Sangoma-number divination

[Springbok]

Frequency of hits. I don't know about this one. You are approaching the hot and or cold number territory there. You can try and experiment and run tests to see whatsup. This is a well-trodden path with many devotees. An interesting variation of the n+1(mark 2) is the ignore the last one or two draws and run a stats report. Predict one or two number pairs. If they don't hit in the next one or two draws then they should hit in the second or third. This should improve accuracy in hitting in the very next draw. I hope to predict the UK Thunderball in 2 numbers for this Saturday. I have been creating numerous files with Thunderball pairings and will post a forecast on the Thunderball thread if I can find the scenarios I am looking for. By the way I feel Endbox needs some assistance in using your software. I must say writing up posts on this site and putting thoughts into print has improved my thinking 1000%. Researching lotteries tend to be an solitary occupation. If you talk to friends and colleagues they think you are mad as "numbers are random, old chap".
Seeing I have hundreds of extra real draws maybe I should raise the barrier for historical highs up to the 600 mark. I will avoid simulated draws at the moment.
I just want to mention, whenever you come across a scenario where you see 0 0 in the last two draws, it eliminates that number pair from appearing in the next draw. Using this technique I have been able to eliminate up to 15 to 16 real numbers in the UK 6/49 lottery. Very rarely do number pairs score three consecutive hits in a row.

 

[Patrick123]

In one sense, it is approaching hot/cold territory, but I think I've explained myself incorrectly.
Where you have a few sets of pairings labeled alpha beta .. through to omega and each one of these has the numbers in pairs as alpha #1, alpha#2 etc.
In my program, by simply using the alpha group, setting the alpha number to 2 and alpha max to 5. I would have the equivalent of the first 5 of your sets for lack of using alpha, beta, etc, I call them alpha set 1, set 2, set 3 and each of these sets has the complete number pairings within which you can see under ball analysis, then clicking the numbers 1-5 with alpha selected in the combination box.
I've expanded this to allow you up to 100 of these sets.
To add to this if you wanted you study triplets and quads simultaneously, you can set the beta group to the number 3 and Delta to 4. What I do, is to set Beta and Delta to 2 as well thus I can work with a potential of 300 of your sets. Alpha max is the number of sets I want to create.
As in your post #80, Omega 16 contains the numbers 31 & 32, whereas kilo 23 contains 29 & 31. My draw probability screen would show a count(weight) of 2 for 31 and ones for 29 and 32 this is simply expanded accordingly based off the Alpha Max setting.

So where you relate to Gamma #3, I would relate to my alpha group set number 5, pairing number 3.

I hope this makes it a bit clearer.

Regards
Patrick

 

[Patrick123]

I must agree with you, being able communicate and share thoughts with someone of similar interest helps to keep your thoughts on the right track, in stead of the usual Lone Ranger style. I've put off doing any real documentation (sorry endbox and excellence) as there is still a lot of chops and changes taking place.

Regards
Patrick

 

[Springbok]

I am going to have a close look at your software this weekend. I am rather excited at my Thunderball prediction I had to search through 5 files before I found it. It seems too good to be true. I still have another 10 files to search. I want to find a confirmation from another file.

 

[endbox]

Hi Patrick123

No need for appologies , your software and this whole topic is fascinating stuff
Take care endbox

 

[Patrick123]

Well, I'm not going to say any numbers for the Thunderball as this will usually start putting doubt in your mind. Rather, the best of luck with your prediction.

I've modified the program again so that you can select a value from n+10>=H down to n+0>=H Modify the value on the Ball Analysis screen then click Draw probability to see the outcome.

