on FL's data, with no guarantee it would work with other states, I was just curious so I made this post. My current strategy is to find out how much the numbers can be reduced and still keep a high hit percentage. For example, I can give you a pattern that uses 20 sets and will hit boxed within ten days 91% of the time. That's not tracking anything. No VTracs, no sums, no outs, etc. A double hits, do a quick workout, and you have twenty sets. It's not a money maker because it's 20 numbers hitting boxed. Sometimes the hit will come in on the next draw, sometimes it takes the full ten days. The point is the reduction. I want to get to straight hits but I'm not there yet. So reducing 333 box combinations down to 20 is a pretty good start, considering the 91% hit rate.

I can get it down to 10 but then the hit rate drops to 72% over ten days. I haven't tested the hit rate on those sets above, need to write the script for that. Those sets were from my script that tests two positions. I'm testing by hand right now in FL and it's looking promising enough to write a script to do a 5+ year test. I'm going to continue testing to see if I can get ten or less numbers that beat the 72%. Once I find the best method, I'll start working on putting them in order so the boxed hits will turn into straights.

I did some very minor testing on a couple other states and it was hitting. You tested ten states and none of them hit? Mind if I ask which states you tested? Here's Florida's November test.

 11/1/2014 DAY 4 0 4 11/5 EVE 11/1/2014 EVE 0 5 2 11/2/2014 DAY 6 8 4 11/2/2014 EVE 5 7 9 11/3/2014 DAY 0 4 4 11/5 EVE 11/3/2014 EVE 3 1 7 11/4/2014 DAY 2 1 9 11/4/2014 EVE 8 5 1 11/5/2014 DAY 1 5 2 11/5/2014 EVE 5 4 0 11/6/2014 DAY 0 4 6 11/6/2014 EVE 7 3 3 11/13 EVE 11/7/2014 DAY 4 0 7 11/7/2014 EVE 1 0 3 11/8/2014 DAY 6 8 4 11/8/2014 EVE 7 6 2 11/9/2014 DAY 8 8 0 11/17 EVE 11/9/2014 EVE 0 5 7 11/10/2014 DAY 6 3 0 11/10/2014 EVE 5 4 0 11/11/2014 DAY 6 6 3 11/17 EVE 11/11/2014 EVE 0 8 7 11/12/2014 DAY 4 7 2 11/12/2014 EVE 3 3 9 11/13 EVE 11/13/2014 DAY 8 3 2 11/13/2014 EVE 1 3 5 11/14/2014 DAY 1 1 8 Miss 11/14/2014 EVE 4 6 0 11/15/2014 DAY 2 7 3 11/15/2014 EVE 3 3 4 11/27 DAY Too long to play? 11/16/2014 DAY 6 2 7 11/16/2014 EVE 4 6 6 11/17 EVE 11/17/2014 DAY 1 0 4 11/17/2014 EVE 8 6 3 11/18/2014 DAY 3 7 0 11/18/2014 EVE 4 8 2 11/19/2014 DAY 6 5 5 11/25 EVE 11/19/2014 EVE 1 5 7 11/20/2014 DAY 9 8 4 11/20/2014 EVE 3 8 0 11/21/2014 DAY 5 6 4 11/21/2014 EVE 0 1 1 12/9 hit but that's a long time to play… 11/22/2014 DAY 4 9 2 11/22/2014 EVE 3 0 5 11/23/2014 DAY 8 0 7 11/23/2014 EVE 8 1 0 11/24/2014 DAY 6 4 7 11/24/2014 EVE 1 1 8 12/9 still too long to play 1s aren't my friend in November 11/25/2014 DAY 9 2 4 11/25/2014 EVE 1 5 4 11/26/2014 DAY 7 5 9 11/26/2014 EVE 3 0 8 11/27/2014 DAY 3 5 1 11/27/2014 EVE 4 8 7 11/28/2014 DAY 5 5 7 11/29 EVE 11/28/2014 EVE 9 2 4 11/29/2014 DAY 0 4 4 12/2 DAY 11/29/2014 EVE 3 7 5 11/30/2014 DAY 2 9 3 11/30/2014 EVE 5 4 5 12/7 DAY

15 Doubles in November. 1 one of them was a total miss. 2 just took to long, wouldn't play a set for that long. One that was a bit too long out, we'll call it a miss. That makes the stats 15 doubles, 4 misses, 11 hits. Which gives us 73.3% hit consistency. So far, no money to be made, but it's a work in progress. Instead of 333 sets, I'm only working with 10 sets now. No guess work. If I can further reduce the sets to 1 or 2, find a way to play them in the right order (for the straight), or a combination of the two. Then I can make money off of it. This is why I kind of disappear for weeks at a time. I go off testing stuff. Soaks up my free time.

