Problematic positions

This may very well apply to other games but I'm concentrating on Pick 3; better odds and all that. I'm curious what other system creator experiences are with the three positions. Position 1 seems to be the easiest to guess at. In all of my trials, this one seems to hit the most often. I'll use a very simplified version of my Deviation System for the example. First, an explanation of what you're looking at:

Pos1/Pos2/Pos3: These are just the winning draws that occurred on the date to the left (for FL).

Dev1-/Dev2-/Dev3-: Simply subtract the number is Pos X from the number above it. For example, for the second row. 2 - 4 = 8 (using lotto math)

Dev1+/Dev2+/Dev3+: Same as the Dev- (above) but it's adding the two positions. Using the same example. 2+4 = 6

Set 1: This just takes each deviation and adds it to the parallel position. So Pos1 + Dev1-. Same example, 2 + 8 = 0 (let's just assume lotto math everywhere)

Set 2: Same as Set 1 but it's using Dev1+. So 2 + 6 = 8

Set 3: Here, I'm subtracting Dev1+ from the parallel position. So 2 - 6 = 6

So Set 1, Set 2, and Set 3 would be the 'playable' numbers.

First, of course, the hits (I'm only looking at straights):
12/9: 8-7-2 hits on 12/16 EVE
12/10: 9-5-3 hits on 12/14 MID
12/11: 8-2-9 doesn't hit until 1/23 EVE
12/14: 7-9-9 finally hits on 1/15 MID
12/16: 2-2-3 gets a hit on 1/1 MID
12/20: 8-8-0 takes it sweet time... 1/29 EVE

I'll stop here and move on to the misses:

12/12: 460 on 12/30 Mid, 894 12/15, 872 12/16 - That sets the score at Pos 2: 1, Pos 3: 2
12/13: c'mon… 718 hits on 12/17, 12/19, and 12/22… moving along - Pos 2: 2, Pos 3: suck it
12/15: 460 and 400 hit on 12/30

I could go on but you get this idea. Forgive the simplified system here, I just needed an example and I know this one the best. Regardless of the system, the question remains. Do you have a column that's easy to predict? Why? And more likely: Do you have a column that's hard to predict? The only answer as to why, that I can come up with is that Position 1 is drawn first, Position 2 second, Position 3 is third. For some reason, that may impact the randomness. If nothing else, they're put into the machine in numerical order. Position 1 gets less time to bounce around. So if that's the case, wouldn't it make sense to stack the picks that are played? For example, if your system gives you a pick of 1-2-3 and you want to increase your odds, you'd +/- each position. So something like this:

But, if Position 1 is typically easier to predict, why not stack the picks like this:

It would give less numbers to play and could possibly increase the odds. Food for thought. I'd love to hear any response.

No comments? Bummer

In any case, I've done some testing. Tried various combinations from +/- each position to overloarding each position. So far, it looks like a simple +/- to each postion gives the best results.

 I read it manual, I just didn't know how to help you since I don't know about programming

I'm good at testing it though! lol

I am not a programmer , but I understantand  permutations.  You try incremental approach bearing in mind  nCr and nPr relation.

Now with 3 digits the permuation is straight forward  3P2 for 6 combos, increase to 4P2  for 12 combos 5P2 for 20 , 6P2 for 30 ...........P(n,r). I want you to consider this logic and modify certain parameters in your data: Supposing you can generate 2 digits out of 3 always (you must have a trigger /code to locate that), then your missing digits becomes optional,  someone beting on distinct play may have the option of deleting  3digits of 3 P2 from the pool  using remaining pool members(7 left), those beting on doubles could consider the highest member of  3P2 as limit of x, eg if my triad for 3P2 is  437 my x is 0 to 7(highest). You may argue that 3P2 is 60 picks considering x as 0-9, but you don't have to do that, you have to be sectional .