Thanks for your patience endbox, I hope it will be rewarded.
Regards
Patrick

 

[Springbok]

Don't worry about doubts in my mind. All we do is using the best available evidence to predict. If it fails we still win as we can gain valuable feedback. How many times have people jumped off cliffs to try and fly. They definitely got feedback. If you don't predict say the thunderball then you miss the chance of feedback. Advantage of posting a prediction is that it makes you very careful that you have made no errors and you have been as careful as possible. If your prediction is not the same as mine we can then discuss why there are differences. I believe my prediction is made with the best available evidence I have so far. I am happy to hear opposing views. Remember " He Who Dares, Wins "

 

[Patrick123]

Well here goes for the Thunderball:
3 - 81% the pair 3 & 15 has reached an all time high for the last 673 draws.
14 - 63% - the pair 14 & 15 have done the same.
15 - 46%
17 - 25%
23 - 53%
24 - 52%
27 - 32%

As for a full row, the program has not progressed that far yet.

Regards
Patrick

 

[Springbok]

Thunderballs go from 1 to 14. There is no 15.

 

[Patrick123]

My apologies, I was predicting the numbers.
The thunderball portion comes through at:
9 - 40%
11- 44%
13 - 42%
14 - 50%

 

[Patrick123]

6 - 62% - 40 of 240 hits
8 - 83% - 10 of 240
9 - 75% - 35
15 - 43% - 13
17 - 70% - 27
27 - 56% - 11
29 - 78% - 30
31 - 60% - 29
32 - 59% - 19
37 - 59% - 17
44 - 70% - 7

I would assume that the frequency of hits would also be an indication of the confidence level.

The definition of hits here, is how many times my program identified the number had the right criteria.

Regards
Patrick

It looks like I must also take into account the numbers of 'hits'.
Regards
patrick

 

[Patrick123]

I was thinking whether I should simply take the percentage & multiply by the # of hits divided by total alpha. In the above exercise, 9 would have shown a 75% * 35/240 = 10.94% followed by 6 at 10.33% with 44 dropping down to only 2.04% or if I need to bias the prediction towards the percentage probability side or the number of hits side?

 

[Excellence]

No need for excuses Patrick. Your software is still evolving so there is no sense in creating documentation right now. I must say that in this period I am very busy so I must catch up whit the post when I can. I hope to give the software a close look this weekend and create the complete picture of what it does (I will probably understand a lot more about the frequency of hits).
Regarding the number of draws analyzed. I observed that the number increased from the last 200 draws to, well more (hope I got that right). If we consider more draws the maximum skip value would naturally increase in most of the cases. Playing numbers that have a skip close to their maximum skip value is like the oldest strategy in lotto games. What I considered innovating in the systems, although at the beginning I was against it, is that the number of draws analyzed is relatively small. Using smaller amount of draws helps finding a certain pattern that we are looking for. It is hard to believe that a number that has the following last 3 sorted skips like this: 13, 14, 18 would get closer to the 18 skip value (n+4, n+2 is even more improbable).
Therefore my believes are: use no more than 300 draws and as lower the probability calculated from the last n+4 or n+2 is, as better (valid for the minimum 300 draws case). Keep in touch, hopefully.

 

[Springbok]

Well call me Suzi Mtetwe and a supper of roast computers . Omega16 has developed rogue tendencies. It has surpassed it previous 400 historical high twice in the last two draws. An 800 draw shows a high of 16 and it surely must come in on Saturday. I have have not been betting on it but will put some money on it for the next 2 draws. I mean surely it cannot beat an 800 draw high. Number pairs do show deviant tendencies on occasions and this appears to be the case. Nevertheless I will profit from it.

* omega # 16 * Hits: 221
------------------------
* Skips >>>>> 14 0 2 3 7 1 3 2 2 0 1 0 2 6 0 1 3 2 0 2 1 0 0 0 2 3 5 1 0 8 6 2 1 0 0 5 1 0 6 0 3 1 5 0 3 4 0 8 1 0 4 5 5 3 0 1 7 3 2 0 1 1 2 7 0 3 0 7 8 0 3 5 0 3 5 4 2 5 2 2 4 0 0 1 2 2 0 2 5 11 1 7 0 0 1 0 1 6 1 0 3 2 2 0 1 1 2 2 0 1 4 7 4 0 6 3 13 2 0 2 0 2 9 0 0 7 1 0 3 0 3 4 4 0 0 3 0 2 2 0 5 1 0 9 0 5 3 1 1 6 7 1 5 7 1 4 4 7 1 1 6 4 10 0 1 8 6 0 0 0 4 3 2 0 1 0 8 0 1 2 1 9 0 0 0 1 3 0 2 4 1 0 4 1 1 0 0 8 4 3 2 0 1 1 0 2 0 8 1 5 5 5 3 3 3 0 16 1 5 4 0 * Sorted Skips: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 6 6 6 6 6 6 6 6 7 7 7 7 7 7 7 7 7 7 8 8 8 8 8 8 8 9 9 9 10 11 13 14 16 ** Skip Median: 2