Depend on how you interpret the hit ratio, '15 doubles for 11 hits' is just observation not a playable scenario.If  the patterns has time frame from one to twenty draws for a hit, how do you correlate the 11 hits? I may pick a pattern and wage for 10 draws and miss most of 11 hits or may change a pattern each double trigger and still lose money.

State tested> Tx(176,492) TN(745,430) VA, Ark, FL(460, only hit),IL(539),IND(631) MO(639), MD(921,210), most of the hits has long time frame  even for a type draw testing (DAY or NIGHT). Test was done on recent 20 draws, back testing going further back will be bias (conditions are different going forward)

NB> I want you to consider this when convenient, If you can predict 2 digits for position 1 for next 5 draws with NP2, you'll hit straight most of the time, the only variable for reduction is N, P is your permutation for a repeat digit. Everybody approaches prediction different, some by stats and tracking, my approach is conceptual and assumptions, I think we can find common ground.

FL

Tue, Dec 16, 2014 Mon, Dec 15, 2014 Sun, Dec 14, 2014 Sat, Dec 13, 2014 Drawing Date Pick 3 Pick 4 Midday Evening Midday Evening 4-6-0 5-5-3-0 2-8-6 7-4-2 9-3-5-5 9-6-2-6 2-4-4 9-7-3 5-8-0-8 2-1-7-6 5-9-8 6-2-1 9-4-2-5 2-7-5-0 3-1-3 0-1-9 9-0-2-0 4-9-3-3 6-6-8 1-7-9 5-2-6-9 3-7-5-4 9-2-8 4-2-1 8-0-7-7 2-8-8-6 1-4-7 5-2-5 1-1-9-2 7-8-5-3 3-3-7 8-3-5 7-2-4-9 2-7-6-9 9-5-6 1-5-7 1-5-6-3 7-4-6-1

Ideal to predict 2 digits for position 1 for next 5 draws, trigger will be position 1 of a draw set.

draw          trigger            next pos 1 digits        prediction         exact hit

956              9                     1,6                         NP2                 157,147,668,179

157              1                     6,0                                                 668

337              3                     9,1                                                 147,928

NB>  For strategic wager with ROI in mind, setting N=10 is not big deal so far as ' degree of certainty' is above 50%, ponder on this if you have time.

adobea78

What is the reason for posting your strategy in Manual's thread as well?

Is there any connection (anything common) between his method and yours?

This is what he wants:

The reason I'm posting is I'm curious if the same set works for other states. I just don't have time to test a bunch of states myself. If anyone feels like testing their state, please post the results here. I don't need exact percentages but if you could give me a general idea of how good/bad it worked for your state, I'd appreciate it.

Florida,

Sun, Dec 14, 2014 2-4-4
Tue, Dec 16, 2014 4-6-0

4: 460, 450, 436, 430, 466, 456, 457, 417, 492, 432

I don't understand your post, the addressee of my post, Manual ,has not complained or deemed it abusive, so why this comment from you?

NB> I want you to consider this when convenient, If you can predict 2 digits for position 1 for next 5 draws with NP2, you'll hit straight most of the time, the only variable for reduction is N, P is your permutation for a repeat digit. Everybody approaches prediction different, some by stats and tracking, my approach is conceptual and assumptions, I think we can find common ground.

Manual ask folks to test some patterns currently in progress, he did ask for opinions to filtering to playable size for straight picks, so how am intruding into his space(ie your inference) when I asked him to consider an ideal at HIS CONVENIENCE!. Manual, is a programmer, and has coded an ideal for me before, so sharing my ideal is selfless , which could be a win-win situation. Why this contention, when the forum is for sharing ideals?

The reason I'm posting is I'm curious if the same set works for other states. I just don't have time to test a bunch of states myself. If anyone feels like testing their state, please post the results here. I don't need exact percentages but if you could give me a general idea of how good/bad it worked for your state, I'd appreciate it.