How many pairs you get from nPr(increase or decrease) with code?  How frequent are these Pairs?  if this section is addressed, then x becomes arbitary(many ways of deducing x).

This just my simple humble opinion, good luck.

Position 1 seems to be the easiest to guess at. In all of my trials, this one seems to hit the most often.

ΨΗΑΤ?

How can one position hit more then the others?

This would be a perfect setup for Carlig to do. She is very good at her trade. Right Carli

I took the avedev of the last 2 NM draws.

I rounded up the 3.5 to 4 and did a +111 rundown.  I used this method recently and it works.

NM draws from last week. Using the avedev function again.

Now I changed to a +123 RUNDOWN.

aND A HIT 560 TWO DRAWS LATER.

Hey, I don't have an explanation for it. That's the reason for this post. I've just gone through several versions of deviations at this point and position 1 seems to hit, spot on, more often than the other two. Position 3 seems to be the worst. It could be one off or five... Just using the above example, the exact single set hit several times withing a month. So $30-$40 would equal a win enough to make a profit except for Position 3 would typically be the culprit for a loss, being off just one.

Is your deviation system specific for a State (FL) or  P3 in general? Are you using historical data in locating Pos 1,2,3? I  am just  curious. If positions are based on historical data, then you need to give a thought about finding  a code or method to generate the first two positions based on draw set which is independent of historical draw. Let me put it this way, say I have a system that generates 3 digits  based on drawn set , and these digits is frequently in positions 1 and 2 , then I will be curious to find a parameter to code this. I believe you are on to something , but historical data would'nt cut it.(digits are random). 

Depending on the state short term historical data will do just fine. 

In NM digits repeat several times in the same position.   Look at the number 2 hitting more than once in position 1 and 3. And the number 7 hit postion one four times since Jan 11th.

I'm using historical data. Thus far, just with FL's data because that's where I live and that's the data I have loaded in my local database. The system could be applied to any state/pick 3 game. However, it does require historical data.

It's very true that digits are random. But, I can run all kinds of tests rather quickly with SQL queries. With that in mind, I can check what deviations are occurring for any given time frame. Some deviations occur more than others, for each position, depending on the time frame used. So, I take the top deviations that are occurring for each position and apply that to the last drawn number. When the top occurring deviations hit together (or close to it if you go +/- whatever), a straight hits. I don't count boxes because that's just luck with this system.

So what you are saying is that position one is more predictable than the other positions?  Not that it hits more because that would be

I did not read the whole thing because too much math is a turn off for me.

I like to just K.I.S. , and i think random is kind of broken in the pick 3, as opposed to the pick 5.

Pick 3 is pseudorandom because of the three machines.

What ever that means.

IN NM and other states the pairs repeat(see the 17 flipped to 71) .  I make a list of recent pairs and then decide on the  freguency as to when and how to use them.

Hi manual, you are on to something,  and like your approach, but want to stretch a bit further. What's the difference between  this deviation system  from  matrix (arrays) using the digit analysis of each state? The digit analysis gives the positions of digits by frequency within certain time frame(You can set range), with this you can set up your matrix 

I was on your site testing 'manual test'( I presume not deviation method), I had  some interesting found, not by entering the draw set, I had to work manually on the set to generate my pairs from  3 digits, which  I then entered and filtered and is promising. You should  have a  a data base( need to generate it manually) , which  should be static, the relation between the drawn digits will pick the best pair.  I want you to do a simple glance/test with the digits 0 and 4.  look at any draw set , if the sequential difference is 0, your best set  is 291 (29-21-91), if is 4 ,your set is 465(46-45-65), so if I had a draw 881(seq. diff 0,1) , I will not enter 188 in the system, but 291(I am not using historical data, I am working with the pool, but in sections!). Now do this test and you'll  see 291 or 465 is mostly in positions 1 and 2. 

Forum is meant for ideal sharing, let me know what u think.

 Your data BASE should be  at most 10 triads