 

[Springbok]

We are going to look at catching the low skip numbers and what type of scenario we should be looking at. We now take note of the median. here the median is 2. That means 50% of the time alpha2 hits within 3 draws(the count starts at 0) . The ongoing skip at the moment is 1. In the previous 2 occasions alpha2 had 5 skips before hitting. This has now happened twice over the median consecutively. Normally a rebalancing should occur and it is due to strike on or below the median. the probability is this number should hit next draw if this rebalancing should occur. In this case it did occur. So you are looking for a number, one draw below the median and has struck two or more times above the median in previous draws.

* alpha2 * Hits: 21
------------------------
* Skips >>>>> 1 5 5 1 6 0 21 0 12 1 3 7 2 0 1 2 1 1 1 2 4 * Sorted Skips: 0 0 0 1 1 1 1 1 1 1 2 2 2 3 4 5 5 6 7 12 21 ** Skip Median: 2

 

[Patrick123]

Thanks Excellence. That is the beauty of the program, you can decide how much historical information to process, as the balls short term characteristics are different from the long term. The portion of the program to identify the due ball that is reaching it's maximum skip value is basically there, and I'm relatively confident that we can start getting it, so we can start attacking the mid, to high frequency balls. Thus I see the full prediction should be broken down into about three sections, The highly due balls that have not been drawn for a while, The balls that were drawn a short while back, and the balls that were drawn within the last 3 or 4 draws.
There is still a lot of testing that needs to take place to to see how a 30,50,200, etc draw analysis impacts the prediction.

I see Springbok is now starting with the next series of instruction, and will start adding this into the program.

Regards
Patrick

 

[Springbok]

Hi Patrick123
There are approximately 1176 pairings out of 49 numbers. Could your programme
work out all those pairings and report on them? With my 14 files I cover 336 pairings. I may be missing some absolute gems which may be present in the missing 840 pairings.

 

[Patrick123]

Hi Springbok, that thought had crossed my mind instead of creating random pairs. I dropped the idea purely because of memory constraints, where I've been storing everything in memory. This is also the main reason why every time I ran an analysis I'd get similar though not exact answers whereas all 1176 pairing should give an exact and reproducible result.

I think it's now time to evolve the program slightly and incorporate a database into it with all the pairing possibilities.
I have picked up that I had a slight flaw in converting the numbers into their equivalent pairing number. I'll be releasing an update within an hour or so.

 

[Springbok]

I created a programme in VB 6 which covered all the 576(I think) pairings for the 5/34 part of the Thunderball. It was huge as I was using my backveld techniques. Vb 6 could not handle the size of this file. If you printed it out it would have stretched from Egoli to Pretoria My programming skills are a bit more sophisticated now and I could reduce the size to 100th of its original length. See what you can do and find out if it is actually an advantage to use all possible pairs for analysis.I am seriously thinking of building a pure number crunching computer using a quad core processor with a 64bit operating system with many, many gigs of memory. Only problem is there a language out there where you can use it to write 4 threads, one for each processor.

 

[Patrick123]

Yep, my train of thought was not there for a minute, until I sat and thought about the issue, I was still thinking of the constraints that 100 random pairing groups puts on the system which is equivalent to 24 pairs * 100 which is already over the 1176 pairs.
I'll be modifying the program, but will be busy the week-end so expect it about Monday, Tuesday.
Regards
Patrick