There's a lot to read into a statement, you can be rigid and interpret the above as  ' yes or no' response, but in discussions, is called digression, he been a programmer may re-valuate his assumptions. I hope your intention is not prevent free expression.

WB Manual

Have you tried to program differently? The regular way is not to follow if you run out of memory. I don't really do pick 3, but found that several ways of programming just fail in cases. In one case Excel beats regular OOP. For pick 3 I never considered boxed, I'd rather play couples as alternative. Can you score on front or rear couples, not boxed?

I am not talking about your 1st post, but next one I replied to.

First one is completely fine, but 2nd one has nothing to do with his strategy and testing his strategy in other states, he is looking for.

 Tue, Dec 16, 2014 4-6-0 6-6-2

Double six evening draw in Florida.

Next few days,

6: 684, 634, 614, 638, 654, 624, 618, 613, 644, 698

Unless you intend on  playing in other states, I wouldn't worry about that. What works in your state, may not work in mine and vice versa. Just try to perfect it for where you are going to play most often. Cool system!!!

Depend on how you interpret the hit ratio, '15 doubles for 11 hits' is just observation not a playable scenario.If  the patterns has time frame from one to twenty draws for a hit, how do you correlate the 11 hits? I may pick a pattern and wage for 10 draws and miss most of 11 hits or may change a pattern each double trigger and still lose money.

Keep in mind, I'm saying this method is a system. It's just a method to reduce numbers. I'm working on my 'system' in a programming approach called Agile. First, I want to reduce the numbers to a playable amount (in my mind, that's ten or less) while still retaining a high hit percentage. That's what I'm working on now. After that, I'll work on turning the boxed hits into straights. As for the hit percentage, I should have been more clear. That 73.3% for November is not a true statistic, just more of a glance to show what I'm seeing with the testing I'm doing. On the other hand, the twenty number patter I mentioned has a hit rate of 92%. I've tested that over five years and it's an overall statistic. Year over year, it may fluctuate up or down but if you were to play it, you could expect a hit 92% of the time, if playing twenty numbers over ten days. Wouldn't make money because the numbers have a fair chance of hitting boxed. Plus, who wants to play twenty numbers...

When I go through and test, I do a lot of it 'by hand', testing patterns by going trough the data myself. If something seems to be working, I write a script to test the pattern out to see how much it hits over 5 days, 10 days, and so on. I start testing with 1 year and go out to 5 (20 sometimes if the script executes quick enough). By coding and using SQL to do my testing, I remove bias and human error. The script goes through, finds a double (since that's the trigger I like to use), applies the current pattern I'm testing, then looks to see if it would have hit in the next ten days. It's a fair amount of work but I like to see the outcome. I think it's amazing that some patters get 167 out of a 1000 for some patterns. Others, that may be very similar, can get up into the 700s hits. If a test goes back 5+ years with similar results, the odds that it suddenly changes is not likely. So I'm fairly confident when I say something hits at 72%, it will. I can back it up with years worth of data. Good question. :)

NB>  For strategic wager with ROI in mind, setting N=10 is not big deal so far as ' degree of certainty' is above 50%, ponder on this if you have time.

You lost me... :( Can you explain that a little better?

mmx1: I don't really mind. Long as the thread doesn't get totally derailed. Thank you. :)

onlymoney: Looks like you're seeing hits here too.  What are the question marks for?

Have you tried to program differently? The regular way is not to follow if you run out of memory. I don't really do pick 3, but found that several ways of programming just fail in cases. In one case Excel beats regular OOP. For pick 3 I never considered boxed, I'd rather play couples as alternative. Can you score on front or rear couples, not boxed?

I may be able to work around it, but it'll go much slower. I don't care for boxed either but it's a smaller starting set. Figured if I could get 333 down to ten or less, then I could worry about ordering. Haven't tried font/back pairs. Might look into that, could help with ordering.

Unless you intend on  playing in other states, I wouldn't worry about that. What works in your state, may not work in mine and vice versa. Just try to perfect it for where you are going to play most often. Cool system!!!

Well, I know it works for FL. I figured if it worked elsewhere, it might help some reduce their playable numbers. Plus, it never hurts to share, discuss, test.

Welcome Guest

Your last visit: Sun, Sep 26, 2021, 7:41 am

# Log In

Forgot your username?

Forgot your password?